The Shaky State of Available Options…

A hearing aid is probably in your future, if you live long enough. And perhaps, even in your present!

This blog shares my misadventures with a hearing aid and explains why the need is near-universal. The following blog offers sound samples and spectral displays, and shows the way to an ideal hearing aid. (That ideal will be far-removed from Madame de Meuron’s picturesque “ear trumpet.”) Here are the main sections:
– Lost: Four Notes!
– My Hearing Aid Experience
– Do We Care About High Frequencies?
– Look at the Data
– What the Data Tells Us

Let’s begin with, not a Lost Chord, but a few Lost Notes…

Lost: Four Notes!

As readers know, I enjoy playing the piano. And sometimes other people enjoy my playing as well.

I have a lovely size M Steinway that has kept me company for twenty-five years. In early 2018 I noticed that something was different: the top four notes on the piano sounded like clinks, not notes. I could hear the percussive sound of the key, but not the musical tone. The highest note I could hear is the key called G#7, which has a frequency of 3322 Hertz. I had lost my hearing above that position.

Of course, hearing a tone is not a yes-or-no thing. Human hearing gradually declines as notes get higher and higher, there’s not a sudden cliff. But subjectively, when I played G#7, I could hear enough of the higher harmonics to perceive it as a musical note. And when I played the next note up, A7, there was a twang of the string but not a tone. So in this case, my frequency cutoff was between those notes, say 3421 Hz.

(Now, almost four years later, I’ve lost two more high notes. My cutoff is now 3048 Hz.)

My Hearing Aid Experience

Now, I know that my hearing is not as acute as it once was. And Nola has nudged me about not hearing words clearly in noisy surroundings. But I could not ignore the clear evidence that I had lost four notes from my Steinway. After all, it only has 88 keys!

So I consulted an audiologist recommended by our family doctor. Because in retrospect I believe they did a poor job, I will allow the audiologist to remain anonymous.

The audiologist had the usual stuff: a soundproof booth and lots of audio equipment. They tested me with pure tones, clear speech, and speech accompanied by noise. And the tests showed that my hearing response was normal up to 1000 Hz, then started dropping. At 4000 Hz, my left ear had sensitivity had fallen by 60 dB and my right ear even more. (FYI, 20 dB is a factor of 100 in reduced sound energy.)

            A Couple of Failures

The audiologist fitted me with a demo pair of “Bi-Cross” hearing aids made by Phonak (CROS and Audéo). With this combination, sound reaching the right ear is sent wirelessly to the left ear, so the left ear gets an amplified dose of sound from both ears. The hearing aid boosts the loudness of the high frequencies, where my hearing is deficient. Because these were up-to-date models, the price for the pair if I decided to buy would be $4195.

Nola and I tested my hearing at four venues: a restaurant; a sports bar with many raised voices; a performance of Pirates of Penzance; and a bar with a loud rock band. My ability to understand Nola’s speech was unchanged, with or without the hearing aid. Moreover, in Pirates I could hear the soprano’s ornamentation equally well with and without the hearing aid. Assessment: zero stars!

I returned the demos and asked for a set of conventional hearing aids instead. Accordingly, they fitted me with a pair of Phonak Audéo models. Moreover, at my request, they unlocked the setting allowing me to adjust the volume. (How annoying: user-available options are turned off by default on all hearing aids!) Nola and I tested this hearing aid and found it to be equally ineffective. Zero stars, again.

            Hindsight is 20-20

Knowing what I know now, what do I wish the audiologist had done? She should have advised me that when a person has significant high-frequency hearing loss, amplifying hearing aids are not useful. What’s needed is to transfer sound energy from high frequencies to lower frequencies. Although the Phonak B90 SoundRecover feature supposedly does this, I don’t think my audiologist knew anything about it, because she never mentioned it to me. She seemed to have only minimal knowledge of the devices she sells.

I mentioned to her what I had read: that the audiologist should test the client with speech and noise when fitting a hearing aid. She said that she doesn’t do that test for a trial fitting. First, you have to buy the hearing aid, or pay them for a separate test. However, if they had run that test, the failure of the aid would have been evident even in their office, saving me the “fun” of testing it myself!

Do We Care About High Frequencies?

You might wonder: do we really care about hearing these high frequencies? After all, men’s voices are mostly in the frequency range of 85 to 180 Hz, while women’s and children’s voices are 165 to 255 Hz. And these ranges are far below the 1000 or 2000 Hz at which significant hearing loss appears in older people.

The problem arises in understanding words. Frequencies below 1000 Hz are adequate to understand vowels very clearly. However, distinguishing words such as “time” and “dime” requires hearing high-frequency sounds that distinguish one consonant from another.

