JP2007295157A - Unit, method and program for data coding, and information recording medium having recorded data coding program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To realize Exp-Golomb coding, required for coding a syntax element in H. 264/AVC, at high speed with a small-scale integrated circuit. <P>SOLUTION: When the data length of an Exp-Golomb coded data does not exceed a bus bit length, the entire coded data are output to the bus at one time. When the data length of the coded data exceeds (2×K-1) times as long as the bus bit length while it does not exceed 2×K times, a first portion is output to the bus after being divided into one or a plurality of times, and a second portion and a third portion are output to the bus after divided into one or a plurality of times. When the data length of the coded data exceeds 2×K times as long as the bus bit length while it does not exceed (2×K+1) times, the first portion is output to the bus after being divided into one or a plurality of times, and a first data not exceeding the bus width, a second data and a third data exceeding an integer multiple of the bus width are output, and the third portion is output to the bus after divided into one or a plurality of times. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The invention according to the present patent application (hereinafter referred to as “the present invention”) is a data encoding device that encodes input data according to a predetermined method and outputs the encoded data to a bus having a predetermined bit length. The present invention relates to a data encoding method, a data encoding program, and an information recording medium on which the data encoding program is recorded.

  As a new video encoding method, H.264 is used. H.264 has been proposed. H. H.264 is a standard for video compression established by ITU-T Video Coding Experts Group (VCEG). ISO / IEC JTC 1 Moving Picture Experts Group (MPEG-4 Part 10 Advanced Coding) H. is integrated with the standard. H.264 / MPEG-4 AVC and H.264. Sometimes referred to as H.264 / AVC.

  In this method, in addition to the conventional compression technology, the compression performance is improved by using new technologies such as sub-macroblock division, intra prediction, deblocking filter, etc. It is said that the compression efficiency is more than doubled compared to the above.

  Such high compression performance is expected to be widely used from low bit rates for applications such as mobile phones to high bit rates of the HDTV class.

  H. In H.264 / AVC, as one of encoding methods for encoding syntax elements, exponential Golomb codes (Exponential Golomb codes), CAVLC (Context-based Adaptive VLC), and CABAC (on-CAC) -Based Adaptive Binary Arithmetic Coding) is specified.

  FIG. 6 shows a block configuration diagram of an Exp-Golomb encoding device 800, which is an example of a conventional Exp-Golomb encoding device that performs Exp-Golomb encoding.

  When syntax element data 810 to be encoded data is input to the Exp-Golomb encoding device 800, the syntax element data 810 is first input to the input control unit 801.

  The input control unit 801 determines a method for generating codeNum (code number) based on the data type of the syntax element data 810.

  There are three methods for generating codeNum: Unsigned, Signed, Mapped, and other methods. Here, three types of Unsigned, Signed, and Mapped are described. Further, the method for generating the codeNum is also determined based on the respective specifications, and is not directly related to the essence of the present invention, and thus detailed description thereof is omitted.

  When the input control unit 801 determines a method for generating codeNum (code number) based on the data type of the syntax element data 810, the syntax element data 810 is sent to the codeNum generation unit 802. The codeNum generation unit is divided into portions for generating codeNum in the respective methods of Unsigned, Signed, and Mapped, and these are referred to as an unsigned unit 821, a signed unit 822, and a mapped unit 823. Each syntax element data 810 is sent to a portion for generating codeNum in each method corresponding to the determined method.

  When the codeNum generation unit 802 generates codeNum, the codeNum is sent to the code generation unit 803. The code generation unit 803 is the center for executing Exp-Golomb encoding.

  The Exp-Golomb encoded data is once sent to the intermediate buffer 804 and accumulated, and sent to the output signal generation unit 805 as necessary and in synchronization with various timings. The data is converted into an output format and output to the data bus 806.

  The Exp-Golomb code is a kind of lossless compression code, and the encoded data is composed of three parts: a leading zero part, an INFO part, and a separating part that separates the leading zero part and the INFO part. .

