초록 ▼

Novel decoding method and apparatus are disclosed for Huffman or variable-length codes. According to the invention, variable-length code-words are divided into two parts, a variable-length prefix part and the remainder of the code-word which is of fixed-length or approximately so. The variable-lengt

Novel decoding method and apparatus are disclosed for Huffman or variable-length codes. According to the invention, variable-length code-words are divided into two parts, a variable-length prefix part and the remainder of the code-word which is of fixed-length or approximately so. The variable-length part consists of a series of 1\s terminated by a single 0 bit. The number of 1\s preceeding the first 0 bit indicates the prefix of a particualr code-word. The code-words are arranged so that the ones with the same prefix have the same length, or as close as possible. With this novel codification, decoding the variable-length code-words is accomplished in two simple steps. In the first step the decoder determines the position of the first 0 bit in the present code-word, using this information to specify a section of the main memory. The second step is to index into that memory section using the remaining part of that code-word, and this retrieve the fixed length representation.

대표청구항 ▼

A method for decoding input variable-length code-words wherein each code-word is divided into two parts, a prefix part consisting of the first 0 bit together with all the preceding 1 bits and a remainder part consisting of all the bits following the first 0 bit, whereby fixed-length binary data-symb

A method for decoding input variable-length code-words wherein each code-word is divided into two parts, a prefix part consisting of the first 0 bit together with all the preceding 1 bits and a remainder part consisting of all the bits following the first 0 bit, whereby fixed-length binary data-symbols corresponding to said input code-words are stored in sections of a decoding table that correspond to the different code-word prefixes, said decoding method comprising the following two steps: first, determining the position of the first 0 bit in each input code-word and thus the code-word prefix, where said prefix specifies a section in the decoding table; and second, using the input bits following the first 0 bit which include the code-word remainder to specify an address in said decoding table section and thus retrieve the data-symbol corresponding to a given input code-word.