Decoder, decoding method, look-up table, and decoding program

Details Decoder, decoding method, look-up table, and decoding program Decoder, decoding method, look-up table, and decoding program

2003-02-03

2004-11-02

JeanPierre, Peguy (Department: 2819)

Coded data generation or conversion

Digital code to digital code converters

C341S051000

Reexamination Certificate

active

06812863

ABSTRACT:

TECHNICAL FIELD
The present invention relates to a decoder, a decoding method, a look-up table, and a decoding program. In particular, the present invention is suitably applied to decode a variable length code.
BACKGROUND ART
There is a conventional method of decoding a variable length code in such a manner that a code bit is read bit by bit from an input bit sequence and symbol corresponding to a bit sequence defined in a code table is output as a decoding result when a match occurs between the read bits and the bit sequence. According to this method, decoding can be performed by using a small storage capacity if only code bits corresponding to symbols are held. However, the time required for decoding is long because of one-bit-by-one-bit reading of code bits from the input bit sequence.
On the other hand, there is a method in which, to perform high-speed decoding, symbols from which a decoding result is obtained and the code lengths of the symbols are registered in a look-up table with respect to all entries (candidates) identified by bit sequences corresponding to maximum code lengths.
FIG. 8
is a diagram showing the configuration of such a conventional look-up table.
Referring to FIG.
8
(
a
), a code table
11
is defined in which, for example, a code bit “1” is assigned to a symbol S
0
, code bits “01” are assigned to a symbol S
1
, code bits “001” are assigned to a symbol S
2
, and code bits “000” are assigned to a symbol S
3
.
In this case, since the maximum code length of the code bits is 3, sequence numbers, the symbols and code lengths are registered in a look-up table
12
with respect to 2
3
entries identified by 3-bit bit sequences, as shown in FIG.
8
(
b
) The sequence numbers may be numbers in decimal notation corresponding to the 3-bit bit sequences.
Bits are read in groups of three bits from the input bit sequence. Each read group of three bits is converted into a decimal number to obtain the sequence number. When the sequence number is obtained, the look-up table
12
is referred to with the sequence number used as a key to obtain the symbol corresponding to the sequence number. The symbol is output as a decoding result while the bit(s) remaining as a result of subtraction of the bits corresponding to the code length of the symbol from bit sequences for 3-bit is returned to the input bit sequence.
For example, if the input bit sequence is “1010010110 . . . ”, the first three bits “101” of this bit sequence is read and this “101” is converted into a decimal number to obtain a sequence number “5”. The look-up table
12
shown in FIG.
8
(
b
) is referred to with this sequence number “5” used as a key to obtain the symbol “S0” and the code length “1” of the symbol “S0”. The symbol “S0” is output as a decoding result.
Also, the code length “1” of the symbol “S0” is compared with the maximum code length “3”. Two bits “01” corresponding to the difference between the maximum code length “3” and the code length “1” of the symbol “S0” are considered unused in decoding of the symbol “S0” and are returned to the input bit sequence “0010110 . . . ”.
Subsequently, the next three bits “010” are read from the input bit sequence “010010110 . . . ” to which the bits have been returned, and the above-described processing is repeated.
According to the above-described conventional method, decoding can be performed only by reading one time the bit sequence corresponding to the maximum code length from the input bit sequence and, consequently, the speed of decoding can be increased.
The above-described conventional method, however, requires registering symbols and the code lengths of the symbols with respect to all entries identified by bit sequences corresponding to the maximum code length and therefore has the problem that the capacity of look-up table
12
is increased.
That is, in the example shown in
FIG. 8
, there are only four symbols but it is necessary to register 2
3
entries of the symbols and the code lengths of the symbols and, therefore, the same symbol is registered in a state of being duplicated a certain number of times.
In particular, if the maximum code length is larger, the number of times one symbol is duplicated when the symbol is registered is increased, resulting in an increase in capacity of the look-up table. For example, if the maximum code length is 16 and if 100 bytes per symbol and one byte for storage of the code length are needed, a capacity of 2
16
×101 bytes is required.
It is, therefore, an object of the present invention to provide a decoder, a decoding method, a look-up table, and a decoding program enabling decoding at a higher speed with a small storage capacity.
DISCLOSURE OF THE INVENTION
A decoder provided to attain the above-described object according to the present invention is characterized by comprising symbol association means for associating each of reference places for encoded symbols with one of the symbols, and reference place identification means for identifying the reference place for each of symbols when the bit sequence assigned to the symbol and corresponding to the maximum code length is cut out.
It is thereby ensured that while the symbols are not registered in correspondence with all entries identified by the bit sequences corresponding to the maximum code length, each encoded symbol can be identified when the corresponding bit sequence is cut out, and it is possible to reduce the storage capacity while maintaining high-speed decoding performance.
A decoder according to another aspect of the present invention is characterized by comprising reference place information acquisition means for obtaining, on the basis of a bit sequence of a predetermined length, reference place information for a decoding result corresponding to a code contained in the bit sequence, and decoding result acquisition means for obtaining, on the basis of the reference place information obtained by the reference place information acquisition means, the decoding result corresponding to the code contained in the bit sequence.
It is thereby ensured that symbols and entries identified by bit sequences having a predetermined length can be associated with each other while avoiding registration of each symbol to be obtained as a decoding result in a duplicated state, thereby reducing the storage capacity while maintaining high-speed decoding performance.
A decoder according to still another aspect of the present invention, which decodes each of encoded symbols, is characterized by comprising an input buffer which reads from an input bit sequence a bit sequence having a length corresponding to the maximum code length of the symbols, a first sequence table in which reference place information for the symbols is registered with respect to each bit sequence having a length corresponding to the maximum code length of the symbols, a second sequence table in which the symbol and the code length of the symbol are registered with respect to each symbol, first reference means for obtaining the reference place information for the symbol corresponding to a code contained in the bit sequence read to the input buffer by referring to the first sequence table on the basis of the bit sequence, second reference means for obtaining the symbol corresponding to the code contained in the bit sequence and the code length of the symbol by referring to the second sequence table on the basis of the reference place information obtained by the first reference means, and return means for returning to the input bit sequence the remaining bit excluding the bit corresponding to the code length of the symbol from the bit sequence read to the input buffer.
It is thereby ensured that each symbol and the code length of the symbol can be obtained from the second sequence table if only reference places for the symbols are registered in the first sequence table by being associated with the bit sequences having a length corresponding to the maximum code length.
Therefore decoding of the variable length code can be performed by making access two times to the first sequence table and the

Affiliated with

Also associated with

Rating

Decoder, decoding method, look-up table, and decoding program does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Decoder, decoding method, look-up table, and decoding program, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Decoder, decoding method, look-up table, and decoding program will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3322213