Abstract

During the analysis of speech signals the evaluation of the basic characteristics of the speech is an important
stage. The basic characteristics of speech are voiced, unvoiced and silence. Such characteristics are
evaluated by calculating zero crossing rate (ZCR) and short term energy (STE). The speech segmentation
system needs a speech signal to be segmented into some basic units like words, phonemes, syllables. In this
paper some of the
steps required for the feature extraction in case of automatic segmentation of speech are discussed. There are
various characteristics of speech such as: voiced, unvoiced and silence. Such basic characteristics of the
speech can be evaluated by the computation of zero crossing rate and short term energy