In this paper a new Huffman tree decomposition technique is presented. The basis for this technique involves a set of encoding schemes. These encoding schemes exploit the technique of parallel computation to speed up the encoding and decoding procedures for the Huffman encoding scheme. The proposed encoding schemes also impose a complex structure on the resulting codewords such that the decoding procedures have the potential to become error correcting ones.