  When a logarithmic value in which 2 of a value obtained by adding 1 to the codeNum is low is obtained and the integer part is M, the leading zero part is composed of M “0” s.

  The separation unit is always composed of one “1”.

  The INFO section is composed of a bit string excluding the most significant 1 when a value obtained by adding 1 to the codeNum is expressed in binary. This can also be referred to as a bit string when a value obtained by subtracting 2 to the power of M from a value obtained by adding 1 to codeNum is expressed by an M-digit binary number.

  In this way, it can be seen that the leading zero part and the INFO part always have the same bit length.

  As an example, it is assumed that the value “21” is input as syntax element data 810 whose syntax element type is “Unsigned” (unsigned integer), and this syntax element data 810 is Exp-Golob encoded. Is shown in FIG.

  Since the syntax element data 810 has a data type of “Unsigned”, codeNum is “21” which is the same as the input value.

  Since the logarithmic value in which 2 of 1 obtained by adding 1 to 21 is low and the integer part is taken out is “4”, the leading zero part of the Exp-Golomb encoded data is 4 digits of 0, That is, “0000”.

  The value obtained by adding 1 to codeNum is “22”, and this value is expressed as “10110” in binary, so that the INFO part is “0110” excluding the highest-order 1.

  Also, subtracting the fourth power of 2 from 22 is 6, and when this is expressed by a 4-digit binary number, the same "0110" is obtained.

  Therefore, Exp-Golomb encoded data is “000010110”.

  Assuming that exponential Golomb encoding is performed on numerical data having a maximum of 32 bits in this way, Exp-Golomb encoded data may have a maximum of 65 bits.

  Therefore, the buffer size of the intermediate buffer 804 needs to be 65 bits. Assuming that the bus width of the data bus 806 is 32 bits, and assuming that 65-bit Exp-Golomb encoded data is output in accordance with the bus width of 32 bits, the output signal generation unit 805 outputs 65 bits. Shift registers are required.

An example of outputting Golomb encoded data in this way is described in Patent Document 1 below.
JP 2004-104159 A

  However, in the above conventional Exp-Golomb encoding device, an intermediate buffer for 65 bits, a shift register, etc. are required to encode 32-bit input data to Exp-Golomb and output it to a 32-bit bus. In order to make an integrated circuit, the circuit configuration has to be complicated and large-scale, and cannot be realized as a simple and small-scale integrated circuit.

  The present invention has been made to solve the above-described problems. It is possible to implement Exp-Golomb encoding, which is necessary for encoding syntax elements in H.264 / AVC, with a simple and small-scale integrated circuit.

  Therefore, specifically, the following configuration is adopted.

  In the data encoding device according to claim 1, an encoding unit that encodes original data, a data length determination unit that determines a data length of data encoded by the encoding unit, and a data length determination unit A dividing unit that divides the encoded data based on the determination result into a predetermined bit length; and an output unit that outputs the divided data to a bus.

  With such a configuration, H.P. It is possible to realize encoding required for encoding syntax elements in H.264 / AVC and the like with a simple and small-scale integrated circuit, and to output the bus efficiently and at high speed.

  In the data encoding device according to claim 2, the encoded data includes first sub data, second sub data, and third sub data, and the first sub data and the third sub data. The data length is the same bit length.

  With such a configuration, in particular, the encoded data includes the first sub data, the second sub data, and the third sub data, and the data lengths of the first sub data and the third sub data are In the case where the bit lengths are the same, encoding can be more effectively realized by a simple and small-scale integrated circuit, and the bus can be output more efficiently and at high speed.

  4. The data encoding apparatus according to claim 3, wherein when the data length of the encoded data does not exceed a predetermined bit length, all of the encoded data is output to the bus at one time, and K is set to 1 When the data length of the encoded data exceeds “2 × K−1” times the predetermined bit length and does not exceed “2 × K” times, the first sub-data is set to 1. Divided into multiple or multiple times and output to the bus, then the second sub-data and the third sub-data are divided into one or more times and output to the bus, and the data length of the encoded data is When the predetermined bit length exceeds “2 × K” times and does not exceed “2 × K + 1” times, the first sub-data is divided into one time or a plurality of times and outputted to the bus. Less than a predetermined bit length and an integral multiple of the predetermined bit length of the second and third sub-data The part that exceeds after outputting to the bus, by dividing the third sub data to one or more times and output to the bus.

