| 摘要 |
A data compression apparatus is provided for compressing an input stream of n-bit data samples, with input sample values being notionally divided into groups containing a number of sample values. The apparatus comprises means 12 (fig.4) for generating a code word for at least each non-zero value input sample such that the code words for samples with values in each group have a most significant bit of one type in the same position, which position is different for different groups, and the sequence of bits of lower order than said most significant bit BS, is different in code words for samples with different values in the same group. A run-length counter 18 generates a run-length count dependent upon the number of samples in a run of zero value input samples terminated by a non-zero value sample and a means 17 geneates a position indicator indicating the position of said most significant bit of one type in the code word for the non-zero value sample terminating the run. A code word (e.g. huffman code) is generated in accordance with the output of the run length counter 18 and the position indicator PI, and the bits BS are appended to the code word so as to identify the particular value within each notionally divided group. <IMAGE> |
