ABSTRACT
Following the first chapter, which introduced basic mathematical notions of signal analysis, we now passed in this chapter to specific techniques for speech processing. This chapter presented a discussion of several speech analysis methods, assuming that our task is to transform a speech signal into a set of parameters that more economically represents the pertinent information in the original speech. Time-domain analysis typically yields simple speech parameters efficiently, which are especially suitable for segmentation purposes, whereas spectral analysis provides the more common approach to an efficient representation of speech information.