  With such a configuration, it is possible to realize encoding with a simple and small-scale integrated circuit using a more specific configuration, and to perform bus output efficiently and at high speed.

  5. The data encoding device according to claim 4, wherein the encoding is exponential Golomb encoding, the first sub-data is a leading zero part, the third sub-data is an INFO part, and the second sub-data Is a separation part of the leading zero part and the INFO part.

  With such a configuration, even when the encoding is Golomb encoding, encoding is more effectively realized by a simple and small-scale integrated circuit, and Golomb-encoded data is output to the bus more efficiently and at high speed. It becomes possible to do.

  In the data encoding device according to claim 5, the predetermined bit length is equal to the bit length of the bus.

  With such a configuration, it is possible to divide the encoded data in accordance with a given predetermined bus width and to output the bus more efficiently and at high speed.

  According to a sixth aspect of the present invention, there is provided a data encoding device comprising a register for storing a predetermined bit length.

  With such a configuration, the bit length to be divided can be stored, the encoded data can be divided efficiently and at high speed using the stored bit length, and the bus can be output more efficiently and at high speed. At the same time, the bit length to be divided can be flexibly changed as necessary.

  In the data encoding method according to claim 7, an encoding procedure for encoding original data, a data length determination procedure for determining a data length of data encoded by the encoding procedure, and a data length determination procedure A division procedure for dividing the encoded data based on the determination result into a predetermined bit length and an output procedure for outputting the divided data to the bus are included.

  With such a configuration, H.P. H.264 / AVC and the like can realize encoding required for encoding syntax elements and the like with a simple and small-scale integrated circuit, and can execute an efficient and high-speed bus output method. Become.

  The data encoding program according to claim 8 is a program that can be executed by a computer, an encoding step that encodes original data, and data that determines a data length of data encoded by the encoding step A length determining step, a dividing step of dividing the encoded data based on the determination result of the data length determining step into a predetermined bit length, and an output step of outputting the divided data to the bus.

  With such a configuration, the above-described data encoding device and data encoding method can be specifically realized and executed using a computer and a program executed by the computer.

  A ninth aspect of the present invention is an information recording medium readable by a computer, in which the data encoding program according to the eighth aspect is recorded.

  With such a configuration, the data encoding program of the present invention can be recorded on a recording medium, stored, used, read out by a computer, and executed.

  In the present invention, H.P. Exp-Golomb encoding, which is necessary for encoding syntax elements in H.264 / AVC, can be realized efficiently and at high speed with a simple and small-scale integrated circuit.

(Embodiment)
Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is an overall schematic diagram of an Exp-Golomb encoding apparatus, which is one embodiment of the present invention. The Exp-Golomb encoding device 100 mainly includes an Exp-Golomb code generation unit 102, an output control unit 104, and a memory unit 103. Exp-Golomb encoding apparatus 100 includes other parts, but is omitted because it is not directly related to the essence of the present invention.

  The Exp-Golomb encoding apparatus 100 is connected to the processor 201 and receives data 211 and a command 212 from the processor 201.

  For example, if this Exp-Golomb encoding device 100 is used in a mobile phone, the processor 201 may be a CPU that controls the entire mobile phone, or may be a CPU that controls image processing of the mobile phone. Or, it may be a CPU that controls communication of a mobile phone. Alternatively, for example, if the Exp-Golomb encoding device 100 is used in a television receiver, the processor 201 may be a CPU that controls the entire television receiver, or controls image processing of the television receiver. It may be a CPU, or it may be a CPU that controls broadcast radio wave reception of a television receiver. This Exp-Golomb encoding device 100 may be used in, for example, the mobile phone and the television receiver as described above, or may be used in a video recording / reproducing device such as a DVD (Digital Versatile Disc). It may be used for PCs (personal computers), PDAs (personal digital assistants), etc., and may be used for various office electrical and electronic devices such as fax machines and copiers. It may be used in various household electrical and electronic devices, and may be used in all other electrical and electronic devices. The description in this specification is only a few examples, and the present invention is not limited to the examples described herein (the same applies to all descriptions in this specification). .

