Persona Medical has been developing hearing aids for 50 years, so we know a lot about hearing loss! What we know is that the majority of hearing loss we see in first time patients is between 2000 Hz and 8000 Hz. In musical terms, that’s two octaves….starting three octaves above middle C on the piano, and extending to five octaves above middle C (256Hz). 2700Hz is the natural peak resonance for humans. No wonder this is where the damage is done. Thousands of years of evolution carefully crafted our ears to focus on the unvoiced consonants of speech that are difficult to project with the efforts of the pressure of our lungs pushing air through our voice box. It was not preparing itself for the onslaught of noise and volume the 20th Century had in store. Yes, our ears naturally resonate the higher frequencies of conversational speech, but certainly were not intended to resonate a Marshall Amplifier already producing 100dB SPL+ of sound! Did our maker not anticipate the industrial revolution and the associated noise evolution, or did He merely predict the invention of Earasers?! The ear is a natural resonator…but not below 1000Hz and very little beyond 5000Hz. Again….keep in mind, that the majority of music will be below 1000Hz!
The ear is non-linear dynamically speaking, and it does not PREFER a FLAT FREQUENCY RESPONSE. The ear actual re-shapes a flat frequency response by adding its own natural resonance. As a matter of fact, Fletcher and Munson realized in the 1930’s that our ears are non-linear as well. SEE: https://en.wikipedia.org/wiki/Equal-loudness_contours
So, what is the right amount of attenuation an earplug should have? As volume increases, our ears are more sensitive to some sounds than others. We know when we listen to soft music, the bass notes and the very high frequencies seem to get lost. What is definitive is that our most sensitive region for hearing (1000Hz) up to our peak resonances (primary 2700Hz, and secondary 4000Hz), are the most sensitive to loudness growth, and the most risky for hearing loss as the SPL for discomfort is achieved at high levels in this region before either the low frequencies or higher frequencies.
It would be important for high fidelity to recognize the ear is capable of “protecting” itself to some degree, even to a great degree if needed. The acoustic reflex alone can be quite pronounced in the low frequencies reducing 10dB to our cochlea simply by reducing the vibration of the ossicles (bones). So why should we adopt a linear ear plug? How can we mimic the effect of a loud environment without causing damage? Maybe we should be reducing the more dangerous frequencies and take into account the non-linearity of the low and very high frequencies in our sensitivity.
Our ears are able to pick up sounds both through our skull (bone conduction) and through our ear canals. The cochlea is mounted in our skulls. When we are in a loud environment, more of our bone conduction is in use than in a conversational environment. If we apply too much reduction to the low frequencies, our own voice will sound very odd due to the increase in our “head voice” bone conducted energy compared to the energy that travels through our ear drum and into the cochlea. Our brain is constantly monitoring the comparison to the two levels (bone vx. air conduction) and making decisions about the sound quality based on this comparison. That is why you think you sound different on recordings…because the recording doesn’t have the bone conducted energy that you hear through your skull/cochlea pathway. Since there is essentially no occlusion effect above 1200Hz, we need an earplug that doesn’t reduce ANY volume below 1200Hz. OK, so maybe that is a little extreme. What is the purpose of the earplug then? We must assume that the environment is significantly loud, and that some reduction is desired.
Now, a study performed by National Acoustic Laboratories, Australia, (NAL) in 1999 confirmed some additional psycho-acoustic phenomenon that muddies the data even more. It appears that simply placing an earplug in your ear causes an “insertion effect” to desire 7dB more gain in the low frequencies in order to maintain the same perception of low frequencies versus 1000Hz. Fascinating! (This is not to be confused with insertion loss…the loss of naturally transmitted energy to the eardrum). Although researchers have not proven the reason for this effect, it may have something to do with the increase in bone conducted energy versus air conducted energy (skull versus eardrum) and the increased awareness of self-generated body sounds including blood flow to the cochlea. What we can say is that this is another bit of evidence telling us not to reduce the low frequencies as much as we may otherwise assume. But, it is becoming more clear to me that earplugs should not be flat attenuators…maybe flat response shape, but not flat attenuators. Even with earplugs, I don’t think the goal is to reduce the volume to conversational levels (65-70dB SPL) but to make them retain the nuance of the music at a higher level without being objectionable, or destroying the ability to hear the layering of different instruments or sounds, localizing, and certainly reducing the likelihood of hearing damage (evidenced by ringing in the ears).
