Dr. Ross on Hearing Loss
Feedback Cancellation Systems and Open-Ear Hearing Aid Fitting
by Mark Ross, Ph.D.
The presence, or threat, of acoustic feedback has long been one of the major problems in the fitting and wearing of hearing aids. Acoustic feedback occurs when some of the amplified sound leaks from the ear canal and is picked up by the hearing aid microphone and then re-amplified. This starts the cycle of leakage and re-amplification (the "feedback loop") that results in the squeal we know as "acoustic feedback."
The traditional solution for reducing acoustic feedback has been to increase the acoustic seal in the ear canal, usually by fabricating tighter, longer, but often more uncomfortable earmolds. For some hearing-impaired people, particularly those with moderate or moderate-to-severe hearing losses, this may take care of the problem. However, there is a limit to the amount of sound isolation that any earmold can provide; even with the tightest mold; given enough amplification, sound is going to leak from the ear canal and will start the feedback cycle.
This would be particularly true for those people with the most severe hearing losses. They are often unable to achieve the desired amplification targets because of the occurrence of acoustic feedback, no matter how well fit the earmold. To minimize feedback, they will often reduce the gain level of their hearing aids, yet squealing may still occur when they chew, talk, put on a hat, or even comb their hair. These people require an effective solution to their feedback problem that entails more than simply fabricating tighter earmolds.
The Occlusion Effect
People whose hearing losses are less than about 40 dB in the lower frequencies have their own set of issues relating to snug fitting earmolds. One of their perennial complaints when they first start wearing hearing aids is that their own voice sounds "hollow" or "booming," as if they're talking in a barrel. This is due to the acoustic phenomenon known as the "occlusion effect." It occurs when an earmold completely fills the outer portion of the ear canal. What this does is trap the amplified, bone-conducted sound vibrations of a person's own voice in the space between the tip of the earmold and the eardrum. Instead of exiting through the ear canal into the environment as would normally occur, the sound is instead reflected back toward the eardrum, thus increasing the loudness perception of a talker's own voice.
The resulting sound experience can be unpleasant. Compared to a completely open canal, the occlusion effect may boost the low frequency energy by 20 dB or even more. Additionally, these people may feel a sense of pressure or blockage when an earmold is inserted. These auditory experiences can be sufficiently disturbing to cause some people to reject hearing aids, and others to obtain much less benefit than they otherwise could have achieved. While the occlusion effect can also be reduced when an earmold (or canal hearing aid) is inserted deep into the ear canal, right next to the eardrum, this often brings its own set of comfort and wearing problems.
The typical solution for the occlusion effect is to vent an earmold, thus permitting the amplified sound to escape into the environment rather than being directed back into the ear canal. (Note: A vent is a channel drilled through the earmold, extending from its external surface to the tip.) Venting, however, presents its own set of quandaries. A vent is designed to permit sound leakage, but this is precisely what we are trying to avoid when confronting the threat of acoustic feedback. The larger the vent, the more the occlusion effect can be reduced, and this is positive. But the larger the vent, the greater the susceptibility to acoustic feedback, and this is bad.
Often, because the occlusion effect can be so intolerable, people will use vented earmolds and then be forced to reduce the gain of their hearing aids in order to eliminate acoustic feedback. Gain reduction is not an appropriate way to eliminate acoustic feedback. While acoustic feedback can be controlled in this way, it is being achieved by compromising desired amplification goals, particularly in the higher frequencies. Clinicians (and their clients) often find themselves trying to achieve a workable balance -between a tolerable occlusion effect and a minimally acceptable pattern of amplification. This is not something that we should have to compromise on. And, as will be discussed below, we may no longer have to.
