JP2010109912A - Image encoder and image encoding method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the accuracy of code amount control by reflecting the binarized symbol length of n-th MB on the quantization parameter of the (n+1)th MB in an image encoder having an MB pipeline configuration. <P>SOLUTION: The image encoder includes: first binarized code amount calculating means for calculating the binarized symbol length of a difference motion vector from a motion vector signal output from motion vector detecting means; third binarized code amount calculating means for calculating the binarized symbol length of an orthogonal transformation coefficient output from the quantizing means; and second binarized code amount calculating means for calculating the binarized symbol length of a signal excluding the difference motion vector and the orthogonal transformation coefficient that constitute a binarized symbol column, from macro block information output from the motion vector detecting means. The code amount control means determines a quantization parameter based on the binarized symbol lengths calculated in the first to third binarized code amount calculating means, and outputs the quantization parameter to the quantizing means. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an image encoding device and an image encoding method, and more particularly to a technique suitable for use in an image encoding device having a macroblock (MB) pipeline configuration.

  In recent years, with advances in digital signal processing, higher integration of LSIs, higher communication speeds, etc., a large amount of digital information such as moving images, still images, and voices has been encoded with high efficiency, and recorded on recording media and communication media. It has become possible to perform transmission. By applying such technology, digital high-definition broadcasting and terrestrial digital broadcasting have been started, and high performance of digital video cameras and the like has been advanced, and image quality has been improved. In recent years, H.264 has been established by ITU-T as an image coding standard. H.264 (Non-Patent Document 1) and the like are attracting attention.

  H. These image encoding processes including H.264 divide an original image into predetermined regions called blocks, and perform processing such as motion compensation prediction, orthogonal transform processing, and quantization processing in units of the divided blocks. Is.

  In many cases, the code length of a stream generated by such an image encoding device varies greatly depending on the spatial frequency characteristics, which are characteristics of the image itself, the scene, and the quantization scale value. Code amount control (rate control) is an important technique for obtaining a decoded image with good image quality in realizing an image encoding apparatus having such encoding characteristics.

  In the code amount control, quantization control is performed in order to bring the generated code amount of a picture close to the target code amount set for each picture. A quantization parameter is controlled for each macro block (hereinafter referred to as MB) having 16 × 16 pixels as one unit, and a generated code length is controlled. For this code amount control, it is necessary to know the generated code length of the MB.

  H. H.264 realizes higher-efficiency encoding than MPEG, which is a conventional image encoding standard. H. There are various reasons why H.264 achieves high-efficiency encoding, one of which is due to entropy encoding. In particular, as a method for realizing highly efficient entropy coding, H.264 is used. In H.264, CABAC can be used as an entropy encoding method. CABAC is Context-based Adaptive Binary Arithmetic Coding.

In CABAC, entropy coding is performed using binary arithmetic coding. For example, Patent Document 1 is known as a prior art relating to binary arithmetic coding.
On the other hand, in an image encoding device mounted on hardware, it is common to increase the speed by constructing a pipeline in units of macroblocks (MB) in order to encode a high-definition image within a certain processing time. It is. MB is one unit of encoding processing.

The MB pipeline will be described with reference to FIG.
In FIG. 4, MB1 to MB3 on the vertical axis are MB1 at the upper left corner of the image, MB2 is right next to MB1, MB3 is right next to MB2, and encoding is in the order MB1 to MB2 and MB3. Done. The stage indicates a motion vector detection stage, an orthogonal transform quantization stage that performs orthogonal transform and quantization, and a binarization stage that performs binarization of CABAC.

  In the motion vector detection stage, a differential motion vector is calculated. In parallel, the macro block partition and prediction mode are determined. Since the contents such as the difference motion vector and the macroblock partition are described in Non-Patent Document 1, detailed description thereof is omitted here.

  In the next orthogonal transform quantization stage, orthogonal transform and quantization are performed. The orthogonal transform coefficient obtained in this stage, the differential motion vector obtained in the motion vector detection stage, the prediction mode, etc. are binarized in the next binarization stage. At the end of the binarization stage (time t1), the binarized symbol length of MB1 is obtained. The binarized symbol length thus obtained is input to a code amount control unit (not shown) at time t1. The code amount control unit determines the quantization parameter of the next MB based on the input binary symbol length.

ITU-T Recommendation H. H.264 JP 2007-124122 A

  However, the following inconvenience occurs in the image encoding apparatus having the MB pipeline configuration. That is, the code amount control unit uses the n + 1th (MB2 in FIG. 4) and subsequent MBs based on the binarized symbol length (binarized symbol length × α) of the nth (MB1 in FIG. 4) MB. Even if the quantization parameter is calculated, the quantization process of the (n + 1) th MB is not in time.

  Referring to FIG. 4, even if the encoding control unit calculates the quantization parameter at the same time t1 based on the binarized symbol length calculated at time t1, and outputs it to the quantization unit, the following The encoding of MB2 has already started and the quantization has also been completed. For this reason, the quantization parameter calculated based on the binarized symbol length can be used only for the quantization of the MB after 2 MB, so that there is a problem that the accuracy of the code amount control is lowered.

  In view of the above-described problems, the present invention allows an image encoding apparatus having an MB pipeline configuration to reflect the binarized symbol length of the nth MB in the quantization parameter for the (n + 1) th MB, It is an object of the present invention to improve the accuracy of code amount control.

