摘要 |
A speech coding system robust to frame erasure (or packet loss) is described. Illustrative embodiments are directed to a modified version of CCITT standard G.728. In the event of frame erasure, vectors of an excitation signal a re synthesized based on previously stored excitation signal vectors generated durin g non-erased frames. This synthesis differs for voiced and non-voiced speech. Duri ng erased frames, linear prediction filter coefficients are synthesized as a weight ed extrapolation of a set of linear prediction filter coefficients determined durin g non-erased frames. The weighting factor is a number less than 1. This weighting accomplishes a bandwidth-expansion of peaks in the frequency response of a linea r predictive filter. Computational complexity during erased frames is reduced thro ugh the elimination of certain computations needed during non-erased frames only. Th is reduction in computational complexity offsets additional computation required fo r excitation signal synthesis and linear prediction filter coefficient generation during erased frames.
|