  The Exp-Golomb code may be used in all such electric / electronic devices, and the Exp-Golomb encoding device 100 may be used as long as the Exp-Golomb code may be used. There is.

  Further, the processor 201 is not necessarily an independent CPU or MPU, and may be a CPU core that is a part of one LSI, a gate array, or the like. Alternatively, a DSP or the like used for a specific purpose may be used.

  The Exp-Golomb encoding apparatus 100 receives data 211 and a command 212 from the processor 201. This data 211 is data to be subjected to Exp-Golomb encoding and other data, and the command 212 is, for example, a command “Convert data 211 to Exp-Golomb code”, and other various commands. Sometimes.

  The Exp-Golomb encoding apparatus 100 uses the Exp-Golomb code generation unit 102, the output control unit 104, and the memory unit 103 in the Exp-Golomb encoding of the received data 211 in accordance with the command 212. To do. Then, the Exp-Golomb encoded data is output to the data bus 301. It is an object of the present invention to realize these processes with a simple and small-scale integrated circuit, and is a feature of the present invention.

  In order to describe the Exp-Golomb encoding device 100 of the present embodiment in more detail, FIG. 2 shows a block configuration diagram of the Exp-Golomb encoding device 100 of the present embodiment.

  As described above, the Exp-Golomb encoding apparatus 100 receives the data 211 and the command 212. This data 211 is, for example, H.264. H.264 / AVC syntax element data.

  Also, as described above, this Exp-Golomb encoding device 100 mainly includes an Exp-Golomb code generation unit 102, an output control unit 104, and a memory unit 103. Exp-Golomb encoding apparatus 100 includes other parts, but is omitted because it is not directly related to the essence of the present invention.

  The Exp-Golomb code generation unit 102 plays a central role in executing Exp-Golomb coding. The Exp-Golomb code generation unit 102 mainly includes an input control unit 121, a codeNum generation unit 122, and a code generation unit 123. The Exp-Golomb code generation unit 102 includes other parts, but is omitted because it is not directly related to the essence of the present invention.

  The Exp-Golomb code generation unit 102 first receives the data 211 and the command 212 input to the Exp-Golomb encoding device 100, and in particular, the input control unit 121 among them. The input control unit 121 receives the input data 211 and the command 212, and determines a method for generating codeNum (code number) based on the data type of the data 211. As a method for generating this codeNum, there are Unsigned, Signed, and Mapped. There are other methods for generating codeNum, but they are not directly related to the essence of the present invention, and will be omitted.

  Then, in accordance with the input command 212, the input control unit 121 sends the code 211 to the codeNum generation unit 122 as a preparatory preparation for encoding the data 211 with Exp-Golomb encoding. At this time, based on the determined method for generating codeNum, the CodeNum for generating codeNum according to the respective data types of the Unsigned unit 124, the Signed unit 125, and the Mapped unit 126 included in the codeNum generation unit 122. The data 211 is sent out by dividing into parts.

  In the codeNum generation unit 122, the Unsigned unit 124, the Signed unit 125, and the Mapped unit 126 receive the received data 211 and generate codeNum according to the respective methods. codeNum is a non-negative integer value, and codeNum is data to be subjected to Exp-Golomb encoding. The codeNum generation method is determined by the type of syntax element that is the subject of each Exp-Golomb encoding, and there are Unsigned, Signed, Mapped, and others as described above, but directly with the essence of the present invention. Since it is not related, further detailed explanation is omitted. Further, in the description herein, codeNum is assumed to be 32-bit data. As described above, the present invention is not limited to this example.

