Abstract

This paper presents a fast efficient codeword search algorithm for vector quantization (VQ) based on the Adaptive Multi-Rate Wideband (AMR-WB) speech codec. AMR-WB is the speech codec standardized by the 3GPP for the GSM and WCDMA 3G system. It is also selected as the ITU-T G.722.2 standards. In AMR-WB speech codec VQ codeword search section, full search algorithm is used to find matching codeword .This requires a huge amount of computation. For large-size codebooks and high-dimensional vector of AMR-WB, computational complexity is significantly higher.The proposed fast codeword search method selects the codeword ,which has the least absolute error component in VQ codebook, as the initial matching codeword. It can help to reduce computational complexity, whereas the reconstructed speech quality almost has no dropping. The article introduces the basic principle of the fast codeword search algorithm in detail. Simulation results show that the algorithm based on AMR-WB speech codec, can reduce the time of vector quantization codebook search and the w-PESQ of nine modes is no significant dropping.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call