Speaker individuality determination by using prosodic features analysis / Shukur Khalid

Speech signal conveys information on several levels. It is an outcome of influence of several physiological, psychological and environmental factors. It contains a message genetically expressed as a sequence of words, information specific to the speaker that produced the speech, and information about the environment in which the speech was produced and transmitted. Speaker specific information includes the identity of the speaker, the gender of the speaker, the language or dialect of the speaker and possibly the physical and emotional condition of the speaker. That richness of information gives a distinctive "identity" to the speech of different individuals. Suprasegmental refers to the time span of the acoustic analysis. Its parameters are spread over several segments as the name suggests. Suprasegmental parameters are also called prosodic parameters. Prosodic is responsible for controlling the intonation, stress, and rhythmic organization of the speech. Typical acoustic suprasegmental measurements are the energy and FO contours. In this research, both prosodic features namely pitch and energy are investigated and analyzed. The purpose is to study how far the normal and disguised speech can be discriminated and to find the best technique for characterizing the individual speakers. Some statistics and prosodic feature extraction computation are conducted through Praat tool and'a comparative analysis between normal and disguised utterances collected from ten different subjects is done. Some variations in pitch and energy features are recorded and from those findings, it is concluded that speaker individuality can be determined whether the subjects speak in normal or disguised manner.