The processes of spectral analysis of frames of signals using spectral models variables in time and signal coding

摘要

A time variable spectral analysis for speech coding based upon interpolation between speech frames. A speech signal is modeled by a linear filter which is obtained by a time variable linear predictive coding analysis algorithm. Interpolation between adjacent speech frames is used in order to express a time variation of the speech signal. In addition, interpolation between adjacent frames secures a continuous track of filter parameters across different speech frames.