The main goal of this work is to develop an alternative methodology for acoustic-phonetic modelling of speech sounds. The approach utilizes a segment-based framework to capture the dynamical behavior and statistical dependencies of the acoustic attribute
