Today, there exist a wide variety of wire and wireless communication networks in which different speech coding standards are adopted. These networks are in the process of integrating into one IP-based next generation network(NGN), and in the integration process the compatibility between different coding standards is becoming a major issue. To address this problem, a transcoding algorithm, which is to convert the encoded bit stream of one speech coder into that of the other, is required.
In this thesis, a transcoding algorithm for the Selectable Mode Vocoder (SMV) and G.723.1 speech coders via direct parameter transformation is proposed. The tandem transcoding is a simple solution to the compatibility problem between different coding standards; however, it is associated with a number of problems such as degradation in speech quality and the increases in computational complexity and delay. These problems can be alleviated by the proposed transcoding algorithm which converts the parameters of one coder to those of the other without going through the complete decoding and encoding process. The proposed algorithm is composed of five parts: parameter decoding, line spectral pair (LSP) conversion, simplified frame classification, pitch conversion for adaptive codebook (ACB) and fixed codebook (FCB) conversion. The proposed algorithm is implemented in the fixed-point format using the C language in the Windows platform. The evaluation results show that the proposed algorithm achieves equivalent speech quality to that of tandem transcoding with reduced computational complexity and delay.