SPEECH POWER .-I XI) ENERGY 



Ml 



In.ulod liy ihc peaks. The resulting; suppression of the peaks in the 

 sound output can readily be detected by an accustomed ear, provided 

 that the whole system is reas<inably free from frequency distortion. 



Avi:ra(.k Powkr 



In Tables 11 and III are summarized the obserxations made u|)<in 



the two speeches which were used in this work. There are two reasons 



for showing theni separately: the two speeches were not spoken in im- 



HK'diate succession; and they differ somewhat in character, the first 



being declamatory while the second is of a more conversational nature. 



This difference is not \Qry great, but should account ne\ertheless, for 



the slightK- higher values in Table II. By taking the weighted mean 



of the first number in both tables, we obtain 7.4 microwatts as the 



long average power in normal speech.' 



TABLE II 



First Speech, 50 Syllables 



Average Power in Microwatts 



TABLE III 

 Second Speech, 72 Syllables 

 Average Power in Microwatts 



Note: The average ratio of the total time in the silent gaps to that consumed by 

 the syllables is 0.55; the syllables average 0.16 sec. 



' Crandall and MacKenzie gave an estimate of 12.5; B. S. T. J., \'ol. 1, .\o. 1; Phys. 

 Rev., Mar. 1922. 