To illustrate this problem, I used my baritone-pitched voice to record several sequences of words, differing only by the final consonant. Here is one of those recordings, for the sequence saa – sab – sac – sad – sap – sat:

Here is a spectrogram that shows the frequency spectrum of each of these words:

A casual viewer would say, hey, those all look pretty much alike! And that’s exactly the point. The cues that distinguish one consonant sound from another are subtle. They depend on features over a wide range of frequencies, especially those above 2000 Hz.

The English language, at least, depends on its consonants for word clarity. So yes, we care about the high frequencies that consonants rely on.

Look at the Data

Is my lousy aging hearing unusual? Sadly, not at all. Consider the following data from Rigters et al. They measured 675 Dutch adults whose mean age was 71.1 years:

In this group, seniors lose about 20 dB of their hearing with each doubling of the frequency above 2000 Hz. The dropoff is sometimes referred to as ski slope hearing loss. And not surprisingly, seniors also lose hearing with advancing age: after age 70, about 10 dB every 12 years.

These are only averages: amazingly, almost 20% of people in their 80s have no loss of high frequency hearing. But unless you’re in that lucky fraction, hearing loss is on its way for you.

Hmm, you might say. What if these Dutch folks are different? Based on many US articles, I think that Americans are the same, although I can’t find an exactly comparable data set here. However, I can find almost the same kind of data reported by Park et al. They compiled data on 15,600 Koreans, screened to represent the entire population of 33,000,000. Their hearing loss data by age, plotted on the same scale as above, appears here:

To create this chart I extracted Park’s data ages lying in the center of Rigter’s age groups. And it yields similar results to those obtained 5,500 miles away in The Netherlands: the seniors lose about 18 dB of hearing with each doubling above 2000 Hz. They lose about 19 dB of hearing with each 16 years of age, which is similar though not exactly the same as in Rigter.

            Hearing Loss by Gender

The Park et al study yields a bonus: data on the differences between men and women in hearing loss with age:

	dB Threshold, Female minus Male at these ages:
	67.5	74.5	83.5
500	-2.5	-3.2	-3.8
1000	-1.9	-3.2	-7.2
2000	0.0	0	0.0
3000	8.9	8.6	7.0
4000	14.0	13.3	10.8
6000	9.5	8.5	6.3

At frequencies from 3000 Hz up, women hear roughly 10 dB better than men of the same age. (The actual average of the 3000, 4000 and 6000 numbers is 9.7 dB.)

Thus we know that the hearing of older people differs by gender in at least two ways:

• Women’s voices are pitched higher, and thus are harder to understand by a person who has lost high-frequency hearing; and
• Women, on average, can hear all voices better than men of the same age.

            Hearing Loss for Art

You might well wonder how Art’s hearing aid test compares with the data above. Here it is, plotted on the same scale, for each of my ears:

Now it’s not at all clear that these numbers directly compare with those given by Rigters and Park. For one thing, the doctor told me that my right ear hearing was flat up to 1 kHz, then dropped off, whereas my left ear was flat up to 3-4 kHz. Whereas these numbers look a lot worse than that. However, generally speaking the steep drop with frequency is consistent with the findings of the researchers.

What the Data Tells Us

The data above suggests the following to me:

• Dropoff with frequency is so profound for older folks that merely amplifying the high frequencies is worthless. Making high pitches thousands or millions of times louder will add random noise. Moreover, the additional pressure delivered to the inner ear may be painful or even damage the ear.
• In addition, it’s difficult to amplify more than about 20 dB without causing feedback noise (whistles). Because many people’s loss of hearing exceeds 20 dB, mere amplification is not enough.
• If high frequencies are critical for speech understanding, and we can’t easily hear high frequencies, we should lower those frequencies.

The last item above gave me an “aha” moment. There are digital apps which can change audio frequencies. The change may be miniscule, so that a pop singer’s recording will sound on pitch. On the other hand, the change may be substantial, to help a singer or instrumentalist record a song in a more convenient key.

I thought: perhaps these modern signal processing tools can also improve speech understanding in a hearing aid. In the next blog I’ll tell you where this “aha” insight took me, in my quest for the ideal hearing aid.

We now know that many older people have such deep hearing loss that hearing amplification isn’t enough. A hearing aid needs higher-tech help, which is the subject of the next blog.

Credits:
– Madame de Meuron with ear trumpet from Wikipedia
– Bulldog from Anna Shvets on pexels.com
– Spectrogram from Edward Ball, Michael Ruiz and Boris Smus on academo.org
– mp3 edits using Audacity.app
– Rigters et al, Progression of Hearing Loss in the Aging Population/ Repeated Auditory Measurements in the Rotterdam Study
– Park et al, Age- and Gender-Related Mean Hearing Threshold in a Highly-Screened Population