The Spectral Characteristics Research of the Voice-Speech Signal in Dysphonia

The object of this research is voice samples (au-dio recordings) of a person with hypotonic and hypertonic dysphonia from healthy voices. The relevance and significance of the research are ensured by the rapid growth the number of people suffering from voice disorders. The work's aim is the developing of a method for analyzing the voice through digital signal processing, allowing to differentiate objectively a voice with hypotonic dysphonia, a voice with hypertonic dysphonia and a voice in normal, taking into account its characteristic changes and violations of the phonation process. The extended acoustic analysis of children's voices recordings, obtained as a result of the experiment, was carried out in the process of work. Within the work, a method for analyzing the voice for the presence of the phonation process violations was developed, taking into account the values of noise characteristics, the type of spectral and cepstral characteristics and the parameter Cepstral Dysphonia Coefficient (CCD), which evaluates the certain pathology's presence.