Abstract: Provided are a method and an apparatus for selectively encoding/decoding point sequences to maximize bit efficiency of a lightweight application scene representation (LASeR) binary stream. The point sequence encoding method includes the steps of: for each point sequence, (a) selecting one of exponential-Golomb (EG) encoding and fixed length (FL) encoding schemes; (b) when the FL encoding scheme is selected, encoding the point sequence using the FL encoding scheme to generate a binary stream; and (c) when the EG encoding scheme is selected, encoding the point sequence using the EG encoding scheme to generate a binary stream. The binary stream includes a flag indicating which encoding scheme is selected and a parameter k, with which the EG encoding can be most effectively performed, when the EG encoding scheme is selected. According to the encoding method, LASeR point sequences can be efficiently encoded and, during a decoding process, a large overhead is not incurred to a decoder (terminal).

Abstract: A system and method to adjust the symbol constellation used to de-map a signal in an optical coherent communications system. A feedback de-mapper adjustment module is configured to compare average Error Vector Magnitude and Bit Error Ratio to a pre-defined table to determine when a constellation de-mapper mismatch occurs. The feedback de-mapper adjustment module then rotates the de-mapper constellation in order to compensate for phase drift.

Abstract: An integrated circuit (IC) having selectable memory elements is provided. The IC includes a logic array block (LAB) disposed within the IC. A plurality of logic elements, having look-up tables functioning as the selectable memory elements is included in the LAB. Within a logic element, a data path that shares multiplexers and drivers when the look-up tables of the logic elements are operated as one of a memory element or a combinational logic device is provided. In addition, a write address decoder is interconnected with the plurality of logic elements through a write bus.

Abstract: A terminal and a base station using an adaptive codebook for a polarization are provided. The terminal includes a rotation matrix generator to generate rotation matrices corresponding to a cross polarization discrimination value (XPD) of a transmission signal and a codebook generator to generate a codebook including result matrices that are generated based on pre-stored block diagonal matrices and the rotation matrices.

Abstract: Managing a string translation table, where a compressor includes information in the table entries to distinguish strings as more or less important. More particularly, the compressor can identify a string in an input stream, determine an age indicator for the string, and record the age indicator in the table. When the string is encountered or identified, a use indicator or counter may be adjusted to indicate use of the entry and timing information related to encountering the entry. Instead of discarding all entries in the string translation table, a portion of the table can be discarded that has less important entries. The determination of whether to discard an entry as being less important can be based on the age indicator and the use indicator.

Abstract: A method and system for decompressing dynamic Huffman coded bit streams is disclosed.

Abstract: The present invention relates to a method for encoding a vector for transmission from a transmitter to a receiver, comprising a step of selecting a first vector in a predefined M-dimensional vector codebook, and at least one refinement step wherein an error vector between the random vector and the first vector is quantised by means of selecting a further vector from a further vector codebook with dimensionality reduced by one relative to the previous step, wherein the M-dimensional vector codebook and the further predefined vector codebook are known to both the transmitter and receiver.

Abstract: A method and an apparatus for encoding/decoding a digital signal are provided. First, a digital input signal is transformed into samples to remove redundant information among signals. Then, a lookup table corresponding to a characteristic of the input signal is selected among a plurality of lookup tables that indicate different numbers of bits allocated for each quantization unit depending on different characteristics of input signals, and the number of bits allocated for each quantization unit is acquired from the selected lookup table. Next, a distribution of samples within each quantization unit is divided into a predetermined number of sections, and the samples are linearly quantized using the allocated number of bits on a section-by-section basis. Thereafter, a bitstream comprised of frames is produced from the quantized samples and predetermined side information so that information about a frame length is stored in the end of frame.

Abstract: An input stream of characters is received. A new string entry is added to a data structure based on a current portion of the input stream if the current portion fails to match any of a plurality of existing string entries in the data structure. An additional string entry is added to the data structure based on a combination of the current portion and a previous portion of the input stream, wherein the additional string entry corresponds to a subset of the combination that fails to match any of the plurality of existing string entries in the data structure. An output stream is generated that includes literal values of the characters and index values corresponding to string entries in the data structure that match portions of the input stream.

Abstract: Provided is a bit stream processor using a reduced table lookup. The bit stream processor includes a bit stream exclusive register in a general purpose register in order to process data of a variable length effectively. Additionally, the bit stream processor an instruction of a table lookup method to which a prefix method is applied and a bit stream exclusive instruction in order to reduce an entire memory size.

