Claims Analysis
Total Claims: 14
Claim 1
Independent Claim
318 words
A method of utilizing a data processor to decode input variable-length code words of the kind which contain variablelength prefixes each uniquely representing a particular code word length, the most frequently occurring input code words having the shortest word lengths and the shortest prefixes, said decoding method comprising the following steps: a. utilizing the leading N-bit portions of the successive input code words (where N is an integer at least equal to the number of bits in the longest of said prefixes) to address selected locations in a first memory of said data processor, each of said first memory locations being allocated to a particular length-indicating prefix and none other; b. storing in a second memory of said data processor, at each of a plurality of locations therein which correspond respectively to addressable locations in said first memory, information that pertains to all of the code words containing the particular length-indicating prefix to which the respective first memory location is allocated, such information including a base address assigned to all code words having that same prefix; c. storing in a third memory of said data processor, at locations therein which are addressable by combinations of predetermined base addresses and selected displacements therefrom, a plurality of fixed-length code words which are the decoded equivalents of at least the most frequently occurring variable-length code words; d. reading a base address from the location in said second memory which corresponds to that location in said first memory which is addressed by the prefix of the input code word that currently is being decoded; e. forming from the base address read in step (d) a final address which incorporates a displacement value denoted by the remaining portion of the current input code word; and f. reading from the particular location in said third memory which is designated by said final address a fixed-length code word corresponding the the current input code word.
Claim 2
Independent Claim
318 words
A method of utilizing a data processor to decode input variable-length code words of the kind which contain variablelength prefixes each uniquely representing a particular code word length, the most frequently occurring input code words having the shortest word lengths and the shortest prefixes, said decoding method comprising the following steps: a. utilizing the leading N-bit portions of the successive input code words (where N is an integer at least equal to the number of bits in the longest of said prefixes) to address selected locations in a first memory of said data processor, each of said first memory locations being allocated to a particular length-indicating prefix and none other; b. storing in a second memory of said data processor, at each of a plurality of locations therein which correspond respectively to addressable locations in said first memory, information that pertains to all of the code words containing the particular length-indicating prefix to which the respective first memory location is allocated, such information including a base address assigned to all code words having that same prefix; c. storing in a third memory of said data processor, at locations therein which are addressable by combinations of predetermined base addresses and selected displacements therefrom, a plurality of fixed-length code words which are the decoded equivalents of at least the most frequently occurring variable-length code words; d. reading a base address from the location in said second memory which corresponds to that location in said first memory which is addressed by the prefix of the input code word that currently is being decoded; e. forming from the base address read in step (d) a final address which incorporates a displacement value denoted by the remaining portion of the current input code word; and f. reading from the particular location in said third memory which is designated by said final address a fixed-length code word corresponding the the current input code word.
Claim 3
Independent Claim
318 words
A method of utilizing a data processor to decode input variable-length code words of the kind which contain variable-length prefixes each uniquely representing a particular code word length, the most frequently occurring input code words having the shortest word lengths and the shortest prefixes, said decoding method comprising the following steps: a. utilizing the leading N-bit portions of the successive input code words (where N is an integer at least equal to the number of bits in the longest of said prefixes) to address selected locations in a first memory of said data processor, each of said first memory locations being allocated to a particular length-indicating prefix and none other; b. storing in a second memory of said data processor, at each of a plurality of locations therein which correspond respectively to addressable locations in said first memory, information that pertains to all of the code words containing the particular length-indicating prefix to which the respective first memory location is allocated, such information including a base address assigned to all code words having that same prefix; c. storing in a third memory of said data processor, at locations therein which are addressable by combinations of predetermined base addresses and selected displacements therefrom, a plurality of fixed-length code words which are the decoded equivalents of at least the most frequently occurring variable-length code words; d. reading a base address from the location in said second memory which corresponds to that location in said first memory which is addressed by the prefix of the input code word that currently is being decoded; e. forming from the base address read in step (d) a final address which incorporates a displacement value denoted by the remaining portion of the current input code word; and f. reading from the particular location in said third memory which is designated by said final address a fixed-length code word corresponding the the current input code word.
