LOUDNESS 



427 



data is concerned with differences in the threshold when Hstenin^ with 

 one ear versus hstening with two ears. 



It is well known that for any individual the two ears have different 

 acuity. Consequently, when listening with both ears the threshold is 

 determined principally by the better ear. The curve in Fig. 20 shows 

 the difference in the threshold level between the average of the better 

 of an observer's ears and the average of all the ears. The circles 

 represent data taken on the observers used in our loudness tests while 

 the crosses represent data taken from an analysis of 80 audiograms of 

 persons with normal hearing. If the difference in acuity when listening 

 with one ear vs. listening with two ears is determined entirely by the 

 better ear, then the curve shown gives this difference. However, 

 some experimental tests which we made on one ear acuity vs. two ear 

 acuity showed the latter to be slightly greater than for the better ear 

 alone, but the small magnitudes involved and the difiliculty of avoiding 



200 500 1000 2000 



FREQUENCY IN CYCLES PER SECOND 



10000 20000 



Fig. 20 — Difference in acuity between the best ear and the average of both ears. 



psychological effects caused a probable error of the same order of 

 magnitude as the quality being measured. At the higher frequencies 

 where large differences are usually present the acuity is determined 

 entirely by the better ear. 



From values of the loudness function G, one can readily calculate 

 what the difference in acuity when using one vs. two ears should be. 

 Such a calculation indicates that when the two ears have the same 

 acuity, then when listening with both ears the threshold values are 

 about 2 db lower than when listening with one ear. This small 

 difference would account for the difficulty in trying to measure it. 



We are now in a position to compare the data of Kingsbury with 

 those shown in Table I. The data in Table I can be converted into 

 decibels above threshold by subtracting the average threshold value in 

 each column from any other number in the same column. 



If now we add to the values for the level above threshold given by 



