182 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 31 



'i 



A particular fork used by Davis had a frequency of 640 cycles per ' 



second. Its loudness when it was struck was about 90 decibels and ' 



the rate of decay was about fi/^ decibels per second. Noises as high I 



as 110 decibels were measured by means of masking observations, j 



Davis has used the method for determining the loudness of a vari- i 



ety of noises over the range of hearing, and obtained results which, as | 



will be seen from Table 1, are in good agreement with audiometer i 



measurements by American observers. I 



RELATION BETWEEN THE LOUDNESS VALUE AND THE MASKING | 



VALUE OF NOISES i 



As the masking effect of a noise is dependent on its composition, | 

 theoretically it is in general not possible to associate the loudness i 

 value of a noise (as determined either physically or by aural match- j 

 ing) and the masking value as determined aurally by one or the other 

 of the available audiometers. As a practical fact, however, it would i 

 appear that for most of the ordinary complex and fairly continuous 

 noises of everyday life, the loudness value exceeds the masking by a ; 

 fairly constant difference, which tends to increase somewhat for ; 

 louder noises, or for those of an intermittent staccato character. 

 (Williams and McCurdy, Journ. Amer. Inst. Electr. Eng., Septem- 

 ber, 1930, and Gait Journ. Acoust. Soc. Amer., July, 1930.) j 



For normal street and interior noises the New York commission j 

 found that on the average the loudness value exceeded the medium- I 

 frequency masking value by about 15 decibels. For a very loud noise j 

 (90 decibels), such as the intense cheering of a large crowd, it would ' 

 appear from the report that the two values differed by 20 decibels. | 

 Davis (loc. cit.), in his tuning-fork experiments, found a like dif- \ 

 ference for a loudness of 110 decibels. He also refers to an approxi- 

 mately linear relation between masking and matching values for | 

 moderately loud noises. f 



In the case of measurements of airplane-propeller noises of very I, 



high intensity made for the Aeronautical Research Noise Subcommit- j 



tee by the National Physical Laboratory the two values differed ' 



by about 20 to 30 decibels, though it will be realized that measure- i 



ments at such intensities are necessarily somewhat rough. \ 



EXAMPLES OF NOISE MEASUREMENT 



Some simple illustrations of the measurement of everyday noises 

 may be of service. 



The loudness of speech ranges between about 40 and 60 decibels, 

 an ordinary conversational tone being about 50 decibels. If, how- 

 ever, the lijDS of an average speaker are within one-half inch of the 

 ear of a person with normal hearing, the latter will receive the speech 

 at a level of about 100 decibels. An average motor horn sounded 



