Method of encoding and decoding data in blocks, and encoding and decoding devices implementing this method

Claim 1

A method for encoding blocks of data words, each block having a fixed number of data words with a maximum of n bits, a position of a data word within a block being diagonally identified by a series of absolute addresses having consecutive values, the encoding method comprising the steps of: classifying data words into n classes, an ith class, where i=1 to n, includes data words having as their most significant non-zero bit their ith bit; computing a relative address for each classified data word by subtracting from an absolute address of a classified data word an absolute address of a data word in the class of the classified data word previously encoded; and determining a transmitted address for each classified data word by subtracting from a relative address of a classified data word a number of data words in classes less than the class of the classified data word which have been encoded since a previous encoding of a data word in the class of the classified data word; thereby encoding the classified data word.

Claim 2

A method for encoding blocks of data words, each block having a fixed number of data words with a maximum of n bits, a position of a data word within a block being diagonally identified by a series of absolute addresses having consecutive values, the encoding method comprising the steps of: classifying data words into n classes, an ith class, where i=1 to n, includes data words having as their most significant non-zero bit their ith bit; computing a relative address for each classified data word by subtracting from an absolute address of a classified data word an absolute address of a data word in the class of the classified data word previously encoded; and determining a transmitted address for each classified data word by subtracting from a relative address of a classified data word a number of data words in classes less than the class of the classified data word which have been encoded since a previous encoding of a data word in the class of the classified data word; thereby encoding the classified data word.

Claim 6

A method for decoding encoded data words and their corresponding transmitted addresses into reconstituted blocks of data words, each block having a fixed number of data words with a maximum of n bits, a position of a data word within a block being diagonally identified by a series of absolute addresses having consecutive values, the decoding method comprising the steps of: determining the class of an encoded data word; storing class by class the encoded data words and corresponding transmitted addresses; reconstituting the ith and greater bits of each encoded data word to reconstitute the value of each encoded data word; reconstituting an absolute address of each encoded data word from a corresponding transmitted address, the step of reconstituting an absolute address comprising the steps of: computing a hypothetical relative address for an encoded data word, in each class, which is to be decoded next, by adding the transmitted address of an encoded data word to be decoded to a number of encoded data words in classes less than the class of the encoded data word to be decoded which have been decoded since a previous decoding of an encoded data word in the class of the encoded data word to be decoded, computing a hypothetical absolute address for an encoded data word, in each class, which is to be decoded next, by adding the hypothetical relative address of an encoded data word to be decoded to the absolute address of a previously decoded encoded data word in the class of the encoded data word to be decoded, and selecting as a reconstituted absolute address a hypothetical absolute address having a minimum value from the hypothetical absolute addresses of the encoded data word, in each class, to be decoded next; and reconstituting a block of data words based on the reconstituted values of encoded data words and corresponding reconstituted absolute addresses, and by inserting zero value data words at absolute addresses in the block having no corresponding reconstituted encoded data word.

Claim 7

A method for decoding encoded data words and their corresponding transmitted addresses into reconstituted blocks of data words, each block having a fixed number of data words with a maximum of n bits, a position of a data word within a block being diagonally identified by a series of absolute addresses having consecutive values, the decoding method comprising the steps of: determining the class of an encoded data word; storing class by class the encoded data words and corresponding transmitted addresses; reconstituting the ith and greater bits of each encoded data word to reconstitute the value of each encoded data word; reconstituting an absolute address of each encoded data word from a corresponding transmitted address, the step of reconstituting an absolute address comprising the steps of: computing a hypothetical relative address for an encoded data word, in each class, which is to be decoded next, by adding the transmitted address of an encoded data word to be decoded to a number of encoded data words in classes less than the class of the encoded data word to be decoded which have been decoded since a previous decoding of an encoded data word in the class of the encoded data word to be decoded, computing a hypothetical absolute address for an encoded data word, in each class, which is to be decoded next, by adding the hypothetical relative address of an encoded data word to be decoded to the absolute address of a previously decoded encoded data word in the class of the encoded data word to be decoded, and selecting as a reconstituted absolute address a hypothetical absolute address having a minimum value from the hypothetical absolute addresses of the encoded data word, in each class, to be decoded next; and reconstituting a block of data words based on the reconstituted values of encoded data words and corresponding reconstituted absolute addresses, and by inserting zero value data words at absolute addresses in the block having no corresponding reconstituted encoded data word.

