Binaural Beamforming: The Natural Evolution
Tech Topic | May 2015 Hearing Review
This paper reports on a new measurement protocol based on the Hagerman-Oluffson Method that yields a sequential directivity index (sDI), and demonstrates that the Siemens binaural beamformer Narrow Directionality in a new 2-microphone device is as effective as in the RIC form factor—the latter of which has been shown in clinical studies to result in better SRT-in-noise for patients in certain noisy environments than those with normal hearing. Additionally, a new binaural beamformer designed for single-microphone devices, like CICs, is shown to have significantly higher sAI-DI values than relying on omnidirectional processing and the natural pinna effect alone.
With the introduction of bidirectional ear-to-ear wireless audio data transmission, each hearing device in a bilateral pair can now operate with input not only from microphones on its own housing, but it also receives acoustic signals detected and processed by the other hearing device. This has opened up new possibilities for advancements in many aspects of hearing instrument processing—most remarkably binaural beamforming technology.
Ear-to-ear wireless audio transmission allows for an algorithm that can provide a spatial focus on speech signals originating from the sides or the back, in addition to the more common frontal patterns. As illustrated in Figure 1, we have been able to create a binaural beamforming algorithm, which can narrow the directional focus of the beam even further so that essentially everything outside of what is immediately in front of the wearer is attenuated (for a review, see Kamkar-Parsi, Fischer, and Aubreville¹). This technology, commercially marketed as Siemens Narrow Directionality, has been shown in two clinical studies to allow hearing-impaired wearers to have a better speech reception threshold (SRT) in noise than those with normal hearing in certain noisy listening situations (see Figure 2 from Froehlich and Powers²).
The benefit we have observed in clinical studies, however, is not reflected in conventional lab measurements used to quantify directionality, such as the directivity index (DI). This is because standard measurement techniques do not reflect the real-world benefit of modern dynamic beamforming technology.
The primary goal of conventional measurements has been to capture directivity in a controlled and reproducible way, accepting deviations from real-world application where necessary. Following this rationale, directionality has often been measured in anechoic chambers, without playing any frontal target signal, or using pure noise signals as the interferer³—all of which do not represent real-world applications.
The Hagerman-Oluffson Method Using Speech Signals
While clinical study results are highly informative, it is not efficient or feasible to use this efficacy measure to assess the effectiveness of every new directional microphone technology for every hearing aid model. Therefore, a lab measurement utilizing measurement conditions close to real-world environments is necessary. This is the only means by which we can quickly assess and quantify directional performance for new beamforming techniques as they are applied in various hearing aid form factors.
This was the motivation behind the development of a new protocol for assessment of directionality in modern hearing aids, which has the benefit of simulating both a target source and an interfering source at the same time using speech signals.³ This protocol uses the established Hagerman-Oluffson Method, but approximates real-world situations, such as noisy restaurants or cocktail parties where the wearer wants to listen to a specific speech source, but the unwanted interfering noise is other people talking in the background. This new measurement method results in a sequential directivity index or sDI. (Authors’ Note: We use the term “sequential” to reflect the Hagerman-Oluffson approach, where signals with phase-inverted components are measured sequentially in order to obtain the final result.)
It has been common over the years to adjust the traditional DI using the articulation index (AI); this gives a greater weighting to DI values in the frequency ranges that contribute the most to speech understanding. This is then termed the AI-DI. The AI calculations also can be used with the sDI, providing us with the sAI-DI. The sDI and the sDI-AI can be interpreted in the same way as the traditional DI, and AI-DI values, and can also be used to quantify and compare directionality for modern hearing aids.
Figure 3 shows the sDI values of the Siemens receiver-in-canal (RIC) instrument used in the clinical studies described previously. Compared with “TruEar” (a modified omnidirectional pattern to approximate pinna effects for RIC and BTE instruments), the 48-channel automatic adaptive monaural directional performance is significantly better, improving directivity across the frequencies. However, the binaural beamformer Narrow Directionality exhibits even higher directional benefit, especially in the higher frequencies important for speech intelligibility.
Quantifying Directionality of ITE Binaural Beamformers
Since the introduction of Narrow Directionality in RIC hearing instruments, this feature has recently been extended to in-the-ear (ITE) devices in various form factors. This means a much wider segment of patients now have access to this feature. Although the technology is the same in the custom devices as in the RIC instruments, can we expect the same benefit?
Considering the microphone placement and inlet-port spacing differences between these form factors, the benefit observed in one may not directly translate to the other. We know this to be somewhat true with traditional directional technology.
Using the new laboratory protocol just described, we can now easily compare performance of different hearing instruments. Therefore, this new method is especially helpful in predicting the real-world performance of the application of this processing in custom hearing aids.
As seen in Figure 4, the same binaural beamformer Narrow Directionality in an in-the-canal (ITC) instrument shows a similar sDI pattern as that of the RIC instrument values displayed in Figure 3. The lab measurements clearly indicate the significant benefit, previously observed for binaural beamformer technology in RICs, can be translated to other hearing aid form factors, thereby benefitting more patients. In fact, the measured sAI-DI from the ITC is 1.3 dB larger than that observed with the RIC instrument.
Since the ITC is seated in the concha rather than on top of the pinna like the RIC, it stands to reason this difference is the result of the added pinna effect. Although these measurement results represent the directional performance from one specific custom hearing aid form factor, it can be expected that custom products of other sizes, such as full-shell or half-shell, will yield similar results.
