The use of waveform asymmetry to identify voiced sounds

A simple speech measurement is described that allows the identification of voiced and unvoiced sounds. In addition the measurement can be used to distinguish between a limited number of different vowel sounds. The measurement consists of examining the difference in magnitude between the positive and negative peaks of the speech waveform. This asymmetry measurement can become completely volume insensitive if preceded by a frequency-dependent phase shifter. Optimization of the measurement for individual speakers is quite easily accomplished.