Claim 4
Independent Claim
318 words
A method of utilizing a data processor to decode input variable-length code words of the kind which contain variable-length prefixes each uniquely representing a particular code word length, the most frequently occurring input code words having the shortest word lengths and the shortest prefixes, said decoding method comprising the following steps: a. utilizing the leading N-bit portions of the successive input code words (where N is an integer at least equal to the number of bits in the longest of said prefixes) to address selected locations in a first memory of said data processor, each of said first memory locations being allocated to a particular length-indicating prefix and none other; b. storing in a second memory of said data processor, at each of a plurality of locations therein which correspond respectively to addressable locations in said first memory, information that pertains to all of the code words containing the particular length-indicating prefix to which the respective first memory location is allocated, such information including a base address assigned to all code words having that same prefix; c. storing in a third memory of said data processor, at locations therein which are addressable by combinations of predetermined base addresses and selected displacements therefrom, a plurality of fixed-length code words which are the decoded equivalents of at least the most frequently occurring variable-length code words; d. reading a base address from the location in said second memory which corresponds to that location in said first memory which is addressed by the prefix of the input code word that currently is being decoded; e. forming from the base address read in step (d) a final address which incorporates a displacement value denoted by the remaining portion of the current input code word; and f. reading from the particular location in said third memory which is designated by said final address a fixed-length code word corresponding the the current input code word.
Claim 5
Independent Claim
356 words
A method of utilizing a data processor to decode input variable-length code words of the kind which contain variable-length prefixes each uniquely representing a particular code word length, the more frequently occurring input code words having the shorter word lengths and the shorter prefixes, the less frequently occurring input code words being of the same length and each having a special prefix common to all such words to which is appended a distinctive fixed-length code word, said decoding method comprising the following steps: a. utilizing the leading N-bit portions of the successive input code words (where N is an integer at least equal to the number of bits in the longest of said prefixes) to address selected locations in a first memory of said data processor, each of said first memory locations being allocated to a particular length-indicating prefix and none other; b. storing in a second memory of said data processor, at locations therein which correspond respectively to the addressable locations in said first memory, information pertaining to the various prefixes, such information including, in the case of the less frequently occurring input code words identified by said special prefix, an indication that such a prefix has been detected; c. storing in a third memory of said data processor the fixed-length code words which are the decoded equivalents of the more frequently occurring variable-length code words; d. reading the information from that location in said second memory which corresponds to the location in said first memory which currently is being addressed by a variable-length code word prefix; and e. performing one of the following steps according to the type of information read from said second memory: e1. if a special code indication is absent, then using the base address corresponding to the prefix in conjunction with the remainder of the input code word to form a final address for said third memory, and reading from said third memory the fixed-length code word thus addressed as the decoded output; or e2. if a special code indication is present, then discarding the special prefix and identifying the remainder of the input code word as the decoded output.
Claim 6
Independent Claim
356 words
A method of utilizing a data processor to decode input variable-length code words of the kind which contain variable-length prefixes each uniquely representing a particular code word length, the more frequently occurring input code words having the shorter word lengths and the shorter prefixes, the less frequently occurring input code words being of the same length and each having a special prefix common to all such words to which is appended a distinctive fixed-length code word, said decoding method comprising the following steps: a. utilizing the leading N-bit portions of the successive input code words (where N is an integer at least equal to the number of bits in the longest of said prefixes) to address selected locations in a first memory of said data processor, each of said first memory locations being allocated to a particular length-indicating prefix and none other; b. storing in a second memory of said data processor, at locations therein which correspond respectively to the addressable locations in said first memory, information pertaining to the various prefixes, such information including, in the case of the less frequently occurring input code words identified by said special prefix, an indication that such a prefix has been detected; c. storing in a third memory of said data processor the fixed-length code words which are the decoded equivalents of the more frequently occurring variable-length code words; d. reading the information from that location in said second memory which corresponds to the location in said first memory which currently is being addressed by a variable-length code word prefix; and e. performing one of the following steps according to the type of information read from said second memory: e1. if a special code indication is absent, then using the base address corresponding to the prefix in conjunction with the remainder of the input code word to form a final address for said third memory, and reading from said third memory the fixed-length code word thus addressed as the decoded output; or e2. if a special code indication is present, then discarding the special prefix and identifying the remainder of the input code word as the decoded output.
