A VOICE AND EAR 



301 



brief and easy to say. The speech output of these same individuals 

 calhng the above phrases was recorded by the process mentioned.^ 

 After adjusting the output of the artificial voice for the close talking 

 position with a condenser transmitter,^ tests were made on the various 

 instruments interspersing both instruments and voices to reduce 

 testing errors and minimize changes in instrument characteristics 

 between tests. The results of individual tests on a given type of 

 instrument with both the human and artificial voices varied over 

 approximately the same range. The average results for all tests for 

 each type of instrument for each distance and for both the human and 



I 2 3 



DISTANCE IN INCHES 



Fig. 6 — Distance loss characteristics. 



artificial voices are shown in Fig. 6. These data show reasonably 

 close agreement between the artificial voice and human voice, both as 

 regards the general level and slope of the distance loss characteristics. 

 In the practical application of the artificial mouth to telephone instru- 

 ment studies the minor discrepancies shown may be minimized by the 

 use of correction factors. A further investigation is being made of 

 the sound field distribution of the artificial mouth in comparison with 

 that of human mouths. 



Tests were made to determine the effect on telephone instruments 

 of the size and shape of the artificial mouth. Response-frequency 

 measurements were made, a deskstand and a handset transmitter 

 being used, each modified to include in the plane of the diaphragm a 

 small condenser transmitter. The latter was used in order that any 

 variation of carbon instruments might be eliminated. Sound pressure 

 was obtained from each of four types of artificial mouth. Three of 

 these employed a long pipe with an inside diameter of about 0.7 inch. 



1 Loc. cit. 



^ "Electrostatic Transmitter," E. C. W'ente, Pliys. Rev., May, 1922. 