Abstract: The present invention relates to a data compression and decompression system and method for lossless compressing digital data. In one preferred embodiment, the method for handling a data stream having a number of data objects comprising a step of performing a compressing process on a data object based on a compression based value to obtain at least one compressed data result, wherein the data object is considered as one numerical value. The device for handling the data stream comprising a compression means for compressing the data objects according to a compression base value. In another preferred embodiment, the method for handling data having a number of data objects comprising a step of performing a compressing process on a data object by obtaining a compression code for the data object according to a compression coding table, wherein the data object is considered as one numerical value.

Abstract: A method includes generating a first symbol string code indicative of one or more positions of a first symbol string within data to be compressed. A second symbol string code indicative of one or more positions of a second symbol string within the data is also generated. The first symbol string code and the second symbol string code are combined into a compressed data code.

Abstract: Simultaneously decoding one or more variable length code symbols including storing in a bit FIFO at least a portion of a variable length coded bit stream; storing a succession of extracted bit fields of predetermined bit length from the variable length coded bit stream in the bit FIFO; defining at least one inspection field in the extracted bit field, each inspection field identifying at least one variable length code symbol; and storing in look-up table a decode value uniquely addressable by each bit combination of the inspection field for reading out the one or more decode values corresponding to the unique address defined by the inspection field.

Abstract: An apparatus of a text image coding and decoding system includes a decomposer to decompose image data to extract a stripe image having one or more symbols, a matching unit to match the symbols with reference symbols of a symbol dictionary, and a dynamic symbol caching unit to generate one or more new symbols when the one or more new symbols of the symbols are not included in the symbol dictionary, and to remove the least used one or more of the reference symbols, so that the symbol dictionary is updated as an updated symbol dictionary with the one or more new symbols and without the least used one or more reference symbols.

Abstract: An image coding apparatus provides a run-length encoding unit RLE1 that subjects quantized coefficients which are obtained by quantizing frequency components of an image signal to a variable length coding process by using a run value Run that indicates the number of successive zero coefficients and a level value Lev that indicates a value of a non-zero coefficient following the zero coefficients. The run-length encoding unit RLE1 includes a reordering unit Lreodr for reordering level values Lev; a variable length coder LVLC for coding reordered level values ROLev by using a code table that is selected according to the value of a quantization parameter QP; a reordering unit Rreodr for reordering run values Run from high frequency component of the quantized coefficients to low frequency component; and a variable length coder RVLC for coding reordered run values RORun by using a code table that is selected according to the number of already-processed run values.

Abstract: The present invention relates to a method, computer program product and system for the compression of a probability table and the reconstruction of one or more probability elements using the compressed data and method. After determining a probability table that is to be compressed, the probability table is compressed using a first probability table compression method, wherein the probability table compression method creates a first compressed probability table. The first compressed probability table contains a plurality of probability elements. Further, the probability table is compressed using a second probability table compression method, wherein the probability table compression method creates a second compressed probability table. The second compressed probability table containing a plurality of probability elements.

Abstract: An improved LZW algorithm creates images at the time of compression, and obtains codes from those images at the time of decompression. At the time of compression, string (x) is obtained from the input stream and code (x) is retrieved from the LZW table for string (x). Assuming that a valid image (x) exists for the code, the system and method of the present disclosure may either: 1) write code (x) to the output stream, or 2) write image (x) to the output stream. The enhanced decompression algorithm of the present disclosure in one embodiment retrieves the code or image and restores the original stream. In addition, 1 bit is removed from code (x+1), and its information is indirectly stored in code (x)/image (x). Thereafter, when the decompression program encounters image (x) in the compressed stream, the decompression algorithm adds binary bit ‘1’ (or ‘0’) to the code (x+1), and adds binary bit ‘0’ (or ‘1’) to code (x+1) when it encounters code (x).

Abstract: A modulation mapper selects respective data blocks (DBs) hierarchically and sequentially in a descending order of a transmission power ratio control by a gain multiplexer; decides, from a resource element group, unused REs to which modulation symbols can be allocated; and decides the allocation of modulation symbols in the REs. A channel resource element mapper performs the mapping of modulation symbols of the respective DBs to the REs on the basis for the decided allocation. A complex-value transmission signal is generated from the group of REs.