Claim 7
Independent Claim
323 words
A method of utilizing a data processor for converting into code words of fixed length a succession of input code words whose respective lengths may vary inversely according to their relative frequency of occurrence, said input code words including prefixes uniquely identifying the respective lengths of their code words, the lengths of the respective prefixes also varying inversely according to their relative frequency of occurrence, whereby the most frequently occurring input code words have the shortest lengths and the shortest prefixes, said method comprising the steps of: a. storing in a first storage unit of said processor, at every location therein which may be addressed by a prefix of an input code word, information pertaining to all code words of the length denoted by the respective prefix, such information including, at least for those input code words of relatively high frequency, an indication of how many bits remain in each of these code words after the prefix thereof has been excluded and also including a base address assigned to all code words having that prefix; b. utilizing the prefix of the current input code word as an address for locating and reading out of said first storage unit the information pertaining to that prefix; c. combining the base address for the current input code word with a displacement value represented by the remainder of that code word (as indicated by the information read out in step (b) to form a final address which is unique to that specific input code word; d. storing in a second storage unit of said processor, at locations therein which are respectively addressable by the various addresses that may be formed in step (c), fixed-length code words respectively representing the input code words from which said final addresses are derived; and e. reading from said second storage unit the fixed-length code designated by the final address formed in step (c) as the output of said code conversion method.
Claim 8
Independent Claim
323 words
A method of utilizing a data processor for converting into code words of fixed length a succession of input code words whose respective lengths may vary inversely according to their relative frequency of occurrence, said input code words including prefixes uniquely identifying the respective lengths of their code words, the lengths of the respective prefixes also varying inversely according to their relative frequency of occurrence, whereby the most frequently occurring input code words have the shortest lengths and the shortest prefixes, said method comprising the steps of: a. storing in a first storage unit of said processor, at every location therein which may be addressed by a prefix of an input code word, information pertaining to all code words of the length denoted by the respective prefix, such information including, at least for those input code words of relatively high frequency, an indication of how many bits remain in each of these code words after the prefix thereof has been excluded and also including a base address assigned to all code words having that prefix; b. utilizing the prefix of the current input code word as an address for locating and reading out of said first storage unit the information pertaining to that prefix; c. combining the base address for the current input code word with a displacement value represented by the remainder of that code word (as indicated by the information read out in step (b) to form a final address which is unique to that specific input code word; d. storing in a second storage unit of said processor, at locations therein which are respectively addressable by the various addresses that may be formed in step (c), fixed-length code words respectively representing the input code words from which said final addresses are derived; and e. reading from said second storage unit the fixed-length code designated by the final address formed in step (c) as the output of said code conversion method.
Claim 9
Dependent Claim
115 words
A code conversion method as specified in claim 3 which is operable also upon less frequently occurring input code words of the type wherein a prefix common to all such code words is combined with an individual fixed-length code word, the information stored in step (a) and read out in step (b) including an indication as to whether or not the input code word is of this type, said method including the following step as an alternative to steps (c) and (e): f. in the case of each input code word that has said common prefix, reading the remainder of said code word exclusive of said prefix as the output of said code conversion method.
