TSVQ base compression for voice data

This paper reports an efficient compression technology for voice data. The technology is targeted for the mobile phone and voice over IP (VOIP) area where the main data is human speech. General compression method is used in such a applications where the main data is human voice. In this paper we used tree structure vector quantization (TSVQ) based clustering method on voice data to extract voice features and compressed the data on this basis. It is developed around Tree Structure Vector Quantization (TSVQ) algorithm. It ensures high speed of encoding and decoding process with acceptable quality compare to existing standard GSM (Global System Mobile) technology. The quality of decompressed voice data confirms high potential in the network society.