Abstract: A two stage compression sub-system for clear channel data. The front stage of the compressing sub-system is an octet based repeat compressor (for example a flag compressor). The second stage is dictionary based compressor (for example Lempel-Ziv (LZ) or Huffmann). Data is compressed using several different techniques, and the technique that provides the best compression is used for each particular packet. For example, each packet can be: a) compress through both compression stages. b) compress through front stage flag compressor only c) compress through back stage dictionary compressor only d) not compressed through either stage (for highly incompressible data) After compression, each packet is provided with a header which specifies the exact method used to compress that packet. At the decoder, the packet header is interrogated to determine how the packet should be de-compressed and the appropriate de-compression is then used.

Abstract: An alpha image encoding and decoding scheme operable according two different modes is disclosed. In the encoding, an alpha image is decomposed into alpha image blocks (600) comprising image elements (610). The blocks (600) are compressed into block representations (700) according to one of the two compression modes. A block representation (700) comprises a color codeword (720), an intensity codeword (730), an alpha codeword (740) and a sequence (750) of image element associated indices indicative of one of the codewords (730, 740). The compression and decompression mode to use for a block (600) is determined based on the alpha codeword (740). In a high alpha-resolution mode, the index sequence (750) comprises alpha indices selecting one of the quantized alpha values (740A, 740B) of the alpha codeword (740) and intensity indices selecting an intensity modifier from a modifier set represented by the intensity codeword (730).

Abstract: A method of compressing instructions in a program may include extracting unique bit patterns from the instructions in the program and constructing a linear programming formulation or an integer programming formulation from the unique bit patterns, the instructions, and/or the size of a memory storage. The linear programming formulation or the integer programming formulation may be solved to produce a solution. The method may include compressing at least some of the instructions based on the solution by storing at least some of the unique bit patterns in a memory and placing corresponding indices to the memory in new compressed instructions.

Abstract: A terminal and a base station using an adaptive codebook for a polarization are provided. The terminal includes a rotation matrix generator to generate rotation matrices corresponding to a cross polarization discrimination value (XPD) of a transmission signal and a codebook generator to generate a codebook including result matrices that are generated based on pre-stored block diagonal matrices and the rotation matrices.

Abstract: The present invention aims to improve the compression ratio of an encoding method using a dictionary for such as LZ77, LZ78 or LZW for a sample string of an audio signal or an image signal. According to the present invention, the samples are aligned from the MSB side (73), and for an NC-th (NC is the character size, for example, eight bits) character C1(i), the samples are partitioned by starting from the MSB side, and a partitioned portion less than NC is added with dummy bits “0” until the size reaches NC to form a character (74). In this processing, lower bits in the samples can be ignored, mask data M1(i) is created in which digits in the ignorable bit and dummy bits for each C1(i) are set to “0” and effective bits (digits) are set to “1” (75), C1(i) is compared with each character D(j) in a character string for an index j in the dictionary using M1(i) to check whether or not there are matches for effective digits portions in C1(i). If there are matches, C1(i) is considered to match D(j).

Abstract: Each symbol in a sequence of codewords is expressed as a voltage level representative of a one or zero, and each codeword includes both levels. Sense amplifiers at the receiver compare all pairs of symbols in each codeword. Comparisons based upon like symbol values provide mid-range sensed voltages, and comparisons based upon disparate symbols values provide relatively higher or lower sensed voltages. Sampling the high and low voltages produces determinate sample values, whereas sampling the midrange voltage produces indeterminate sample values. Sampling the sensed voltages thus produces a mix of indeterminate and determinate values for each codeword. Codewords are selected such that the number and placement of the determinate sample values identify each codeword, and subsequent codewords are selected so the nodes supporting the high and low sensed voltages sampled to obtain the determinate sample values transitioned to their current voltage from either the intermediate voltage or did not transition.

Abstract: An encoding processing apparatus includes a first storing section for storing first encoded information and second encoded information, a second storing section for storing a table indicating association relation between the first encoded information and the second encoded information, an arithmetic section for calculating the second encoded information by reading the first encoded information stored in the first storing section and searching the table stored in the second storing section, a third storing section for storing by associating the first encoded information previously read from the first storing section and the second encoded information, a first control section for reading the second encoded information associated with the first encoded information from the third storing section, and a second control section for storing by associating the first encoded information with the second encoded information in the third storing section.

