A VOICE AXD EAR 



303 



of the mouth opening. To determine the effect of increasing the 

 distance between the transmitter and the mouth, measurements were 

 made with the deskstand t>pe of instrument with results as shown in 

 Fig. 8. 



Measurements were made of the distortion of the sound field from 

 each of the four types of mouth mentioned which was caused by the 

 introduction of different types of transmitter in the field. The 

 effects were measured immediately to one side of the mouthpiece. 

 It will be seen from Fig. 9 that for frequencies below 2000 cycles there 

 is good agreement for all types of mouth and both types of instrument. 

 Above 2000 c.p.s. variations appear, but in general the largest differ- 

 ences occur as before for the extreme form of artificial mouth, namely, 

 the pipe without termination. 



500 1000 



FREQUENCY IN CYCLES PER SECOND 



10,000 



Fig. 8 — Effect on station transmitters as distance from mouth is varied. 



Summarizing the results obtained with the artificial mouth de- 

 scribed, although the general requirements have not been completely 

 satisfied, a rather close approximation has been realized. The re- 

 sponse-frequency characteristic does not vary more than ±1.0 db 

 from 250 to 3000 c.p.s. nor more than ± 2.0 db from 100 to 7500 c.p.s. 

 The mouth is capable of delivering an output of an intensity equivalent 

 to loud talking without appreciable non-linear distortion. The dis- 

 tribution of the sound field is similar to that of the real mouth. As 

 judged by tests on several different forms of mouth of quite different 

 sizes and shapes, the indications are that the introduction of objects 

 in the sound field of the artificial mouth chosen distorts that field in 

 about the same manner as occurs with the real mouth. Speech 

 reproduced by the artificial voice sounds natural. Comparative 



