Efficient Modeling of Dominant Transform Components Representing Time-Varying Signals

An accurate representation of time-varying signals using mixed transforms has been reported in [1], The mixed transform representation uses partial sets of basis functions from the discrete Fourier transform (DFT) and the Walsh Hadamard Transform (WHT). The location, magnitude and phase of the transform components have to be specified for proper signal reconstruction. A least squares (LS) IIR algorithm, in the transformed domains, which fits each of the retained subsets of the complex transform components accurately, is presented in this letter. The IIR function, while characterized by real coefficients about twice the number of the retained complex transform components, carries enough location, magnitude and phase information for accurate signal reconstruction. To illustrate the technique's accuracy and efficiency, its application to model the DFT part of a voiced speech segment is given.