294 



AUDITORY BIOPHYSICS 



similar way other frequencies may be explored, giving an essentially 

 different, though analogous, graphical relation. These data unquestion- 

 ably prove that the normal ear is a reliable loudness meter. 



It has been found by Stevens and Davis [1938] that the data repre- 

 senting the 1000-cycle loudness function quite accurately expresses the 



100,000 

 50,000 



10,000 

 5,000 



o 



£ 



a. 

 c 



g 1,000 

 a 



c 



1 500 



-J 



100 



50 



20 30 40 50 60 70 80 90 

 Intensity level in decibels above threshold 



Fig. VII-20. The loudness function. This shows how the experienced loudness 

 varies with the intensity of the stimulus at the frequencies indicated. These stand- 

 ard curves were adopted from data by Fletcher and Munson [1933]. (By courtesy 

 of the American Standards Association, New York.) 



data for the frequency range 700 to 4000 cycles. In general, at lower 

 frequencies loudness increases faster than in the higher range of audible 

 frequencies for the same change in intensity. Thus an increase in 

 intensity of 20 db (50 to 70 db), i.e., a tenfold increase in pressure ampli- 

 tude of the sound stimulus of a 100-cycle tone (Fig. VII-20), will increase 

 the loudness fifty fold, but for the 1000-cycle tone this change produces 

 only a fourfold increase in loudness. In general, in the 100- to 4000- 

 cycle frequency band (vowel sounds 300 to 2500 cycles) at an intensity 

 level of about 60 db, the high-frequency tones are heard as much as ten 

 times louder than the low-frequency tones. 



