Acoustic-statistical analysis of voice quality

Abstract

The investigation deals with long-time analyses of: (1) fundamental frequency, (2) intensity above 1000 Hz relative to the total intensity, and (3) total intensity. Data are sampled 20 times per second. For each parameter, the mean and the coefficient of variance are calculated. The method seems useful for comparisons of changes in prosody taking place e.g. during voice training. Especially the coefficient of variance seems to be a good measure for the dispersion of pitch, "voice quality" , and intensity.