Speech Testing with Audiometeric Earphones
by Larry Revit, hearing scientist
New data suggests a possible need to revise standard practice.
Here's a clinical tidbit which will come as no surprise to some, but which may shock others. Because Frye Electronics has just introduced the FONIX FA-10 Hearing Evaluator, I have been assigned the task of devising instructions for using the "Hearing Aid Simulator" feature. Operating this feature requires presenting speech through the audiometer, so, being the "hearing scientist", I soon began asking questions -- like what, acoustically, is being presented to the client during speech audiometery, and how does it relate to hearing aids?
Because almost all the talk about hearing aids nowadays relates to insertion gain, the above questions boiled down to "What is the equivalent insertion gain, the above questions boiled down to "What is the equivalent insertion gain of a speech audiometer?" I found some data in JASA1 that let me calculate the sound pressure level (SPL) at the average eardrum, in response to a constant SPL into a speech audiometer having a flat frequency response and using a TDH-39 earphone. This is no more and no less than the "real-ear aided response" (REAR) of a speech audiometer: the audiometer, in this case serving as the hearing aid. Because "real-ear insertion-gain response" (REIR) is defined as the difference between the aided and unaided real-ear responses, I had only to find an appropriate real-ear unaided response (REUR) to subtract from the already known aided response. For this I used the mean, diffuse-field data from Killion, Berger, and Nuss2. The results are shown in the accompanying figure. The bold curve shows that the high frequencies are attenuated relative to the lower frequencies.
What is the practical implication of these results? The implication is that audiometeric speech tests are given under acoustic conditions that attenuate the important speech frequencies above 1000 Hz by as much as 10 dB (in the average ear). So we are giving our clients an additional 10 dB of high-frequency hearing loss before we determine their abilities to hear speech!
Recommendations: The hearing health care profesion might consider re-examining the validity of this practice, and could consider the possible alternatives of:
- a. Electronically compensating speech audiometers for the "insertion loss" of audiometric earphones, or else
- b. using earphones that reproduce the natural earcanal resonance.
In the meantime, for those interested in experimenting, using the "6dB" or "12dB" slope settings of the Hearing Aid Simulator in the mid to high frequencies than using the "Off" setting. Let me know what you come up with!
Footnotes
- Cox, RM (1986). NBS-9A coupler-to-eardrum transformations: TDH-39 and TDH-49 earphones. Journal of the Acoustical Society of America, 79(1):120-3.
- Killion, M.C., Berger, E.H., and Nuss, R.A. (1987). Diffuse field response of the ear. Journal of the Acoustical Society of America, 81(1) Suppl: S75