| 摘要 |
<p>PURPOSE:To reduce the size of a lookup table(LT) by converting a string of a high-order bit of an input code into a more dense identification code string in a 1st lookup table for space compression, combining the identification code and a low-order bit of the input code in a 2nd LT to encode the combined code. CONSTITUTION:A Huffman code string whose maximum code length is 16 bits is divided into m-bit and (160m) bits to generate a 1st LT3 and a 2nd LT4 when an identification code has 8 bits. In this case, a 1st Huffman code and a Huffman code length (size) are loaded by setting an ID set at first to be 1. When the size is m-bit or over, n=0 is set and when the size is less than m-bit, n=12-(size) is set. Then the value (n) is used to write the ID for 2<n> times to consecutive 2<n> addresses in the memory. The Huffman code and its size are loaded, and when the size is 0, the processing is terminated and when the size is other than 0 and the high-order m-bit of the Huffman code are already set, the relation of ID=ID+1 is set, and the succeeding Huffman code and size are loaded and the similar processing is repeated.</p> |
