A Fast ACELP Codebook Search Method Based on Pulse Replacement and Tree Pruning

The algebraic code excited linear prediction (ACELP) algorithm has been adopted by many speech coding standards, due to low complexity and high quality in its analysis-by-synthesis optimization. One of the optimum ACELP codebook search methods is depth first tree search. Two methods based on pulse replacement and pruning of tree are already proposed for improvement the depth first tree search method. In this paper, the improved pulse replacement and combinational methods are proposed. The combinational method originates from two pulse replacement and pruning of tree methods. These two methods not only reduce the overall complexity but also somewhere have higher efficiency than depth first tree method. The efficiency and complexity of these two methods are examined on GSM-EFR coder in compare with depth first tree, pulse replacement and pruning of tree methods.