| 摘要 |
A method comprising entropy encoding into bits a set of data values, and packing into storage the entropy encoded bits by reversing the bits of words with unknown length and keeping in blocks the words with known lengths. For instance, in an entropy encoded data set that uses both Huffman coding and zero run coding, the class code may be reversed in bit order from right to left rather left to right while the words of known length such as the zero run code and Huffman pointer are stored left to right in blocks. This data arrangement is particularly useful in an MMX based machine.
|
