a, i, and sometimes y

Marshall Chasin
July 21, 2015

Well, not “y” but it sounded better that way.  Y is actually a confusing letter with a long history of stupidity.  It was frequently interchanged with the old English consonant “eth” which looked like a backwards “p” but when the printing press came along, the top part of the “p” had a splotch of ink which destroyed the entire letter.  To resolve this, printers gradually disconnected the top of the “p”, but then it looked like a backwards “y” which is why we erroneously sometimes say “ye” for “the” as in “ye olde pub”… its actually pronounced “the olde pub”,but that’s another story. Eventually someone (from Germany) had the great idea to just write “th” if they wanted the “th” sound.

This is a re-visitation of one of the more important clinical issues that perhaps people learn in school, but like many things, quickly forget once the exam is over.

A major reason why a client may reject hearing aids or hearing protection is because of the occlusion effect.  It is real, easy to measure, and you can even do it without a probe tube microphone system.

The occlusion effect is an enhancement of low frequency sound energy below 1000 Hz.  It explains why our own voices sound so lousy when we try an occluding hearing aid or stick hearing protection in our ears.  It’s like walking around with a constant cold.

This is a real and measureable effect- it is not something that just belongs in an audiology text book (although it’s interesting how many audiology textbooks don’t have a discussion of this).

Here is how it goes:  Ask your client to say the two vowels, /a/ and in ‘father’ and /i/ as in ‘beat’.  If /i/ is louder then /a/, DO NOT LET THEM LEAVE YOUR CLINIC.  If /a/ is the same as /i/ , then all is good.  I realize that this is not great poetry and that Robert Frost and my grade 5 teacher would be shaking their heads, but nevertheless, it is true.  if someone can come up with better poetry, I’m all ears.

The vowel /i/, being a high vowel, has a very low frequency first formant (first resonant peak).  This is true of all high vowels including /u/ as in ‘boot’, but generally /i/ and /a/ are used because both are roughly the same loudness.  The vowel /u/ is rather quiet to compare directly with /a/.  It is the low frequency first formant that undergoes the greatest enhancement with an occluded ear.  In contrast, mid and low vowels such as the vowel /a/ as in ‘father’ have higher first formants.

Typically the first formant of /i/ (and /u/) are about 125 Hz whereas that of /a/ is around 500 Hz.  For the vowel /a/ there simply is no significant energy below 500 Hz.

Low frequency enhancement when uttering the high vowel /i/ borrowed from https://hearinghealthmatters.org/hearthemusic/2012/deep-earmold-impressions/

Low frequency enhancement when uttering the high vowel /i/ https://hearinghealthmatters.org/hearthemusic/2012/deep-earmold-impressions/  from a guest blog by Patty Johnson

With the occlusion effect, the sound level of the vowel /i/ can be enhanced by over 20 dB when the ear is plugged up.  In contrast, the sound level of the low vowel /a/ only undergoes several decibels of enhancement during occlusion of the ear  with the figure to the left being from a guest blog by Patty Johnson.

Summary:

Unoccluded /i/ and /a/ are the same loudness.

Occluded /i/ is much louder than /a/.

Clinical relevance:

If /i/ is much louder than the /a/ with the hearing aid or the hearing protector in place, then this occlusion effect will most likely cause the wearer to be back in your office in a day or two.

The two main solutions are to make the earmold bore longer such that it sits adjacent to the bony portion of the ear canal thereby preventing the buildup of low frequency sound pressure, or using a vent of sufficient size.  Both of these solutions can be problematical.

Extending the length of the earmold bore down to the bony portion of the ear canal can be uncomfortable with some people.  This part of the ear canal is the only place in the human body where skin sits directly on bone, without any layers of fat.  Many people, including me, find long bore hearing aids or hearing protectors to be rather uncomfortable.

Use of an air vent that allows the excess low frequency sound to exit the ear canal can be a great solution but it does compromise the amount of low frequency amplification from hearing aids and the low frequency attenuation from hearing protectors.

The exact diameter (and length) of the vent can be crucial to balancing the tradeoff between lack of occlusion effect and maintenance of the frequency response of the amplified or attenuated spectrum.  If the bore is long (perhaps “almost” into the boney portion of the ear canal), then only a 1.4 mm vent is required to get rid of the occlusion effect.  If the ear canal bore length is shorter, then a larger air vent will be required.

A happy solution, at least for me clinically, is to make a long bore (but depending on the individual, possibly not into the boney portion of the ear canal) and have some venting.  The largest vent that I can get away with, before compromising the low frequency response and not encountering feedback, is one that will minimize client complaints.

And oh yes – what if we do have a real ear measurement system?

This is the procedure to measure the occlusion effect.

  1. Calibrate the REM device in the normal fashion.
  2. Perform a REUR (unaided real ear canal resonance measure)
  3. Disable the reference microphone and speaker. Depending on the manufacturer, this may be done differently. In the case of Audioscan/Verifit, set the stimulus level to “0 dB”. In the case of the Frye system, set the stimulus level to “off”. In both cases, the reference microphone is disabled and the loudspeaker is turned off.
  4. Have your client utter the vowel /i/ as in ‘beet’ and perform a sweep or run.
  5. Have your client put on the hearing aid or hearing protector, and do the same thing.
  6. The difference in the lower frequencies between step 4 and 5 is a measurement of the occlusion effect.
  7. Clinically increase the vent size until the difference between 4 and 5 has been minimized.

Leave a Reply