Directional microphone technology has stood the test of time for 40 years as one of the few methods that improves speech understanding in background noise. This benefit has been supported in evidence-based reviews,1 but improvements in directional technology are still needed.
Most modern directional hearing aids are automatic and adaptive. Automatic ones are based on the signal detected by the classification system, and directional technology is activated or deactivated (omnidirectional amplification) for particular listening situations. In adaptive hearing aids, the directional polar pattern is automatically altered to maximize speech understanding for different speech or noise conditions, which now even includes switching to anti-cardioid when speech is from behind the hearing aid user.2
When counseling patients on the benefits of directional technology, we often mention specific listening situations where benefit can be expected—parties, noisy restaurants, large groups. In general, we talk about listening situations where the speech and noise are relatively loud, approximately 70 dB SPL or greater. But there are many troublesome listening situations, also with an adverse signal-to-noise ratio, where the speech and noise signals are not as intense, maybe only 60 or 50 dB SPL, or even softer.
THE TWO FACES OF NOISE
We all know that directional microphone hearing aids are designed to reduce noise, and, they do this quite well in many listening situations. Ironically, though, directional technology also creates noise at the same time it is attempting to reduce it because its low-frequency sensitivity is less than an omnidirectional microphone. The noise is an unavoidable consequence of the “delay-and-sum” procedure employed by directional microphones, something often referred to as a “low-frequency roll-off” because of its appearance in the standard coupler response curve.
If the hearing aid is to have essentially the same amount of gain in the low frequencies in the directional mode as it does for the omnidirectional setting, then additional amplifier gain must be applied in the low frequencies to compensate. Depending on the design of the product, this could be an additional 10-15 dB or more.
While this process will then provide adequate gain for important incoming low-frequency speech signals, it also will amplify the microphone's internal noise by an equal amount, also centered in the lower frequencies. It's possible that the amplified microphone noise will be louder than the external noise in some listening situations.
Because of this internal noise, directional microphones are typically engaged automatically only when the ambient noise floor exceeds a certain intensity level, usually around 55-60 dB SPL. Some manufacturers use even higher activation thresholds. This prevents users from hearing the amplified microphone noise that can be louder than the environmental noise.
Even at a setting around 60 dB SPL, directional microphone noise can still be problematic for some users who have good low-frequency hearing. Because traditional directional instruments produce higher noise levels, users often can only take advantage of directionality when the noise floor of the listening situation is relatively high. This limitation needs to be addressed because directionality can be quite beneficial for many soft speech and soft noise listening situations.
SOFT LEVEL DIRECTIVITY
Soft Level Directivity can improve the signal-to-noise when the speech and noise are soft. Directional microphones can be activated automatically at a much lower noise level, and the extent of directivity is adjusted to keep the microphone noise below the environmental noise level. Because less directivity means less directional microphone noise, the softer the noise level, the less directional the hearing instrument will be. (Figure 1.)
When the environmental noise floor drops below even the omnidirectional microphone noise level, the user will hear the microphone noise. Whenever the ambient noise floor is between the noise level of omnidirectional and full directional microphone noise, however, the user can benefit from Soft Level Directivity because the microphone noise does not exceed the noise floor. The user can take advantage of at least some directivity in low-noise situations and hopefully achieve better speech intelligibility, but the amplified circuit noise will not be annoying or even noticeable.
This feature works independently in four frequency bands. (Figure 2.) Soft Level Directivity “adds” directivity for ambient levels below the noise level of the full directional microphone. Without Soft Level Directivity, the only choice is omnidirectional or microphone noise at these levels.
Soft Level Directivity is activated by default for any directional mode; there is no need to switch it off for fixed directional modes. It is possible to adjust the activation threshold at which the microphones switch from the omnidirectional to directional mode to take full advantage of this fitting option in the automatic microphone mode and to fine-tune its function.
This value, noted in dB SPL with activation threshold settings of 48, 54, and 60 dB SPL, is the level that the overall environmental noise has to reach before the directional microphone is activated. Most hearing aids equipped with this feature will be able to provide noise-free directionality already below 48 dB. It is important to select 48 dB to get the benefit of Soft Level Directivity in the automatic microphone mode.
The 54 dB SPL default setting is preferred by most patients. A 48 dB SPL threshold works for those who require more directivity in low-level noise. The highest setting of 60 dB SPL is for patients who prefer the omnidirectional mode in most situations, for those who need to hear environmental sounds at average levels from all directions (localization or safety assurance), or those who prefer the greater loudness sensation often present with omnidirectional amplification.3
Electroacoustic testing indicated appropriate automatic switching to the directional pattern for soft inputs, but we also conducted behavioral speech-in-noise testing at the Hearing Aid Laboratory of the University of Iowa to determine the magnitude of the soft-level directional advantage.
Fifteen experienced hearing aid users with bilateral downward sloping sensorineural hearing loss were fitted bilaterally for experimental testing with Siemens Pure 700 and 701 hearing aids. Both hearing aids were in the fixed directional microphone mode, but only the Pure 701 employs Soft Level Directivity.
Wu and Bentler obtained polar plots with the signal cancellation technique for two ambient noise levels.4 (figure 3.) As expected, the hearing aid without Soft Level Directivity shows an omnidirectional pattern for 50 dB, while the directional pattern is already similar to the full cardioid with Soft Level Directivity. Only negligible differences exist at 65 dB between the hearing aids.
The hearing aids were fitted to the NAL-NL2 prescriptive target for each participant for behavioral testing, and real ear insertion gain (REIG) verification was conducted using the Audioscan Verifit probe-microphone system for pink-noise inputs of 55, 65, and 75 dB SPL (using target values derived from the CONNEXX fitting software). Other than Soft Level Directivity, all other parameters were the same for the hearing instruments. Testing was also conducted in the omnidirectional setting to establish a baseline.
The speech material used came from the Hearing In Noise Test (HINT), with aided testing conducted in the soundfield. The conventional speech-shaped background noise of the HINT was presented at 50 dB SPL from 180° azimuth. The HINT sentences (two lists per condition) were presented adaptively from a loudspeaker at 0° azimuth.
The mean HINT reception threshold for speech (RTS) scores is plotted in benefit compared with the mean RTS scores for the omnidirectional condition. (Figure 4.) There was a significant advantage for soft-level directional (p≤001) compared with the conventional directional mode. When Soft Level Directivity was employed, a mean HINT RTS advantage of 6.2 dB was observed (compared with the omnidirectional setting). When the conventional directional microphone was used, there was no significant advantage (p=0.79); only a 0.8 improvement over omnidirectional was observed.
ConnexxFit, an alternative prescriptive fitting method, shows a similar gain to the NAL-NL2 for average and loud inputs but less gain for soft inputs. This modification facilitates acceptance for new hearing aid users during initial acclimatization. The same participants in the Soft Level Directivity test were programmed for ConnexxFit in a separate memory of the Pure 701, and were tested using the HINT with Soft Level Directivity activated.
The mean Soft Level Directivity advantage in this test is now only 2.7 dB. (Figure 5, compared with Figure 4 results.) While this is still superior to omnidirectional (p=.03), it is substantially less than the mean 6.2 dB RTS advantage observed when the NAL-NL2 algorithm was employed. This illustrates the known but sometimes forgotten fitting concept that the desired speech signal must be adequately audible to obtain the full benefits of directional technology. Some patients, however, will not “accept” the amount of gain employed by the NAL-NL1 for soft inputs, and a compromise might be necessary. Ideally, more gain can be added so that the patient will obtain the additional benefit of the directional processing for soft input signals.