Automatic Directionality in Single-microphone CICs
If we trace the history of directional hearing aids, we see that, for the first few decades of their use, the directional processing was accomplished with a single directional microphone, or a single microphone that could be switched from directional to omnidirectional.4 This was possible using a pressure-gradient microphone that was sensitive to sounds from both sides of the membrane via a front and back inlet port. The directional effect was accomplished by placing a delay in the sound path for the rear inlet. The delay needed to be consistent with the time required for sound to travel externally between the inlet ports (for a detailed explanation, see Arentsschild and Frober5 and Nielsen6). Some of these early instruments were always directional, but if the input from the rear inlet was completely closed (eg, with a switch on the hearing aid), the processing became omnidirectional.
In design theory, the single-microphone approach has appeal, as less space is required. However, there is still the issue of two inlet ports and appropriate port spacing. This has limited the application of the directional design to BTEs, and at the smallest, the ITC style.
The question then becomes, with today’s effective wireless streaming of audio signals between hearing aids, would it be possible to develop a directional algorithm using only one microphone in each hearing aid? And if so, since only one input port would now be required, could this technology be applied effectively in small completely-in-the-canal (CIC) instruments? The answer to both questions is “yes.”
Having bilateral CIC instruments with bidirectional ear-to-ear wireless audio data transmission implies that each hearing device in a bilateral pair operates with input detected and processed by both hearing aids, and it is possible to design a binaural adaptive beamforming algorithm that incorporates head shadowing effects. More specifically, by carefully weighting and combining both available microphone signals, and by imposing an appropriate optimization criterion for the adaptive weighting rule, it is possible to generate an enhanced output signal where interfering lateral noise sources can be efficiently attenuated, while the frontal desired speaker signal remains untouched.
As a result, directionality (or more precisely, a beam toward the front direction) is produced, which was never possible in the past with CIC instruments. This of course has the potential to improve speech understanding in adverse noisy environments, presuming the user is able to face the target signal in much the same way as previously accomplished with BTE instruments (Figure 5).
This new binaural single-microphone beamformer can provide significant improvement over the pinna effect alone (Figure 6). Whereas the pinna only provides directionality in the higher frequencies, not only does the binaural beamformer enhance the pinna effect, it also provides substantial directionality for the lower frequencies.
It is also important to note that this new binaural single-microphone beamformer is automatic and adaptive, just like the binaural beamformers in the two-microphone systems. This means it only activates when the listening situation requires it. Otherwise, the device remains in the omnidirectional mode, allowing the wearer to retain more spatial awareness from all around.
Unlike the mechanically generated single-microphone directionality in the past, it does not require wearer interaction. Additionally, because it is entirely accomplished via wireless transmission and signal processing rather than mechanical components, it can be implemented in the smallest CICs.
Modern beamforming technology has long outgrown traditional methods of quantifying directionality. A new measurement protocol (based on the Hagerman-Oluffson Method) has now been developed that is better able to approximate real-world noisy listening situations, such as restaurants. As a result, this is a better indicator of the potential practical benefit for patients in their daily lives.
Using this new protocol, the binaural beamformer Narrow Directionality in the new two-microphone custom device was shown to be as effective as in the RIC form factor—the latter which has been shown in clinical studies to result in better SRT-in-noise for patients in certain noisy environments than those with normal hearing.
Finally, we introduce a new binaural beamformer designed for single-microphone devices such as CICs. This new feature, marketed as Siemens binaural OneMic directionality, offers automatic, adaptive directionality, and is shown to have significantly higher sAI-DI values than relying on omnidirectional processing and the natural pinna effect alone.
A follow up article will describe clinical study results conducted with custom instruments with Narrow Directionality and CICs equipped with binaural OneMic directionality. Both features can be found in Siemens Insio binax hearing aids.
Kamkar-Parsi H, Fischer E, Aubreville M. New binaural strategies for enhanced hearing. Hearing Review. 2014;21(10):42-45. Available at: http://www.hearingreview.com/2014/10/new-binaural-strategies-enhanced-hearing
Powers T, Froehlich M. Clinical results with a new wireless binaural directional hearing system. Hearing Review. 2014;21(11):32-34. Available at: http://www.hearingreview.com/2014/10/clinical-results-new-wireless-binaural-directional-hearing-system
Aubreville M, Petrausch S. Directionality assessment of adaptive binaural beamforming with noise suppression in hearing aids. International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2015. In press.
Mueller HG, Johnson RM. The effects of various front-to-back ratios on the performance of directional hearing aids. J Am Aud Soc. 1979;5(1):30-34.
Arentsschild J, Frober B. Comparative measurements of the effect of a directional microphone in the hearing aid. J Audiologic Technique. 1972;11:27-36.
Nielsen H. A comparison between hearing aids with directional microphone and hearing aids with conventional microphone. Scand Audiol. 1973;2:173-176.
Rebecca Herbig, AuD, is Manager & Editor of Scientific Marketing, and Matthias Froehlich, PhD, is head of Product Management Audiology for Sivantos GmbH, Erlangen, Germany.
Correspondence can be addressed to: email@example.com
Original citation for this article: Herbig, R, Froehlich, M. Binaural Beamforming: The Natural Evolution. Hearing Review. 2015;22(5):24.