Slaying the Myths in Variable Length Coding
GR Higgie · 1992
Variable length coding has long been seen as a useful technique for data compression in communications, but it has suffered from the generally held belief that the encoding and decoding process is difficult because of the variable length nature of the codewords, and that the difficulty of codeword synchronisation after a data corruption precludes the use of variable length codes when such corruption is probable. This paper outlines the results of recent research into the self-synchronising variable length T-Codes. One result of this research has been the development of an extremely simple but fast technique for encoding and decoding variable length codes. A second result has shown that for any particularly application it is possible to select a T-Code which offers as much code compression as the ideal Huffman code, but which typically has an average codeword synchronisation delay of only 1.5 codewords. Taken together, these two results refute the above generally held beliefs which have previously restricted the use of variable length codes.