FK;. 1. Illustrating one method which has been used in the past to show how vocal resonances change with time. This is a series of harmonic analyses of successive periods of the vowel in the word "out." 19

By treating the vocal tract as a series of cylindrical sections, or acoustic lines, it is possible to use transmission line theory in finding the resonances. With constants uniformly distributed along each section, resonances appear as modes of vibration of the tract taken as a whole. Thus, the fundamental mode of the smaller cavity may be affected considerably by a higher mode of the larger; and in addition, higher resonances are found without postulating additional cavities. This is an advantage over the lumped constant treatment, where it is necessary to postulate a different cavity for each resonance, and where the interaction terms in the equation do not include the higher modes of vibration. Under the distributed treatment, dimensions for each vowel may be taken from x-ray photographs of the vocal tract. The calculations then yield at least three resonances which lie in the frequency regions known for the vowel, from analyses of normal speech. Dependence of the different resonances upon the different cavities is discussed in some detail in the paper. An electrical circuit based on the transmission line analogy has been made to produce acceptable vowel sounds. This circuit is useful in confirming the general theory and in research on the phonetic effects of articulator movements. The possibility of using such a circuit as a phonetic standard for vowel sounds is discussed.

The damping constants of vowel formants, as expressed in the half-power bandwidths, have been studied from the reports of several investigators. The accuracies of methods of measurement used have been estimated through application of the same methods to an electrical analog of the vocal tract. Two new sets of measurements are presented, one on spectrograms of real speech, the other making use of an artificial larynx, applied to real subjects. It is concluded that, while wide individual variations occur, the average bandwidths for male voices probably lie in the lower range of those which have been reported.

A single speaker in a seated position repeated a fifteen-second sample of connected speech, while r.m.s. pressure measurements were made in thirteen frequency bands, and at seventy-six positions, in different directions and distances. The results are applicable to intelligibility and microphone placement problems. They show, in general, the greater variation with direction at higher frequencies. Directivity due to the size of the mouth opening appeared to enter above 5600 cycles per second, the axis at these frequencies being about 45° below the horizontal, in front. Frequencies below 1000 cycles per second were found strongest directly downward from the lips, or nearly so. The power radiated in different directions has been calculated, and a summation gives a spectrum of the total speech power emitted by the mouth. It is proposed that similar spectra for other speakers may be obtained from pressure measurements at a single point, using the relations discovered for this speaker. The necessity for protecting a microphone used close to the mouth, from the puffs of air accompanying the speech, is demonstrated and explained.

A new experimental artificial larynx, which makes use of transistors and miniaturized components to provide a voice for those who have lost the use of their vocal cords by surgical removal or paralysis, is described. The larynx operates by introducing a substitute for the sound of the vocal cords into the pharyngeal cavity by means of a vibrating driver held against the throat. The acoustic principles of normal and artificial speech production that were followed in arriving at the new design are presented, along with descriptions of the transistor circuit and its operating characteristics.