A low delay speech coding system at 4.8 kb/s

Abstract

Low-delay speech coding has drawn much attention because of its many potential applications. A lot of low delay speech coders have been proposed in the past few years. But almost all of them focus on the speech bit rate between 8 kb/s to 16 kb/s. The authors present a low delay speech coder operating at 4.8 kb/s which provides good speech quality with a coding delay only about 2.5 msec. This new low delay coder uses backward LPC analysis to remove the short-term correlation from speech and adopts a new approach to predict the long term prediction (LTP) coefficients based on a "dual-decision" scheme. At the excitation part, the authors exploit a new excitaton model which contains a sinusoid signal codebook and a stochastic codebook to excite the whole speech system. The synthesized speech of this coder has comparable quality to that of FS-1016 CELP coder while having a delay constraint more than an order of magnitude smaller than that of the FS-1016 coder.<>