On improving Tunstall codes

Though many compression methods are based on the use of variable length codes, there has recently been a trend to search for alternatives in which the lengths of the codewords are more restricted, which can have useful applications, such as easier processing and fast decoding. This paper explores the construction of variable-to-fixed length codes, which have been suggested long ago by Tunstall. Using new heuristics based on suffix trees, the performance of Tunstall codes can in some cases be improved by more than 40%.

Research highlights
► The construction of variable-to-fixed length codes based on suffix trees.
► The dictionary is based on the substrings that actually appear in the text.
► The obtained compression performance is better than that of Tustall codes.