Abstract:
The introduction of CELP (Codebook Excited Linear Prediction) speech coding provided an efficient way to compress speech data to 4.8 kbps with high quality, but with the disadvantage of great computational complexity required for real-time processing. In this paper, we investigate the overall computing complexity of the CELP speech coders and we propose some particular structure of codebook involved in residual speech coding, in order to obtain a significant reduction of the complexity.