Some advances on speech analysis using generalized dimensions

Nonlinear systems based on chaos theory can model various aspects of the nonlinear dynamic phenomena occuring during speech production. In this paper, we explore modern methods and algorithms from chaotic systems theory for modeling speech signals in a multidimensional phase space and extracting characteristic invariant measures such as the generalized fractal dimensions. Such measures can capture valuable information for the characterisation of the multidimensional phase space since they are sensitive on the frequency that the attractor visits different regions. Further, we integrate some of these chaotic-type features with the standard linear ones (based on cepstrum) to develop a generalized hybrid set of shorttime acoustic features for speech signals and demonstrate its efficacy by showing slight improvements in HMMbased phoneme recognition without the use of any language model.