Properties of low augmentation level T-codes

Initial investigations of the re-synchronisation properties of fifth augmentation level, variable length, self-synchronising T-codes are described. It has been found from simulation studies that within the fifth augmentation level family there are considerable differences in the resynchronisation delay statistics of individual code sets, with some code sets consistently outperforming others with the same codeword length distribution. Furthermore, the code sets with the shortest average re-synchronisation delays are generally those with the shortest average code-word length (i.e. the most efficient). The coding efficiency of these code sets has been found to be within 1% of the efficiency of a Huffman code optimised for the particular information source.