  In this way, the codeNum generation unit 122 sends the codeNum generated according to each data type of Unsigned, Signed, and Mapped to the code generation unit 123. Then, the code generation unit 123 performs encoding from codeNum to Exp-Golomb code. Similar to the code generation unit 803 in the conventional technology described above, the code generation unit 123 is the center for executing Exp-Golomb encoding.

  The procedure of Exp-Golomb encoding in the code generation unit 123 and the data subjected to Exp-Golomb encoding are the same as those of the code generation unit 803 in the above-described conventional technology. Exp-Golomb encoded data is composed of three parts: a leading zero part, an INFO part, and a separating part that separates the leading zero part and the INFO part. When a logarithmic value in which 2 of a value obtained by adding 1 to the codeNum is low is obtained and the integer part is M, the leading zero part is composed of M “0” s. The separation unit is always composed of one “1”. The INFO section is composed of a bit string excluding the most significant 1 when a value obtained by adding 1 to the codeNum is expressed in binary. This can also be referred to as a bit string when a value obtained by subtracting 2 to the power of M from a value obtained by adding 1 to codeNum is expressed by an M-digit binary number. Assuming that exponential Golomb encoding is performed on numerical data having a maximum of 32 bits in this way, Exp-Golomb encoded data may have a maximum of 65 bits.

  However, in the present embodiment, not all of the Exp-Golomb code that may be 65 bits at the maximum is generated, but only the INFO part is generated, and only this INFO part is sent to the intermediate buffer 131 of the memory part 103. .

  Then, the data length of the Exp-Golomb code having a maximum length of 65 bits is sent to the division determination unit 141 of the output control unit 104.

  The division determination unit 141 of the output control unit 104 determines the division method of the Exp-Golomb encoded data while looking at the data length of the Exp-Golomb encoded data, and the intermediate buffer 131 of the memory unit 103 The output signal generator 132 is controlled to output Exp-Golomb encoded data to the 32-bit data bus 301.

  The division size at this time can be set to 32 bits, which is the bit length of the data bus 301, for example. As a result, Exp-Golomb encoded data can be divided in accordance with the bus width of the data bus, and the bus can be output more efficiently and at high speed. However, matching the division size to the bit length of the data bus 301 is one example, and the present invention is not limited to this example.

  At this time, this division size can also be stored and stored in, for example, a register held by the division determination unit 141. As a result, the bit length stored and saved can be read out, and the encoded data can be divided efficiently and at high speed, and the bus output can be performed more efficiently and at the same time. The bit length to be divided can be flexibly changed as necessary.

  Next, when the data length of the data encoded by Exp-Golomb does not exceed the bit length of the bus, the data length of the data encoded by Exp-Golomb encodes “2 × K−1” of the bit length of the bus. ”Times and not exceeding“ 2 × K ”times, and the data length of Exp-Golomb encoded data exceeds“ 2 × K ”times the bit length of the bus, and“ 2 × K + 1 ”. When the division method is not exceeded, the division determination unit 141 of the output control unit 104 determines the division method, the intermediate buffer 131 and the output signal generation unit 132 of the memory unit 103 are controlled, and the Exp− A control method for outputting Golomb encoded data to the 32-bit data bus 301 will be described.

  First, a case where the data length of Exp-Golomb encoded data does not exceed the bit length of the data bus 301 will be described. In other words, this is when the sum of the leading zero part, separation part, and INFO part of the Exp-Golomb code does not exceed 32 which is the bit length of the data bus 301, and the data length of the INFO part does not exceed 15 bits. It's time.

  An outline of the data structure and output method at this time is shown in FIG.

  At this time, the division determination unit 141 determines that all Exp-Golomb codes can be output to the data bus 301 at one time and all Exp-Golomb codes are output to the data bus 301 at one time. Then, the intermediate buffer 131 is instructed to send all the INFO data to the output signal generator 132 at a time. As a result, the INFO section having a bit length N (where N is an integer greater than 0 and equal to or less than 15) is sent to the output signal generation section 132 as it is.