Abstract: A decoder for decoding a message using an arbitrary-side growing (ASG) Huffman tree including a plurality of codewords and symbols includes a database, a processing module, and a bit pattern matching module. The database stores a plurality of parameters corresponding to the ASG Huffman tree divided into several sub-trees according to a Hashemian cut operation and a bits-pattern-Xor (BPX) operation. The plurality of parameters include a bit pattern of a sub-tree of the ASG Huffman tree divided according to the BPX operation and a code length of the bit pattern. The processing module receives a bit stream corresponding to the ASG Huffman tree from an encoder. The bit pattern matching module selects bits from the bit stream according to the code length of the bit pattern and decodes the selected bits into a symbol in the ASG Huffman tree by comparing the selected bits with the bit pattern.

Abstract: A method for encoding an input file into an output file that is compressed so that the number of bits required to represent the output file is less than the number of bits of the input file. The encoding method includes the parsing of the input file into a series of data items, the data items having an order and collectively corresponding to the input file. The encoding method compares the series of data items against a static dictionary having at least mappings between terminal sequence pointers and representations of data items.

Abstract: A coding apparatus includes a fourth layer codebook that shows N number of codes indicating uniquely respective N integers that increment one by one; and first to third layer codebooks that show M number of codes indicating uniquely respective M integers that are a subset of the N integers, and codes a digital signal using any one of the first to fourth layer codebooks. The coding apparatus does not need to do rescaling even when switching one of the first to fourth layer codebooks into another of them.

Abstract: The described embodiments provide a system that encodes a sequence of integers using a variable-length compression technique. During operation, the system scans the sequence of integers and observes the sizes of the integers to determine a threshold value, K, from the observed sizes. For a given integer of length N bits, if N?K is greater than zero, the system generates a tag for the encoded integer comprising a sequence of N?K zeros followed by a one, and generates a set of remaining bits for the encoded integer as a sequence of the N?1 least-significant bits which make up the integer. Otherwise, the system generates a tag for the encoded integer as a single one, and generates a set of remaining bits for the encoded integer by padding the N bits which make up the integer with zeros so that the set of remaining bits is K bits in length.

Abstract: A digital signal encoding method and apparatus using a plurality of lookup tables. The method includes: preparing a plurality of lookup tables storing numbers of allocated bits for encoding frequency bands of an input signal according to a characteristic of the input signal in a predetermined number of addresses; dividing an input signal in the time domain into signals in predetermined frequency bands; calculating address values of the frequency bands; selecting one of the plurality of lookup tables according to the characteristic of the input signal; extracting numbers of allocated bits of addresses having the calculated address values from the selected lookup table with respect to the frequency bands and allocating the numbers of bits to the frequency bands; and generating a bitstream by quantizing the input signal according to the numbers of allocated bits.

Abstract: Provided is an H.264 Context Adaptive Variable Length Coding (CAVLC) decoding method based on an Application-Specific Instruction-set Processor (ASIP). The H.264 CAVLC decoding method includes determining a plurality of comparison bit strings on the basis of a table of a decoding coefficient, storing lengths of the comparison bit strings in a first register, storing code values of the comparison bit strings in a second register, comparing an input bit stream with the comparison bit strings based on the lengths and code values of the comparison bit strings, and determining value of the decoding coefficient according to a result of comparison between the input bit stream and the comparison bit strings. The method extracts a decoding coefficient using a register in an ASIP without accessing a memory and prevents a reduction in speed caused by memory access, thereby increasing the decoding speed of an H.264 decoder.

Abstract: A method and an apparatus for performing multiple-input multiple-output (MIMO) wireless communications are disclosed. A Node-B may receive an index to a pre-coding matrix in a single user MIMO (SU-MIMO) pre-coding codebook from wireless transmit/receive units (WTRUs) and adaptively perform one of SU-MIMO or multi-user MIMO (MU-MIMO) based on a predetermined criterion. Channel information for performing MU-MIMO may be obtained based on the pre-coding matrix of the SU-MIMO pre-coding codebook. A rank requested by the WTRU may be overridden if the unitary MU-MIMO codebook is a subset of the SU-MIMO pre-coding codebook. If not, a MU-MIMO pre-coding matrix with a largest correlation to the pre-coding matrix may be selected. A WTRU may send a pre-coding matrix for transmission to the WTRU along with a preferred interference matrix. A WTRU may send rank information and multiple right singular vectors for MU-MIMO.