In addition to reducing or eliminating the occlusion effect, there are several other relevant acoustic and audiological implications of venting. As already noted, a vented earmold permits the amplified low frequencies to escape from the ear canal. It does this by opening up a less resistant acoustic path for these low frequencies to exit the ear canal rather than continuing forward to the eardrum. In reality, therefore, it is mainly the amplified higher frequencies that are actually transmitted through the middle ear. The larger the vent, the more the low frequencies are shunted out of the ear canal and the more the amplification focus becomes the higher frequencies. These acoustic effects of vented earmolds have long been applied in hearing aid fitting practices. However, vented earmolds not only emphasize the amplified higher frequencies; they also permit the natural reception of the low frequencies directly through the vent to the eardrum.
Here, too, as with the occlusion effect, the reality situation has often necessitated a compromise between the size of the vent for the necessary low frequency reduction and the amplification goal for the higher frequencies. Hearing aid users with relatively good low frequency hearing require a larger vent for maximum low frequency reduction. But since this larger vent increases sound leakage, and thus feedback, it becomes difficult to meet amplification targets for the more impaired higher frequencies. But this is exactly where most of the sound amplification is required for people with this type of hearing loss. The resulting compromise may necessitate more low and less high frequency amplification than is desirable. So it can be a fitting dilemma.
Vents also have a number of advantages in addition to those already reviewed, such as permitting the natural aeration of the ear canal so that it is not perennially moist. To be precise, however, it is not the vent itself that produces these acoustic and non-acoustic effects, but the fact that we are reducing the impact of inserting a foreign body (the earmold) into the ear canal. For people with lesser degrees of hearing loss, a vent is just one step toward obtaining all the acoustic and comfort benefits of a completely "open" ear canal. People with the most severe hearing losses, for whom earmolds will be required for the foreseeable future, need to receive feedback-free amplification at the target output levels. For both these groups, an electronic solution for feedback is necessary in order to realize these goals. Earmold modification by itself will not do it.
Feedback Cancellation (FBC) Systems
Digital signal processing is now permitting us to approach the goal of reaching our amplification goals (or targets) without the limitations imposed by acoustic feedback. The first electronic feedback suppression systems worked by reducing the degree of amplification at the feedback frequencies. Thus, for example, in response to acoustic feedback at some high frequencies, the hearing aid would automatically reduce the overall amplification (gain) at these high frequencies. Or the hearing would "notch out" the offending frequency by markedly reducing the gain around that point. Thus, if the feedback frequency were about 2200 Hz, the gain of the aid would be reduced, perhaps from 2000 to 2400 Hz. While both of these methods worked, in that more hearing aid gain was possible before the squealing point was reached, the consequence was less audibility at frequency locations where the person may have required more.
An optimal feedback cancellation or suppression circuit will reduce acoustic feedback without any undesirable modifications of the hearing aid's frequency response. A number of manufacturers now include this capability in their hearing aids. While each company has its own proprietary algorithm, they all apparently have this one feature in common. They all permit additional gain before the onset of acoustic feedback and they evidently manage this without any modification in the frequency response. Research has shown that it is possible to achieve 10 dB, or possibly even more, overall added gain to a hearing aid before the onset of feedback. From my perspective, this is a significant technological breakthrough in hearing aid design.
Generally, feedback cancellation (FBC) circuits continually monitor the output of the hearing aid to determine whether some portion of the amplified signal contains elements that have the acoustic characteristics of acoustic feedback. When it does, the feedback circuit first determines the frequency, amplitude, and phase of the feedback component and then generates signals of opposite phase that will cancel (or markedly reduce) the feedback component. Since acoustic feedback is often a complex signal (like a tone with a series of harmonics), the cancellation process requires a complex solution, since more than one frequency is involved. This has to be done very quickly and has to be done adaptively. That is, since the characteristics of acoustic feedback often change (when chewing, talking, sitting in an armchair, etc.), the system must continually generate solutions to the changing feedback frequencies.
While an FBC cancellation circuit should not modify the frequency response of a hearing aid, there is the possibility that it may generate some audible distortion during its rapid operation. This is because the system may mistakenly try to eliminate some portion of a desired signal (musical or tonal sounds), thus altering the perceived quality of the signal. A hearing aid user may report hearing an additional tone (I do, when my microwave beeps come on), or a brief exposure to some kind of audible distortion. It is a possibility that consumers should be aware of when trying a hearing aid with one of the newer feedback cancellation circuits.