  The output signal generator 132 receives all the INFO data sent from the intermediate buffer 131, and adds a leading zero part having the same bit length N and a separating part consisting of 1 bit “1”. Then, an instruction is issued so as to generate an Exp-Golomb code and output the Exp-Golomb code to the data bus 301 at a time.

  In accordance with this instruction, the output signal generation unit 132 receives the data of all the INFO units sent from the intermediate buffer 131, and outputs a leading zero part having the same bit length N and a separation unit consisting of 1 bit “1”. In addition, an Exp-Golomb code is generated, and the Exp-Golomb code is output to the data bus 301 at a time. As described above, the data length of the Exp-Golomb encoded data in this case is 2 × N + 1 bit length, and this bit length is 32 or less, so that the Exp-Golomb encoded data is data at a time. It is possible to output to the bus 301.

  Next, when K is a natural number, the data length of the data encoded by Exp-Golomb exceeds “2 × K−1” times the bit length 32 of the data bus 301, and increases by “2 × K” times. Explain when not exceeding.

  As described above, in the present embodiment, the maximum bit length of codeNum is assumed to be 32 bits, and the maximum bit length of Exp-Golomb encoded data is assumed to be 65 bits. That is, the data length of the Exp-Golomb encoded data cannot exceed (2 × 2−1 = 3) times the bit length 32 of the data bus 301. Therefore, in this embodiment, this condition corresponds to K = 1, that is, the data length of the data encoded with Exp-Golob is 32 (2 × 1-1 = 1) with the bit length 32 of the data bus 301. ) Times greater than (2 × 1 = 2) times.

  Specifically, this is when the data length of Exp-Golomb encoded data is 32 bits or more and 63 bits or less, in other words, when the data length of the INFO section is 16 bits or more and 31 bits or less. To do.

  FIG. 4 shows an outline of the data structure and output method at this time.

  At this time, the division determination unit 141 first causes the intermediate buffer 131 to wait to send the INFO unit to the output signal generation unit 132.

  While this intermediate buffer 131 is waiting for the INFO section to be sent to the output signal generation section 132, the output signal generation section 132 is informed of the bit length N of the INFO section (where N is 16 or more and 31 It is instructed to divide the leading zero part having a bit length corresponding to (the following natural number) into the data bus 302 by dividing it into several times. In the present embodiment, as described above, it is assumed that the bit length N of the INFO part is a natural number of 16 or more and 31 or less, so that the leading zero part can be output to the data bus 302 at one time. .

  When the output of the leading zero part is finished, next, the division determination unit 141 instructs the intermediate buffer 131 to send the INFO part to the output signal generation part 132. In accordance with this instruction, the intermediate buffer 131 sends the INFO unit to the output signal generation unit 132. As described above, the data length N of the INFO portion is 16 bits or more and 31 bits or less.

  Then, the division determination unit 141 receives the INFO unit sent from the intermediate buffer 131 to the output signal generation unit 132, adds a separation unit in front of the INFO unit, and divides this into several times. Instruct to output to the data bus 302.

  In the present embodiment, as described above, since the data length N of the INFO part is 16 bits or more and 31 bits or less, the data length of the part to which the separation part is added before the INFO part is 17 bits or more and 32 bits or less. A portion obtained by adding a separation unit to the INFO unit can be output to the data bus 302 at one time.

  Finally, when K is a natural number, the data length of the data encoded by Exp-Golomb exceeds “2 × K” times the bit length 32 of the data bus 301 and does not exceed “2 × K + 1” times. Explain about time.

  As described above, in the present embodiment, the maximum bit length of codeNum is assumed to be 32 bits, and the maximum bit length of Exp-Golomb encoded data is assumed to be 65 bits. In other words, the data length of the Exp-Golomb encoded data cannot exceed (2 × 2 = 4) times the bit length 32 of the data bus 301. Therefore, in this embodiment, this condition corresponds to K = 1, that is, the data length of the data encoded by Exp-Golomb is (2 × 1 = 2) times the bit length 32 of the data bus 301. Only when (2 × 1 + 1 = 3) times is not exceeded.

