Perception of Speech by Deafened Listeners and Its Relation to the Design of Hearing Aids
Harvey Fletcher · The Journal of the Acoustical Society of America · 1952
This paper describes a method of calculating the articulation scores which a talker-listener pair will obtain in terms of the response of the hearing aid, the audiogram (both bone and air) of the listener, and talking level of the talker, and the proficiency factor of the talker-listener pair. This method was used to calculate the articulation scores for all the talker-listener pairs and the six kinds of hearing aids used in the test at Harvard made by Davis and his associates. The calculated and observed results agree within the experimental error. The philosophy underlying the method of calculation is used to predict the ideal type of response to be used by a deafened listener. It leads to the following simple rule. The ideal response R in db gain at any frequency is given by R = βA − 23βB + αI, where βA is the hearing loss by air conduction, βB is hearing loss by bone conduction, and αI is the over-all gain (same for all frequencies) to bring the speech level at the listener's ear to its optimum value. If the values of βA − 23βB at the three frequencies 500, 1000, and 2000 chs are examined and the average of the two highest values called β2, then the maximum value of αI is given by αI(max) = 52 − β2. The optimum gain αI is, in general, less than αI(max), and so a means of lowering the gain in the set must be provided. When αI(max) calculates to be negative, then the optimum gain αI is equal to αI(max). These gains are all with reference to a speaker 100 cm from the listener, and speaking so that his average level at the listener's ear is 66 db. If the speaker moves farther away or talks softer, then it is counted a negative gain or a loss. If he moves closer or talks louder it is counted as a gain.