Comparison of Khasi Speech Representations with Different Spectral Features and Hidden Markov States

In this paper,we present a comparison of Khasi speech representations with four different spectral features and novel extension towards the development of Khasi speech corpora.These four features include linear predictive coding(LPC),linear prediction cepstrum coefficient(LPCC),perceptual linear prediction(PLP),and Mel frequency cepstral coefficient(MFCC).The 10-hour speech data were used for training and 3-hour data for testing.For each spectral feature,different hidden Markov model(HMM)based recognizers with variations in HMM states and different Gaussian mixture models(GMMs)were built.The performance was evaluated by using the word error rate(WER).The experimental results show that MFCC provides a better representation for Khasi speech compared with the other three spectral features.

A unified direct approach for discretizing fractional-order differentiator in delta-domain

Fractional-order differentiator is a principal component of the fractional-order controller.Discretization of fractional-order differentiator is essential to implement the fractionalorder controller digitally.Discretization methods generally include indirect approach and direct approach to find the discrete-time approximation of fractional-order differentiator in the Z-domain as evident from the existing literature.In this paper,a direct approach is proposed for discretization of fractional-order differentiator in delta-domain instead of the conventional Z-domain as the delta operator unifies both analog system and digital system together at a high sampling frequency.The discretization of fractional-order differentiator is accomplished in two stages.In the first stage,the generating function is framed by reformulating delta operator using trapezoidal rule or Tustin approximation and in the next stage,the fractional-order differentiator has been approximated by expanding the generating function using continued fraction expansion method.The proposed method has been compared with two well-known direct discretization methods taken from the existing literature.Two examples are presented in this context to show the efficacy of the proposed discretization method using simulation results obtained from MATLAB.