  Specifically, this is when the data length of Exp-Golomb encoded data is 64 bits or more and 95 bits or less, in other words, when the data length of the INFO section is 32 bits or more and 47 bits or less. To do. In addition, as described above, in this embodiment, the maximum bit length of codeNum is assumed to be 32 bits, and the maximum data length of Exp-Golomb encoded data is assumed to be 65 bits. Only when the data length of the INFO portion is 32 bits and the data length of the data encoded with Exp-Golomb is 65 bits.

  An outline of the data structure and output method at this time is shown in FIG.

  Also at this time, the division determination unit 141 first causes the intermediate buffer 131 to wait to send the INFO unit to the output signal generation unit 132.

  While this intermediate buffer 131 is waiting for the INFO section to be sent to the output signal generation section 132, the output signal generation section 132 is preceded by a leading zero section having a bit length corresponding to the bit length N of the INFO section. Is output to the data bus 302, and when there is data corresponding to a fraction less than the bus width in the leading zero part, it is instructed to leave it as it is. In the present embodiment, as described above, it is assumed that the bit length N of the INFO part is only 32. Therefore, all of the leading zero part can be output to the data bus 302 at one time, and the leading zero can be output. No data corresponding to a fraction less than the bus width remains in the part.

  When the output of the leading zero part is finished, next, the division determination unit 141 instructs the intermediate buffer 131 to send the INFO part to the output signal generation part 132. In accordance with this instruction, the intermediate buffer 131 sends the INFO unit to the output signal generation unit 132. As described above, the data length N of this INFO part is also only 32 bits.

  Then, the division determination unit 141 receives the INFO unit sent from the intermediate buffer 131 to the output signal generation unit 132, and when there is a remaining data corresponding to a fraction less than the bus width of the leading zero part, What is the number of times when the remainder of the leading zero part and the 1-bit “1” that is the separation part are added after that, and then the part exceeding the integer multiple of the data bus width 32 of the INFO part is added? An instruction is given to divide the data into data and output it to the data bus 302.

  In the present embodiment, as described above, there is no remaining portion of data corresponding to a fraction less than the bus width of the leading zero part. Further, since the data length N of the INFO part is 32 bits, there is no part exceeding an integral multiple of the bus width 32 of the data bus 302 of the INFO part.

  Therefore, next, only 1 bit “1” of the separation unit is output to the data bus 302.

  Next, an instruction is given to divide a portion of the data bus 302 that is an integral multiple of the bus width 32, which is the remaining portion of the INFO portion, into a size corresponding to the bus width 32 and output it to the data bus 302.

  In accordance with this instruction, the output signal generation unit 132 outputs the remaining part of the INFO unit, that is, the entire INFO unit received from the intermediate buffer 131 to the data bus 302 while dividing the INFO unit by the bus width 32 of the data bus 302. . In this embodiment, since it is assumed that the data length of the INFO section is 32 bits, all of the INFO sections received from the intermediate buffer 131 can be output to the data bus 302 once.

  In this way, the maximum bit length of Exp-Golomb encoded data is 65 bits, but there is no need to have a 65-bit buffer memory or shift register inside the memory unit 103, and Exp- It is possible to reduce the circuit scale of the entire Golomb encoding device.

  In addition, since data can be output in units of 32 bits to the data bus 302, the processing in the subsequent circuit can also use the same processing as other encoding results.

  In the present embodiment, the description has been made assuming that the maximum length of data in the INFO section is 32 bits and the maximum length of Exp-Golomb encoded data is 65 bits. However, a data size larger than this data size is described. At this time, the above description can be applied in the same manner, and thus detailed description is omitted.

