The other day, I was setting up a fairly sophisticated and unusual paging voice alarm system and was checking out the electronic head end of the system when a member of the client’s team commented on the residual noise the system made. Because the speakers are located in an area where the background noise level never drops below 60dBA, and most of the time is considerably greater than this, I was surprised that normal electronic residual noise could be heard, unless something was amiss. (I had designed the system to deliver 75 to 95dBA of speech at the listener locations, so hearing some form of residual noise with the system working flat out was possible, though quite unlikely.)
As a first step, I thought I would check out the residual noise level at the amplifier rack. I connected a spare, identical paging speaker to one of the outgoing lines. No obvious hiss, hum or buzz. However, the amplifier rack was located in a fairly noisy computer equipment room. All of the other racks had noisy fans that were permanently in an excited state.
Therefore, I stuck one of my specially tuned, calibrated stereoscopic listening probes (i.e., my right ear) up close to the test loudspeaker to listen to the aforementioned noise. Not a jot, but what I did hear was a very high-frequency tone. Although there was no way this could be heard in the working area, I was intrigued as to what it could be. Because I happened to have my trusty handheld spectrum analyzer with me, I switched it on and was not totally surprised to see the 20kHz bar light up. Clearly, this was the system surveillance/fault monitoring tone that I was seeing and, quite remarkably, hearing.
I know that, although my hearing is better than average, to still be able to hear 20kHz was a little surprising…and, for a few minutes, I felt quite proud. However, common sense prevailed, and I decided it must be a sub-harmonic that I was hearing. Out came the analyzer again but, no, the 20kHz was clearly the dominant component. Then I read the level and discovered that I had been listening to around 100dB at 20kHz, so it’s not so surprising that I could still hear this!
That got me to wondering about our ability to hear high-frequency sounds. As we age (or become more experienced audio engineers, as I like to think of it), our hearing ability naturally decreases as presbycusis sets in. It is estimated that by the age of 65, about 30% to 35% of the population will experience hearing loss. Indeed, around 20% of the population as a whole has noticeable hearing loss. One recent study showed the population percentages affected by hearing loss as follows:
• 11%, ages 44 to 54
• 25%, ages 55 to 64
• 43%, ages 65 to 84
Put into global context, the WHO (World Health Organization) estimates that about 500 million people suffer significant impairment from presbycusis.
Aging, of course, is not the only factor that affects what we might hear. Apart from congenital hearing defects, common drugs, such as aspirin or other ototoxic drugs (including some antibiotics and anti-inflammatory drugs), can affect our hearing, as, indeed, can alcohol, hypertension and diabetes. Smoking has also been linked to hearing loss. This, of course, is in addition to other noise-trauma stresses that we may have subjected our ears to while gaining our audio “experience.”
Interestingly, it is possible to isolate and determine the average presbycustic hearing loss, given the age and sex of a listener. (For some reason, males suffer more age-related hearing loss than females. I’ll leave you to figure out a reason for that!) The average hearing-loss data (i.e., change in hearing threshold) is given ISO 7029.
Looking at the equal loudness curves (a typical example is shown in Figure 1) indicates that even those with “normal” hearing require a signal at 20kHz to be about 15dB more than a 1kHz tone in order to sound equally as loud.
Presbycusis is not a uniform loss of hearing ability over the audible range, but is highly frequency dependent (as, indeed, are most other forms of hearing loss). Most hearing loss data relates to standard audiometric studies, so tends to have a high-frequency limit of 8kHz. Some studies do extend the range to 16kHz or beyond, but there is little information relating to the effect of age at these higher frequencies. (I guess the powers that be do not think it is normal to subject your ear(s) to a 100dB of 20kHz sound!).
Figure 2 shows a typical set of male hearing-loss curves that I have adapted from ISO 7029. As the graphs show, once you have reached your 50s, most people will have a significant loss at 8kHz, and even at 4kHz your hearing is likely to be about 15dB down. These figures relate to the lower end of the 50s range, so by 55, on average, your hearing will be about 22dB down at 4kHz and 31dB down at 8kHz…and that’s assuming that you’ve given up smoking, drinking and other nefarious ototoxic vices.
The problem with hearing is that once it’s gone, it’s gone! It does not recover. Although some research suggests that upping your vitamin intake along with other supplements can slow down the rate of loss, this is still not totally proven. (Perhaps smoking a carrot would be a neutral thing to do, ontologically speaking!) Alternatively, you can just crank up the level. Now, where have I heard that before?
If you would like to learn more about your hearing and its implications for audio perception and intelligibility, why not come to the AES Sound Reinforcement conference (aes.org/conferences/59) being held in Montréal, Canada, in July, where leading hearing expert and researcher Dr. Poppy Crum will be talking about hearing perception, and I will be taking a different look at speech intelligibility. I hope to see you there. Meanwhile, on the basis that mice can hear up to 70kHz or 80kHz (the non smoking, non-aspirin-taking ones, that is), I’m off to see if I can build a better mousetrap.