180 ANNUAL EEPOET SMITHSONIAN INSTITUTION, 19 3 2 



The Western Electric audiometer offers a choice of eight standard 

 pure notes from 64 to 8,192 cycles per second generated by a valve 

 oscillator. (An electric buzzer is used in an earlier model first 

 adapted in 1925 from an audiometer primarily developed for the 

 measurement of hearing.) The instrument is graduated in steps of 

 5 decibels. The telephone receiver is provided with an off-set ear 

 plate, so that the noise to be measured enters by the same ear as is 

 applied to the telephone receiver. 



In a third type of audiometer, the valve oscillator is replaced by a 

 gramophone record and electrical pick-up. This also will furnish at 

 will a selection of from three to six different pure notes of high, 

 medium, and low frequency, or, if preferred, of " warbling " notes, 

 the frequency of a complete warble being about six times a second. 

 In practice, a 3-band warbling record may have the following fre- 

 quency limits: 1,500-5,600, 750-1,500, and 250-750 cycles per 

 second. With this audiometer, also, the same ear is normally used 

 for both the standard note and the noise to be measured. 



With all the various types of audiometer, the reading is best ap- 

 proached by a series of progressive comparisons alternating on either 

 side of the critical value. When making measurements one should, 

 as it were, try to focus on loudness, and endeavor to ignore such fre- 

 quency differences as there may be. Unless one ear is abnormal, it 

 does not seem to matter a great deal for ordinary requirements 

 whether the standard note and the noise enter by the same ear or 

 by different ones. 



If a noise contains only two or three major components, one ob- 

 server may be tempted unconsciously to match on one component, 

 whereas another observer will seize on another component, with con- 

 sequent disagreement in their results. In such cases there may be an 

 advantage in having a choice of frequency for the standard note. 



In general, however, experience indicates that most people find 

 themselves able after a little practice to obtain fairly consistent re- 

 sults with even the simpler forms of audiometer, at any rate with 

 noises which are more or less continuous. On the whole, masking 

 results are easier to obtain than matching or equality values. In 

 the United States, masking results are usually preferred, affording 

 as they do a measure of the degree of " deafening " or the raising 

 of the threshold limit for the particular frequency or band of 

 frequencies used. 



CLICKER EXPERIMENTS ON NOISE MEASUREMENT 



The writer was led in 1929 to make some rough experiments on 

 noise measurement, by the aid of a flexed steel-strip " clicker," such 

 as is sometimes used by lecturers. The note of the clicker is high 



