590 



BELL SYSTEM TECHNICAL JOURNAL 



of senii-vowol sounds (r, /, w;, n, ng) closely reUited to the \()\vel 

 group, and \ielding readih' a determination of tiicir "rharacteristir 

 frequciicies." 



There are two physical theories of vowel production; and these 

 two theories suggest different methods of analyzing the vowel sounds 

 into components of simpler nature. These two points of view we 



1000 2000 J<-00 4000 



Frequency 



Fig. I — Energy eiislribution; composite riir\c for male and female voices 



shall briefly consider along historical lines. We are indebted to 

 Helmholtz for the greatest single contribution to the study of the 

 vowels, in that he gave a complete diagram of the characteristic fre- 

 quencies of the vowels (ref. 2, pp. 103-109), which w-as based on his 

 celebrated experiments in analysis and synthesis by means of the 

 Helmholtz resonators. But in connection with his scheme of char- 

 acteristic frequencies he took up the theory of Wheatstone (1837) that 

 these fre(iuencies are true harmonic components of the cord tones, 

 which were rcenforced by resonance in the oral cavities. Some later 

 jihysicists have followed this so-called harmonic or steady state theory 

 of the \()wel sounds, notably Miller (reference 3, pp. 239-243) who 