Given the recent popularity of "open ear" fittings, all of which would likely require some kind of FBC circuit for optimal performance (I counted advertisements for 15 such hearing aids in a recent trade journal), it would be helpful if independent researchers directly compared their effectiveness. The only study that I have been able to find that directly compared the additional gain possible before the onset of feedback from six hearing aids of different manufacturers was one completed at Starkey Laboratories (The Hearing Review, April 2006).
In the Starkey study, three aspects of FBC circuits were compared for six hearing aids from six different manufacturers. In the first comparison, the six aids were compared in regard to the additional gain possible before the onset of feedback. This is the most basic comparison and the one that would most directly affect listeners. The second dimension evaluated was termed "entrainment." This occurs when the FBC system mistakenly tries to cancel a desired tonal input, producing a form of audible distortion. The third area evaluated was the adequacy of the feedback circuit when confronted with an object (such as a phone, hand, hat, etc.) when it was moved closer to or further from the hearing aid. Such movements often produce undesirable sounds in real life.
In the key performance dimension added gain before feedback the results indicated a fairly substantial range of added gain across manufacturers, from 3.5 dB to about 16 dB, with four of the six reaching about 10 dB or more. I find this the most compelling result: being able to add 10 more dB before the onset of feedback can be extremely helpful for many hearing aid users.
The hearing aids also differed in the other two aspects that were evaluated, entrainment and object interference. Here, two of the systems displayed "objectionable" performance during the test (not the same ones for the two dimensions). Another study, this one by GN Resound, also cautions us that some FBC systems may produce unacceptable distortion by-products (added gain before feedback was not reported in this study). We don't know, however, how troublesome these performance dimensions would actually be in a real-life situation. This has yet to be investigated.
What we can conclude from these studies is that while FBC systems do clearly work, there may well be significant differences between the effectiveness of the various systems in different hearing aids. They also apparently differ on the production of artifacts that are perceived as distortion. It would be helpful for consumers if hearing aid companies were required to report, according to some standardized procedure, the results of objective measures of their FBC system. This is now done for other electroacoustic dimensions, in accordance with procedures developed by the American National Standards Institute (ANSI). As the industry moves into wider inclusion of FBC systems in hearing aids, measurements for this hearing aid performance dimension should be reported as well.
It is convenient to discuss the practical implications of an FBC system in terms of two general hearing loss categories, those with severe to profound hearing losses and those with relatively good hearing in the low frequencies. The major advantage for those with the most severe hearing losses is that an FBC system can help them reach target amplification goals without the necessity of tighter earmolds that, in any event, may not do the job. Being able to realize an additional 10 or 15 dB of gain before the onset of feedback may be enough to take care of the problem. For others in this hearing loss category, an FBC system, by controlling acoustic feedback, permits the inclusion of a small pressure vent in the earmold. By equalizing the air pressure on both sides of the earmold, such a vent reduces the "stuffy" feeling and helps to keep the ear canal dry.
There are even more potential advantages of an FBC system for those with mild to moderate hearing losses. Because it is now possible to "open" the ear somewhat, either by using a larger vent in the earmold, or by eliminating the traditional earmold entirely, the occlusion effect can be eliminated without acoustic feedback becoming a limiting factor. Furthermore, the more the ear canal is opened, the greater the extent of unamplified natural sounds that enter the ear canal. Additionally, the more the ear is opened, the more the normal resonance of the external ear canal (about a 15 dB boost centered around 3000 Hz) can contribute to the overall hearing sensation.
