Browse Prior Art Database

Variable Length Code Word Encoder Decoder

IP.com Disclosure Number: IPCOM000053001D
Original Publication Date: 1981-Aug-01
Included in the Prior Art Database: 2005-Feb-12
Document File: 2 page(s) / 39K

Publishing Venue

IBM

Related People

Jones, I: AUTHOR

Abstract

A variable-length code-word encoder/decoder is described which allows a fixed algorithm, implemented in hardware or software, to generate any desired code length structure using stored coefficient tables. The translation process is divided into two steps to reduce table size. The fixed length Source character is first translated by table look-up to Rank; then translations between Rank and Code can be performed by means of a simple algorithm, which uses a table specifying the number of codes of each length.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 62% of the total text.

Page 1 of 2

Variable Length Code Word Encoder Decoder

A variable-length code-word encoder/decoder is described which allows a fixed algorithm, implemented in hardware or software, to generate any desired code length structure using stored coefficient tables. The translation process is divided into two steps to reduce table size. The fixed length Source character is first translated by table look-up to Rank; then translations between Rank and Code can be performed by means of a simple algorithm, which uses a table specifying the number of codes of each length.

Huffman's well known algorithm devises a tree which, given the set of probabilities of all Source characters, gives the length of code word for each character and for which the average code length is minimized. The actual assignment of codes is arbitrary subject to the requirement that no code forms the prefix of any other code word.

Table 1 shows examples of coding. Eight items are listed in order of decreasing Frequency. Column H shows codes assigned strictly from the Huffman tree, column J shows codes, of the same length as in column H, assigned to increase in numerical value as the length increases (viewed as numbers with the binary point in Front of the code). See original. Table 1 The structure of code J in Table 1 is summarized in Table 2. See original. Table 2

The algorithm for deriving the code J from the Rank R and vice versa utilizes only the list of numbers N(L) of codes of each length L. J(R) is defined by...