Claim 10
Dependent Claim
357 words
Apparatus for decoding into fixed-length code words a succession of input code words each represented by serially ordered bits, at least some of said input code words being of a type produced by a variable-length encoding process and respectively having variable-length prefixes, each such prefix being unique to a particular code word length, said apparatus comprising: a. an argument register; b. means for entering into said argument register at least that much of the leading portion of an input code word which contains the prefix thereof; c. an associative memory containing stored code words each corresponding uniquely to a respective length indicating prefix, said associative memory being operable in response to the argument stored in said argument register to generate a match-indicating signal distinctively related to the prefix included in said argument; d. a storage unit having word-storing locations therein respectively corresponding to said code-length indicating prefixes and respectively addressable by the match-indicating signals which are generated by said associative memory in response to such prefixes, each of said locations storing a word which contains a base address unique to the respective prefix and other information indicating which, if any, of the bit positions in an input code word of that length is occupied by bits exclusIve of that prefix; e. a data register for storing the base address of the word in said storage unit which is designated by the match-indicating signal from said associative memory; f. means responsive to the other information in said designated word for forming in said data register a new address which is determined in part by said base address and in part by the remainder of the current input code word exclusive of its prefix; g. a memory having storage locations therein which are addressable in accordance with addresses that may be stored in said data register, each such location storing a fixed-length code word corresponding to the input code word that produced the respective address stored in said data register; and h. means for reading from said memory (g), as the decoded output of said apparatus, the fixed-length code word designated by the address stored in said data register.
Claim 11
Dependent Claim
357 words
Apparatus for decoding into fixed-length code words a succession of input code words each represented by serially ordered bits, at least some of said input code words being of a type produced by a variable-length encoding process and respectively having variable-length prefixes, each such prefix being unique to a particular code word length, said apparatus comprising: a. an argument register; b. means for entering into said argument register at least that much of the leading portion of an input code word which contains the prefix thereof; c. an associative memory containing stored code words each corresponding uniquely to a respective length indicating prefix, said associative memory being operable in response to the argument stored in said argument register to generate a match-indicating signal distinctively related to the prefix included in said argument; d. a storage unit having word-storing locations therein respectively corresponding to said code-length indicating prefixes and respectively addressable by the match-indicating signals which are generated by said associative memory in response to such prefixes, each of said locations storing a word which contains a base address unique to the respective prefix and other information indicating which, if any, of the bit positions in an input code word of that length is occupied by bits exclusIve of that prefix; e. a data register for storing the base address of the word in said storage unit which is designated by the match-indicating signal from said associative memory; f. means responsive to the other information in said designated word for forming in said data register a new address which is determined in part by said base address and in part by the remainder of the current input code word exclusive of its prefix; g. a memory having storage locations therein which are addressable in accordance with addresses that may be stored in said data register, each such location storing a fixed-length code word corresponding to the input code word that produced the respective address stored in said data register; and h. means for reading from said memory (g), as the decoded output of said apparatus, the fixed-length code word designated by the address stored in said data register.
Claim 12
Dependent Claim
123 words
Apparatus as set forth in claim 5 which is further adapted to decode input code words of a special type in each of which a bit string of fixed length is appended to a length-indicating prefix that is common to all input code words of that type, the information contained in each of the words stored in storage unit (d) including an indication as to whether or not such word is addressable by said common prefix, said apparatus further including: i. means responsive to an indication that the currently designated word in storage unit (d) is being addressed by said common prefix for causing the fixed-length bit string appended to that prefix to be read out of said apparatus as its decoded output.