Claim 9

A method for encoding and decoding blocks of data words, each block having a fixed number of data words with a maximum of n bits, a position of a data word within a block being diagonally identified by a series of absolute addresses having consecutive values, the encoding and decoding method comprising the steps of: encoding data words for transmission comprising the steps of: classifying data words into n classes, an ith class, where i=1 to n, includes data words having as their most significant non-zero bit their ith bit; computing a relative address for each classified data word by subtracting from an absolute address of a classified data word an absolute address of a data word in the class of the classified data word previously encoded; determining a transmitted address for each classified data word by subtracting from a relative address of a classified data word a number of data words in classes less than the class of the classified data word which have been encoded since a previous encoding of a data word in the class of the classified data word, thereby encoding the classified data word; and transmitting class by class and in an order of increasing absolute addresses within each class, encoded data words and their corresponding transmitted addresses, the encoded data words of each ith class having their ith and greater bits omitted from transmission, the transmitted addresses being encoded according to a variable length code; decoding the transmitted encoded data words to reconstitute the blocks of data words comprising the steps of: determining the class of an encoded data word; storing class by class the encoded data words and corresponding transmitted addresses; reconstituting the ith and greater bits of each encoded data word to reconstitute the value of each encoded data word; reconstituting an absolute address of each encoded data word from a corresponding transmitted address, the step of reconstituting an absolute address comprising the steps of: computing a hypothetical relative address for an encoded data word, in each class, which is to be decoded next, by adding the transmitted address of an encoded data word to be decoded to a number of encoded data words in classes less than the class of the encoded data word to be decoded which have been decoded since a previous decoding of an encoded data word in the class of the encoded data word to be decoded, computing a hypothetical absolute address for an encoded data word, in each class, which is to be decoded next, by adding the hypothetical relative address of an encoded data word to be decoded to the absolute address of a previously decoded encoded data word in the class of the encoded data word to be decoded, selecting as a reconstituted absolute address a hypothetical absolute address having a minimum value from the hypothetical absolute addresses of the encoded data word, in each class, to be decoded next; and reconstituting a block of data words based on the reconstituted values of encoded data words and their corresponding reconstituted absolute addresses, and by inserting zero value data words at absolute addresses in the block having no corresponding reconstituted encoded data word.

Claim 10

A method for encoding and decoding blocks of data words, each block having a fixed number of data words with a maximum of n bits, a position of a data word within a block being diagonally identified by a series of absolute addresses having consecutive values, the encoding and decoding method comprising the steps of: encoding data words for transmission comprising the steps of: classifying data words into n classes, an ith class, where i=1 to n, includes data words having as their most significant non-zero bit their ith bit; computing a relative address for each classified data word by subtracting from an absolute address of a classified data word an absolute address of a data word in the class of the classified data word previously encoded; determining a transmitted address for each classified data word by subtracting from a relative address of a classified data word a number of data words in classes less than the class of the classified data word which have been encoded since a previous encoding of a data word in the class of the classified data word, thereby encoding the classified data word; and transmitting class by class and in an order of increasing absolute addresses within each class, encoded data words and their corresponding transmitted addresses, the encoded data words of each ith class having their ith and greater bits omitted from transmission, the transmitted addresses being encoded according to a variable length code; decoding the transmitted encoded data words to reconstitute the blocks of data words comprising the steps of: determining the class of an encoded data word; storing class by class the encoded data words and corresponding transmitted addresses; reconstituting the ith and greater bits of each encoded data word to reconstitute the value of each encoded data word; reconstituting an absolute address of each encoded data word from a corresponding transmitted address, the step of reconstituting an absolute address comprising the steps of: computing a hypothetical relative address for an encoded data word, in each class, which is to be decoded next, by adding the transmitted address of an encoded data word to be decoded to a number of encoded data words in classes less than the class of the encoded data word to be decoded which have been decoded since a previous decoding of an encoded data word in the class of the encoded data word to be decoded, computing a hypothetical absolute address for an encoded data word, in each class, which is to be decoded next, by adding the hypothetical relative address of an encoded data word to be decoded to the absolute address of a previously decoded encoded data word in the class of the encoded data word to be decoded, selecting as a reconstituted absolute address a hypothetical absolute address having a minimum value from the hypothetical absolute addresses of the encoded data word, in each class, to be decoded next; and reconstituting a block of data words based on the reconstituted values of encoded data words and their corresponding reconstituted absolute addresses, and by inserting zero value data words at absolute addresses in the block having no corresponding reconstituted encoded data word.

Claim 11

An apparatus for encoding blocks of data words, each block having a series of data words comprising a maximum of n bits, the apparatus comprising: means for determining a most significant non-zero bit in each data word; n first memory means for storing in an ith first memory, where i=1 to n, i-1 th and lower bits of data words determined to have as their most significant non-zero bit their ith bit; means for computing an absolute address for each data word based on a position of a data word in a block; means for computing a relative address for each data word stored in the n first memory means, by subtracting from the absolute address of a stored data word the absolute address of a data word previously stored in the same ith first memory means; means for counting the number (Ci) of data words stored in each of the 1 to i-1 first memory means since a previous storage of a data word in the ith first memory means; means for computing a transmitted address for each stored data word by subtracting from the relative address of a data word stored in the ith first memory means the number (Ci); and n second memory means for storing the transmitted addresses of the corresponding data words stored in the n first memory means.