Abstract: In one embodiment, a computing device may compress a text file for storage by dynamically creating an encoding table. A storage device of the computing device may store an encoding table populated by string characters from a text file. A processor of the computing device may encode the text file by replacing a string character in a text string of the text file with a table index position of that string character in the encoding table. The processor of the computing device may decode the text file by replacing the table index position with the string character at the table index position in the encoding table.

Abstract: A method for encoding data packets includes providing an encoding scheme for coding source data units into encoded data units; establishing a first look-ahead table for the source data units; providing a data packet including a first source data unit and a second source data unit; encoding the first source data unit to generate a first encoded data unit; indexing the first look-ahead table using the first source data unit to determine a balancing capability of the first encoded data unit for balancing a running disparity; and encoding the second source data unit to generate a second encoded data unit using the balancing capability of the first encoded data unit.

Abstract: A method of determining Unicode values corresponding to the text in digital documents includes: providing a digital document containing information related to the text in the document, the information including at least one set of data selected from the group consisting of: the numerical character code comprised by a single byte value or a sequence of multiple bytes, the glyph name corresponding to the character code for simple fonts, the code-to-Unicode mapping provided by a ToUnicode CMap, and font outline data embedded in the document; obtaining the information related to the text from the document; and determining the Unicode values corresponding to a specific code of a specific font on a per-glyph basis by executing a cascade of determination steps for each code separately, the cascade being executed in a predetermined sequence using different sources of information.

Abstract: The present invention relates to methods and apparatus for preventing power imbalance in a multiple input multiple output (MIMO) wireless precoding system. According to one aspect of the present invention, a codebook is constructed with a first subset of codewords that are constant modulus matrices, and a second subset of codewords that are non-constant modulus matrices. A mapping scheme is established between the first subset of codewords and the second subset of codewords. When a unit of user equipment feeds back a first codeword that is a non-constant modulus matrix, the Node-B may replace the first codewords with a second codeword that is selected from the first subset of codewords and that corresponds to the first codeword in accordance with the mapping scheme.

Abstract: An embodiment of the invention relates to a process for compression of executable code by a microprocessor, comprising steps consisting of decomposing the executable code into words; compressing each word of executable code, each compressed word of executable code comprising a part of predefined fixed length and a part of variable length whereof the length is defined by the part of fixed length; and combining all the parts of fixed length and all the parts of variable length of the words respectively into a block of parts of fixed length and in a block of parts of variable length, the respective positions of at least certain parts of variable length in the block of parts of variable length being saved in an addressing table.

Abstract: A method for encoding an input file into an output file that is compressed so that the number of bits required to represent the output file is less than the number of bits of the input file. The encoding method includes the parsing of the input file into a series of data items, the data items having an order and collectively corresponding to the input file. The encoding method compares the series of data items against a static dictionary having at least mappings between terminal sequence pointers and representations of data items.

Abstract: Systems and methods for optimizing system performance in variable length decoding systems are described. Embodiments are described in which decode tables are analyzed and elements of the tables sorted by probability of occurrence. Storage of elements can be determined by probability of occurrence and embodiments of the present invention can optimize system efficiency by storing most likely entries into fast-memory and least likely entries in slowest memory. In certain embodiments, a single large table is provided that cannot fit into decoder fast-memory. In some embodiments, individual elements can be optimized for storage in fast-memory by selecting more frequently occurring entries or groups of entries into decoder memory.

Abstract: An information retrieval architecture for performing a multi-dimensional search for a lookup value associated with a set of input values, the set of input values organized into one or more fields, the information retrieval architecture including a plurality of classification modules, each classification module storing the lookup values, each lookup value being associated with a set of input values; and a preprocessing module which receives a set of input values and selectively limits search of the plurality of classification modules to a subset of the classification modules.

Abstract: An apparatus for processing a signal and method thereof are disclosed. Data coding and entropy coding are performed with interconnection, and grouping is used to enhance coding efficiency. The present invention includes the steps of obtaining index information and data and entropy-decoding the index information and identifying an entropy table corresponding to the entropy-decoded index information and entropy-decoding the data using the identified entropy table.

Abstract: A system and method that process encoded data, wherein the encoded data is an encoded video stream. The encoded data may be buffered and the characteristics of the buffered encoded data may be determined and used to select a decoding table from a plurality of decoding tables. An encoded word may then be decoded using the selected decoding table. The output of the decoding table may be the corresponding decoded word and the size of the encoded word. The size of the encoded word may be utilized to shift the buffered encoded data. The decoded word may be sent to the appropriate portion of the system.

