A CELP codebook and search technique using a Hopfield net

A new method of structuring a ternary excitation codebook for a CELP coder is introduced in which the main codebook is divided into several sub-codebooks. It is shown how a modified Hopfield neural net is capable of searching such a codebook and selecting a near optimum code-vector. Hence an algorithm is derived which offers considerable computational savings over previously proposed codebook search methods. Results are presented demonstrating the net's performance and its effect on the subjective quality of the coded speech.<>