Discrete Systems and Digital Signal Processing with MATLAB, Second Edition free download
this matlab toolbox supports the definition of a class of filter structures as state-space models and uses several established techniques such as sliding window analysis (swa) and power spectral density (psd) to evaluate its performance, compare its design parameters to the performance of other, compared to other filters. the performance results are then used to generate analysis code and test results. the design parameters can be adjusted by the user and these results can be displayed on a color computer screen in an interactive way. the analysis results in the form of a table and a graph can also be generated directly from the table. the matlab filter design module also automatically generates vhdl code, test benches, and simulation results.
in the matlab filter design module, the parametric filter design optimization algorithms are available for dc, the two-sided bandwidth, the order and the two-sided frequency performance. the toolbox also includes the use of least squares methods for estimating parameters such as the frequency, order and center frequency. autocorrelation results from a sliding window can be used for optimizing the order and center frequency. the mfcc (mel-frequency cepstral coefficients) classifier uses mfcc parameters as input features for classifying speech, words and phonemes. the first step is the mfcc feature extraction based on the mel-frequency spectral scaling of the speech signal. after the mfcc feature extraction, the mfcc will be used as a classifier that analyzes signals to determine which class a particular speech sample belongs. the mfcc classifier has been used successfully in speech recognition for a number of widely diverse automatic speech recognition systems.