Claim 12

An apparatus for encoding blocks of data words, each block having a series of data words comprising a maximum of n bits, the apparatus comprising: means for determining a most significant non-zero bit in each data word; n first memory means for storing in an ith first memory, where i=1 to n, i-1 th and lower bits of data words determined to have as their most significant non-zero bit their ith bit; means for computing an absolute address for each data word based on a position of a data word in a block; means for computing a relative address for each data word stored in the n first memory means, by subtracting from the absolute address of a stored data word the absolute address of a data word previously stored in the same ith first memory means; means for counting the number (Ci) of data words stored in each of the 1 to i-1 first memory means since a previous storage of a data word in the ith first memory means; means for computing a transmitted address for each stored data word by subtracting from the relative address of a data word stored in the ith first memory means the number (Ci); and n second memory means for storing the transmitted addresses of the corresponding data words stored in the n first memory means.

Claim 14

An apparatus for decoding blocks of encoded data words, the decoding apparatus receiving a series of encoded data words having their most significant non-zero bit and greater bits omitted, the decoding apparatus receiving transmitted addresses, encoded according to a variable length code, corresponding to the encoded data words, the encoded data words being received in groups of encoded data words having an equal number of bits; the decoding apparatus comprising: means for receiving the transmitted addresses and decoding the transmitted addresses according to a variable length code; means for reconstituting the most significant non-zero bit and greater bits of each encoded data word; n first memory means for storing the reconstituted data words, data words having as their most significant non-zero bit their ith bit being stored in the ith first memory means, where i ÷1 to n; n second memory means for storing the decoded transmitted addresses corresponding the data words stored in the n first memory means; means associated with the n first and second memory means for reconstituting an absolute address for each stored data word, the means for reconstituting: counting a number (Ci) of data words stored in each of the 1 to i-1 first memory means since a previous storage of a data word in the ith first memory means, computing a hypothetical relative address for a stored data word, in each n first memory means, which is to be read next, by adding the transmitted address of a stored data word to be read to the number (Ci), computing a hypothetical absolute address for a stored data word, in each n first memory means, which is to be read next, by adding the hypothetical relative address of the stored data word to be read to the absolute address of a stored data word previously read from the same ith first memory means as the stored data word to be read, and selecting as a reconstituted absolute address a hypothetical absolute address having a minimum value from the hypothetical absolute addresses of the stored data words, in each n first memory means, to be read next; and means for reading from the n first memory means the stored data word corresponding to the reconstituted absolute address and storing the data word in a third memory means in order of increasing reconstituted absolute addresses to reconstitute a block of encoded data words.

Claim 15

An apparatus for decoding blocks of encoded data words, the decoding apparatus receiving a series of encoded data words having their most significant non-zero bit and greater bits omitted, the decoding apparatus receiving transmitted addresses, encoded according to a variable length code, corresponding to the encoded data words, the encoded data words being received in groups of encoded data words having an equal number of bits; the decoding apparatus comprising: means for receiving the transmitted addresses and decoding the transmitted addresses according to a variable length code; means for reconstituting the most significant non-zero bit and greater bits of each encoded data word; n first memory means for storing the reconstituted data words, data words having as their most significant non-zero bit their ith bit being stored in the ith first memory means, where i ÷1 to n; n second memory means for storing the decoded transmitted addresses corresponding the data words stored in the n first memory means; means associated with the n first and second memory means for reconstituting an absolute address for each stored data word, the means for reconstituting: counting a number (Ci) of data words stored in each of the 1 to i-1 first memory means since a previous storage of a data word in the ith first memory means, computing a hypothetical relative address for a stored data word, in each n first memory means, which is to be read next, by adding the transmitted address of a stored data word to be read to the number (Ci), computing a hypothetical absolute address for a stored data word, in each n first memory means, which is to be read next, by adding the hypothetical relative address of the stored data word to be read to the absolute address of a stored data word previously read from the same ith first memory means as the stored data word to be read, and selecting as a reconstituted absolute address a hypothetical absolute address having a minimum value from the hypothetical absolute addresses of the stored data words, in each n first memory means, to be read next; and means for reading from the n first memory means the stored data word corresponding to the reconstituted absolute address and storing the data word in a third memory means in order of increasing reconstituted absolute addresses to reconstitute a block of encoded data words.