Abstract
A new Huffman coding and decoding technique is presented. There is no need to construct a full size Huffman table in this technique; instead, the symbols are encoded directly from the table of code-lengths. For decoding purposes a new Condense Huffman Table (CHT) is also introduced. It is shown that by employing this technique both encoding and decoding operations become significantly faster, and the memory consumption also becomes much smaller compare to the normal Huffman coding/decoding.