The direct reception of external sounds is of particular importance to those people with relatively good hearing in the low frequencies. They no longer need to depend upon the hearing aid to replace the direct sounds that an earmold had blocked from entering their ear canal. In an open ear fitting, all of these factors come into play: the natural perception of the low frequencies, the natural resonance of the ear canal, and the actual amplification pattern of the hearing aid. It takes a real-ear measure in order to observe the sum of all these interacting acoustic effects. In such a measure, through the use of a probe-tube microphone, the actual sound pressure in the ear canal is measured while the hearing aid user is exposed to an appropriate sound stimulus. This is the only way I know of to observe the combined influence of all these factors. Unfortunately, none of the articles I consulted included this measure and I really don't understand why not.
The ultimate in open-ear hearing aids appears to be the thin-tube open fitting. In this type of fitting, a thin, narrow-diameter, almost invisible tubing connects a small behind-the-ear (BTE) hearing aid to some sort of non-occluding ear bud. The marketing for this type of tubing has stressed its invisibility, evidently an appeal that some people find compelling. They should know that the acoustic effect of a narrow-diameter tube is to reduce the intensity of the high frequencies being delivered through the tube. However, the fitting algorithms of hearing aids using the thin-wall tubing include instructions to increase the amount of high frequency amplification in order to compensate for any loss during the transmission process through the narrow tubing. Because of the FBC system, the additional amplification required to make this compensation would not, hopefully, increase the susceptibility to acoustic feedback. Thus, it is claimed that both the advantages of an open ear and an appropriate amplification goal can be achieved. However, it would be very informative if the results of real-ear measures were displayed to verify this assumption. Hearing aid users should know if they are making any kind of acoustical sacrifice in order to achieve their goal of an "invisible" hearing aid. At the present time, given all we know about the acoustical effects of various tubing diameters, the burden is on the manufacturers to demonstrate that thin-tube fittings do not require hearing aid users to make acoustical compromises.
Recently another kind of open-ear fitting has been introduced, one that does not require a thin wall, narrow-diameter tubing to convey the sound from the hearing aid to the ear canal. Or, indeed, any type of tubing at all. This sort of aid may also reduce, if not eliminate, the need for a feedback cancellation system. This is a behind-the-ear (BTE) hearing aid in which the receiver is externalized and inserted into, but does not occlude, the ear canal. Because the microphone is located further away and somewhat isolated from the receiver (being in the ear canal) than it is with conventional hearing aids, susceptibility for acoustic feedback is reduced. Tubing effects can be ignored, since no tubing is involved in the sound delivery path from the receiver right into the ear canal. This arrangement does not affect any of the other features and programs that may or may not be included in a hearing aid; it simply changes the location of the hearing aid's "loudspeaker."
Personally, I find this development intriguing. Twenty-six years ago, I co-authored a research paper (Volta Review, January l980) in which Raymond Cirmo and I externalized the receiver on three different BTE hearing aids and embedded them into full-concha insta-molds. I was the subject. We compared the frequency response and gain before feedback when the aids were worn traditionally and with the receiver embedded in the mold. In addition to a smoother frequency response in the modified aids, we noted that the gain could be increased up to the maximum possible on all three aids without experiencing feedback. The actual increase was 7 to 13 dB though it undoubtedly could have been more. I've been waiting for years to see this concept picked up commercially, and I'm pleased to see that it has been.
By current standards, mine was rather a primitive study, but still it illustrates, I think, the potential advantages of externalizing the receiver (of course only BTE hearing aids would be suitable). With the newer generation of this type of hearing aid, because the receivers are now so small, it is possible to also obtain the acoustic advantages of an open ear fitting. With the further contribution of an FBC system with this type of hearing aid, I do believe that it is now feasible for just about anyone to receive feedback-free amplification at target amplification goals. .
In summary, every once in a while, there is a new development in hearing aids that I think portends a significant improvement in hearing aid performance. One such development is the feedback cancellation (FBC) circuit, the implications of which impact on people with all types and degrees of hearing loss. The other, not really new, but very much improved has been the practice of externalizing the receiver on BTE hearing aids. (Ironically, some of the very first generation of BTE hearing aids included an external button receiver, simply because there was insufficient room for a receiver in the hearing aid case itself.) I'm sure we will be seeing more of both of these developments in future years.