This is a little understood cause and effect. A recent study showed that 75% or more of young adults experienced ringing in their ears after attending clubs or certain venues where music was playing. Yikes! Typically, ringing in the ears is accompanied by a temporary threshold shift which has been shown to be a pre-cursive indicator to eventual sensorineural hearing loss!
By studying the natural resonance of the average ear canal, we can also see a secondary peak of 9dB at around 4000Hz and gradually reducing to no additional resonance beyond 8000Hz. (Some people also experience the above mentioned “insertion effect” at 3000Hz but not as often as the lower frequencies.) What we can say is that these higher frequencies, above the secondary peak, especially beyond 8000Hz, also show an improved loudness growth function. After all….just add on another octave and you are practically at the maximum higher frequency range of hearing. Loudness growth approaches infinity. In other words, the higher frequencies also have a natural built-in attenuator. But, here is the catch. We know that higher frequencies are rarely sustained for any period of time….at least that which remains audible. The lower threshold of hearing in the higher frequencies is easily masked by high intensity, low frequency sounds. For example, the bass drum or bass guitar will easily overtake the sustained resonance from a high frequency cymbal roll. If we reduce the high frequencies by the same amount as the peak frequencies, we get a very unnatural sound. The higher frequencies have a wider range in which the nerves of the cochlea overlap to receive the signals. Thus, there is more power perceived for noise stimuli in high frequencies than for pure tone stimuli. We assume for this discussion, that much of the music generated above 4000Hz is broader band and not pure tones. For relative reference, the highest piano note only goes to 4186Hz.
These very high frequencies are used for clarity of speech, the unvoiced consonants such as “th” “f” “sss.” These sounds should not be over-attenuated as their ability to be vocalized is limited and thus, even with amplification, they are difficult to reach the uncomfortable levels of most humans. If they are reduced by too much, speech definition is compromised, cymbals become dull and everything that should be snappy, becomes dull. The open, airy feeling is lost. Often times, we associate reverberation with a large room and the lack of reverberation with a very small, closed in container. One is rich in quality, the other lacks it. Because the ear sums the high frequencies due to the way the nerves in the cochlea operate, the equal loudness contours underestimate the frequencies above 1000Hz. In practice, with our situation being music, you can actually raise the equal loudness contours above 1KHz (Suzuki, 2002). This is reason to let the ear do a little more of the work in protecting itself against loud high frequencies…because it has some headroom to spare.
In most musical situations, a normal listener would not enjoy, nor tolerate much more than 10-15dB’s more sound pressure than other normal hearing listeners uncomfortable limit. That 10dB would represent double the volume perceptually. Thus, 10dB seems that a good place to start. If we start with 10dB attenuation at 500Hz…. where do we go at 1KHz and 2.7KHz…etc? Remember that our ears naturally resonate the most at 2700 Hz and the equal loudness contours show the ear is most “sensitive” to loudness at the 2700Hz region. When we amplify sounds, our ears have no way to notch out this region.
The real danger is in the 1K-4KHz region. Now, how much should we adjust those frequencies? Take a look at our natural resonance, about 18dB of amplification at 2700Hz for conversational levels. But, we don’t need that resonance for 95dB SPL sounds! You can see from the equal loudness contours, an attempt by our ears to begin to dampen and restore a flatter response. Our ears just aren’t doing enough to protect us here. Maybe pre-industrial revolution we would be fine, but the intensity of high frequency sounds we generate, and the short and repetitive durations add up to trouble. An earplug which reduces 15-25dB’s at the peak frequency would be desirable if simply to remove the gain from the natural amplifier we carry with us that was, for the sake of survival, intended to amplify soft sounds for hunting, and for the sake of communication, amplify the higher sounds for speech clarity at conversational levels or lower.