Determination of the area function of individual vocal tract from average for sustained vowels

Area functions of vocal tracts, for a certain sustained vowel, have considerable individual variations though an average shape is observed associated with each vowel sound. In this paper, two algorithms are presented for the determination of the functions with the objective to detect the malfunctioning muscles/nerves controlling the shape of the vocal tract. In the first method an iterative technique is used; the average area function associated with the vowel is the starting point and this area function converges to individual. In the second method the glottal wave is obtained along with the speech signal and the former is used to separate the transfer function of the vocal tract (V(z)). A new simple low-cost technique, using the stethoscope and a microphone, is proposed to obtain the glottal wave.