Claim 13
Dependent Claim
651 words
Apparatus for decoding into code words of fixed length a succession of input code words at least some of which are of variable length, said variable-length code words respectively having variable-length prefixes each uniquely identifying a particular code word length, none of said prefixes constituting the beginning of a longer prefix, and none of said variable-length code words constituting the beginning of a longer code word, said apparatus comprising: a. an associative memory having rows and columns of bit-storing memory cells, said rows respectively storing the bits of the various prefixes, with the bit in the leading position of each prefix occupying a common registration column, there being at least as many cells in each row as there are bits in the longest prefix, each of said cells being capable of assuming either one of two binary states or else a third state in which its setting is considered to be without binary significance, this third state being assumed by any cell that does not store a prefix bit; b. an argument register for said associative memory having positions therein for storing the respective bits of an argument, with the position at which the leading bit of the argument is stored corresponding to said registration column, said register having a capacity for storing at least the number of bits contained in the longest prefix; c. ingating means operable for causing the bits of a bit stream containing the successive input code words to be serially entered into said argument register; d. shift means operable to shift the bits stored in said argument register progressively through selected numbers of positions in the direction of said leading bit position, with exit from said register being accomplished by any bit when it leaves said leading bit position; e. a readable storage unit controlled by the output of said associative memory, said storage unit containing rows of bit-storage cells, each such row of cells being allocated to a corresponding row of cells in said associative memory and containing cells which store a representation of the respective prefix length, other cells which store a representation indicating the number of remaining bits in any code word of the length denoted by said prefix, and still other cells storing a representation of a base address assigned to code words of that length; f. means for operAting said associative memory to find a match between the prefix currently stored in the leading portion of said argument register and the prefixes stored in the respective rows of said associative memory; g. means for reading out of said storage unit the items of information stored in the row of cells which corresponds to the matching prefix found by said associative memory; h. data registers for respectively storing the prefix length, remainder length and base address read out of said storage unit during each such association; i. means responsive to the setting of said prefix length register to operate said shift means (d) and ingating means (c) for shifting out of said argument register the prefix bits of the current input code word and for entering an equal number of bits into said argument register; j. means responsive to the setting of said remainder length register to operate said shifting means (d) and ingating means (c) for shifting out of said argument register the remaining bits, if any, of the current input code word and entering an equal number of bits into said argument register; k. address forming means operable to form a final address which is determined in part by the initial setting of said base address register and in part by the remainder, if any, which is read out of said argument register by means (j); and l. a decoded word store addressable by the final address formed by means (k) to furnish a fixed-length code word equivalent to the input code word from which said final address was derived.
Claim 14
Dependent Claim
651 words
Apparatus for decoding into code words of fixed length a succession of input code words at least some of which are of variable length, said variable-length code words respectively having variable-length prefixes each uniquely identifying a particular code word length, none of said prefixes constituting the beginning of a longer prefix, and none of said variable-length code words constituting the beginning of a longer code word, said apparatus comprising: a. an associative memory having rows and columns of bit-storing memory cells, said rows respectively storing the bits of the various prefixes, with the bit in the leading position of each prefix occupying a common registration column, there being at least as many cells in each row as there are bits in the longest prefix, each of said cells being capable of assuming either one of two binary states or else a third state in which its setting is considered to be without binary significance, this third state being assumed by any cell that does not store a prefix bit; b. an argument register for said associative memory having positions therein for storing the respective bits of an argument, with the position at which the leading bit of the argument is stored corresponding to said registration column, said register having a capacity for storing at least the number of bits contained in the longest prefix; c. ingating means operable for causing the bits of a bit stream containing the successive input code words to be serially entered into said argument register; d. shift means operable to shift the bits stored in said argument register progressively through selected numbers of positions in the direction of said leading bit position, with exit from said register being accomplished by any bit when it leaves said leading bit position; e. a readable storage unit controlled by the output of said associative memory, said storage unit containing rows of bit-storage cells, each such row of cells being allocated to a corresponding row of cells in said associative memory and containing cells which store a representation of the respective prefix length, other cells which store a representation indicating the number of remaining bits in any code word of the length denoted by said prefix, and still other cells storing a representation of a base address assigned to code words of that length; f. means for operAting said associative memory to find a match between the prefix currently stored in the leading portion of said argument register and the prefixes stored in the respective rows of said associative memory; g. means for reading out of said storage unit the items of information stored in the row of cells which corresponds to the matching prefix found by said associative memory; h. data registers for respectively storing the prefix length, remainder length and base address read out of said storage unit during each such association; i. means responsive to the setting of said prefix length register to operate said shift means (d) and ingating means (c) for shifting out of said argument register the prefix bits of the current input code word and for entering an equal number of bits into said argument register; j. means responsive to the setting of said remainder length register to operate said shifting means (d) and ingating means (c) for shifting out of said argument register the remaining bits, if any, of the current input code word and entering an equal number of bits into said argument register; k. address forming means operable to form a final address which is determined in part by the initial setting of said base address register and in part by the remainder, if any, which is read out of said argument register by means (j); and l. a decoded word store addressable by the final address formed by means (k) to furnish a fixed-length code word equivalent to the input code word from which said final address was derived.