  The Golomb encoding apparatus 100 according to the present embodiment executes one embodiment of the Golomb encoding method according to the present invention. The Golomb encoding apparatus 100 includes a microcomputer (not shown). It is an embodiment of the Golomb encoding program of the present invention that controls the computer to execute Golomb encoding. Also, the program that is one embodiment of the Golomb encoding program of the present invention is recorded on a computer-readable information recording medium, and this information recording medium is an information recording medium that records the Golomb encoding program of the present invention. This is one embodiment.

  In the present invention, H.P. Exp-Golomb encoding required for encoding syntax elements in H.264 / AVC can be realized by a simple and small-scale integrated circuit, and its industrial applicability is extremely large. .

1 is an overall schematic diagram of an Exp-Golomb encoding device that is one embodiment of the present invention. The block block diagram of the Exp-Golomb encoding apparatus of embodiment of this invention The figure which shows the outline | summary of the data structure and output method when the data length of an INFO part does not exceed 15 bits The figure which shows the outline of the data structure and output method when the data length of the INFO section is 16 bits or more and 31 bits or less The figure which shows the outline of the data structure and output method when the data length of an INFO part is 32 bits Block diagram of a conventional Exp-Golomb encoding device The figure which shows a mode that syntax element data is Exp-Golomb encoding.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100,800 Exp-Golomb encoding apparatus 102 Exp-Golomb code generation part 103 Memory part 104 Output control part 121,801 Input control part 122,802 codeNum generation part 123,803 Code generation part 124,821 Unsigned part 125,822 Signed Unit 126, 823 Mapped unit 131, 804 Intermediate buffer 132, 805 Output signal generation unit 141 Division determination unit 201 Processor 211 Data 212 Command 301, 806 Data bus 810 Syntax element data

Claims (9)

An encoding unit for encoding the original data;
A data length determination unit for determining a data length of data encoded by the encoding unit;
A division unit that divides the encoded data into a predetermined bit length based on a determination result of the data length determination unit;
An output unit for outputting the divided data to a bus;
A data encoding device. The encoded data is composed of first sub data, second sub data, and third sub data,
The first sub-data and the third sub-data have the same bit length.
The data encoding apparatus according to claim 1. When the data length of the encoded data does not exceed the predetermined bit length, all of the encoded data is output to the bus at a time,
When K is a natural number,
When the data length of the encoded data exceeds “2 × K−1” times the predetermined bit length and does not exceed “2 × K” times, the first sub-data is set once or a plurality of times. Dividing the output into the times and outputting to the bus, then dividing the second sub-data and the third sub-data into one or more times and outputting to the bus,
When the data length of the encoded data exceeds “2 × K” times the predetermined bit length and does not exceed “2 × K + 1” times, the first sub-data is set once or a plurality of times. Dividing and outputting to the bus, the portion of the first sub data less than the predetermined bit length and an integer multiple of the predetermined bit length of the second sub data and the third sub data After outputting the portion to the bus, the third sub-data is divided into one or more times and output to the bus.
The data encoding device according to claim 2. The encoding is exponential Golomb encoding;
The first sub-data is a leading zero part;
The third sub-data is an INFO section;
The second sub-data is a separation part between the leading zero part and the INFO part.
The data encoding device according to claim 3. The predetermined bit length is equal to the bit length of the bus;
The data encoding device according to any one of claims 1 to 4. A register for storing the predetermined bit length;
The data encoding device according to any one of claims 1 to 5. An encoding procedure for encoding the original data;
A data length determination procedure for determining the data length of the data encoded by the encoding procedure;
A division procedure for dividing the encoded data into a predetermined bit length based on a determination result of the data length determination procedure;
An output procedure for outputting the divided data to a bus;
A data encoding method comprising: A program executable by a computer,
An encoding step for encoding the original data;
A data length determining step of determining a data length of the data encoded by the encoding step;
A division step of dividing the encoded data into a predetermined bit length based on a determination result of the data length determination step;
An output step of outputting the divided data to a bus;
A data encoding program. An information recording medium readable by a computer,
An information recording medium on which the data encoding program according to claim 8 is recorded.

Images