Abstract: A method of designing a look-up table of a finite-state encoder, applied to a finite-state encoder, comprises steps of: determining a bit length of a legal output codeword derived from the finite-state encoder and a restrict condition; collecting a plurality of legal output codeword satisfied the restrict condition; determining a bit length of an input codeword derived to the finite-state encoder according to the amount of the plurality of the legal output codeword; determining the amount of states in the finite-state encoder according to the bit length of the input codeword and the bit length of the legal output codeword; dividing the plurality of legal output codeword to a plurality of subset according to the amount of the states in the finite-state encoder and a specific mathematic equation; determining the amount of the legal output codewords in the plurality of subset; and, completing the look-up table through determining a relationship among a present state, a previous state, and a corresponding subset,

Abstract: A method for compressing information is provided that includes receiving a plurality of bits associated with a communications flow and determining if a match exists between the plurality of bits and a selected one of a plurality of bit patterns. A header bit segment that corresponds to the selected bit pattern may be communicated in place of the plurality of bits associated with the communications flow.

Abstract: A signal compression optimization system between a communication network and wireless user equipment advantageously selects an optimized decompressor when feasible for reduced content processing latency and otherwise selects a virtual machine decompressor, such as a Universal Decompressor Virtual Machine (UDVM) that interprets the received decompression bytecode. Since the UDVM is not optimized for any particular decompression algorithm and suffers by the requisite delays associated with analyzing each statement in the bytecode before execution, being able to avoid use of the UDVM whenever possible enhances user experience in presenting wirelessly received signaling messages or media content.

Abstract: Apparatus and methods are described for network content in dictionary-based (de)compression. Content for compression is parsed into discrete constructions, such as phrases, words, etc. The discrete constructions are passed to a searching engine to locate network information at one or more network locator identities, such as URI's (URL's, URNs) that correspond to the discrete constructions. Upon locating the network information, a dictionary of entries corresponding to the content is created. The content is encoded from the dictionary by indicating raw or relative offsets into the network information per each of the network locator identities. Decoding occurs anti-symmetrically to the encoding. In this manner, the vastness of network content, e.g., the world wide web, is leveraged to provide relevant dictionaries for de-/encoding. Searching for candidate dictionaries, scoring same and safeguarding against volatile dictionary content are other noteworthy features.

Abstract: Disclosed are a Multiple User Multiple Input Multiple Output (MU-MIMO) codebook design method, and a communication device using the codebook. A MU-MIMO codebook design method includes analyzing beam patterns of candidate vectors included in a predetermined candidate codebook, and eliminating at least one of the candidate vectors based on the beam patterns of the candidate vectors to generate the MU-MIMO codebook for a MU-MIMO system being comprised of the remaining vectors.

Abstract: A method of processing an encoded data stream comprises determining one or more data strings of interest; wherein the data string comprises a predetermined sequence of characters; encoding (3) the or each data string using the same encoding that was used to encode to the data stream; and searching (4) for the encoded data string in the encoded data stream.

Abstract: An adaptive compression technique which is an improvement to Lempel-Ziv (LZ) compression techniques, both as applied for purposes of reducing required storage space and for reducing the transmission time associated with transferring data from point to point. Pre-filled compression dictionaries are utilized to address the problem with prior Lempel-Ziv techniques in which the compression software starts with an empty compression dictionary, whereby little compression is achieved until the dictionary has been filled with sequences common in the data being compressed. In accordance with the invention, the compression dictionary is pre-filled, prior to the beginning of the data compression, with letter sequences, words and/or phrases frequent in the domain from which the data being compressed is drawn. The letter sequences, words, and/or phrases used in the pre-filled compression dictionary may be determined by statistically sampling text data from the same genre of text.

Abstract: In an adaptive transform coding system and/or and adaptive transform decoding system, coding efficiency in the case where a small number of quantized values having large absolute value are present, is improved. The adaptive transform coding system codes the small number of quantized values having large absolute values and other quantized values are coded separately. More particularly, the adaptive transform coding system includes a selector (6) discriminating the small number of quantized value having large absolute value from other quantized value, a pulse coding means for coding the small number of quantized values having large absolute values (8) and the pulse decoding means (16) for decoding the same, a coding means (7) for coding the quantized value other than those having large absolute values and a decoding means (15) decoding the same, and a synthesis means (18) for synthesizing the small number of quantized values having large absolute value and other quantized values.