  An image coding apparatus according to the present invention is an image coding apparatus comprising quantization means, code amount control means, binarization means, and binary arithmetic coding means, and is output from the motion vector detection means. First binary code amount calculation means for calculating a binary symbol length of a differential motion vector from a motion vector signal, and third for calculating a binary symbol length of an orthogonal transform coefficient output from the quantization means. The binary symbol length of the signal excluding the differential motion vector and the orthogonal transform coefficient constituting the binarized symbol sequence from the macroblock information output from the binary code amount calculating means and the motion vector detecting means Second binary code amount calculating means for calculating is separately provided from the binarizing means, and the code amount control means is calculated by the first to third binary code amount calculating means. Valuation symbol And outputting said quantization means determines the quantization parameter based on.

  The image coding method of the present invention is an image coding method comprising a quantization step, a code amount control step, a binarization step, and a binary arithmetic coding step, and is output in the motion vector detection step. A first binary code amount calculating step for calculating a binary symbol length of a difference motion vector from the motion vector signal, and a third for calculating a binary symbol length of an orthogonal transform coefficient output in the quantization step. And calculating the binary symbol length of the signal excluding the differential motion vector and the orthogonal transform coefficient constituting the binarized symbol sequence from the macroblock information output in the binary code amount calculating step and the motion vector detecting step. A second binary code amount calculation step to be calculated separately from the binarization step, and in the code amount control step, calculation is performed in the first to third binary code amount calculation steps. Determine a quantization parameter based on the binarized symbol length is and outputs the quantization step.

  The program of the present invention is a program that causes a computer to execute an image encoding method including a quantization step, a code amount control step, a binarization step, and a binary arithmetic encoding step. A first binary code amount calculating step for calculating a binary symbol length of a differential motion vector from an output motion vector signal, and calculating a binarized symbol length of an orthogonal transform coefficient output in the quantization step Binarization of a signal excluding the difference motion vector and the orthogonal transform coefficient constituting the binarized symbol sequence from the macro block information output in the third binary code amount calculation step and the motion vector detection step A second binary code amount calculation step for calculating a symbol length, and in the code amount control step, in the first to third binary code amount calculation steps, Characterized in that to execute the picture coding method to be outputted to the quantization step determines the quantization parameter into the computer on the basis of the binarized symbol length issued.

  According to the present invention, the quantization parameter calculated based on the binary data code amount of MB (binary data code amount × α) can be used in the quantization of the next MB in the encoding order, As compared with the conventional method, it is possible to realize an image encoding device that further increases the accuracy of code amount control.

(First embodiment)
Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram showing a configuration example of an image encoding apparatus to which the present invention can be applied.
Before describing the image encoding device 100 of the present embodiment, the operation of CABAC encoding processing and code amount control when entropy encoding is performed using CABAC will be described.
First, CABAC encoding processing will be described with reference to FIG. 2. CABAC includes a binarization unit 201, a context calculation unit 202, a context memory 203, and a binary arithmetic encoding unit 204.

  The binarization unit 201 inputs information D21 such as a difference motion vector, an orthogonal transform coefficient, a prediction mode, etc., performs binarization processing determined by the standard, and generates a binarized symbol sequence D22. Since the binarization processing method is described in Non-Patent Document 1, detailed description thereof is omitted here.

  The context memory 203 stores symbols having a high probability of occurrence (most probable symbols, hereinafter referred to as MPS) and 460 occurrence probabilities for each context index.

  The context calculation unit 202 calculates a context index based on the binarized symbol, reads the MPS and the occurrence probability from the context memory 203 using the context index as an address, and inputs the MPS and the occurrence probability to the binary arithmetic encoding unit 204.

  The binary arithmetic encoding unit 204 performs arithmetic encoding processing based on the binary symbol, MPS, and occurrence probability. Note that the binary arithmetic coding unit 204 performs arithmetic coding for each binarized symbol, and it is necessary to update the MPS and occurrence probability of the context memory each time. It is known that parallelization is impossible.

  In the coding apparatus using the above-described CABAC system entropy coding, when code amount control is performed, it is difficult to return and reflect the generated code length to the code amount control at an early stage due to the characteristics of CABAC. Therefore, a method has been proposed in which the generated code length of the binary arithmetic encoding unit 204 is estimated from the code length of the binarized symbol output from the binarizing unit 201 (see, for example, Patent Document 1). Specifically, the arithmetic code length is estimated using the ratio α between the binary symbol length of the immediately previous coded picture and the arithmetic code length as the binary symbol length.

Next, the operation of code amount control when entropy encoding is performed using CABAC will be described with reference to FIG.
As shown in FIG. 3, when a difference image D31 between a predicted image and an input image generated by a motion compensation unit (not shown) is input, the orthogonal transform quantization unit 301 performs orthogonal transform and quantization on the input image. Apply. Then, information D32 obtained by combining this transform coefficient and information for each macroblock (hereinafter, macroblock information) such as the prediction mode is output to the binarization unit 302.

  The binarization unit 302 performs the same operation as the binarization unit 201 described above, and receives the conversion coefficient and macroblock information D32 as input, and the binarized symbol length for each binarized symbol D33 and macroblock (MB). D34 is generated.

  When the binarized symbol length D34 is input from the binarizing unit 302, the generated code length estimating unit 304 calculates the estimated arithmetic code length D35 using the ratio α described above, and outputs it to the code amount control unit 305. To do.

  The code amount control unit 305 determines a quantization parameter for quantizing the next MB based on the estimated arithmetic code length D35 input from the generated code length estimation unit 304 for each MB, and an orthogonal transform quantization unit The quantization parameter D36 is output to 301. The orthogonal transform quantization unit 301 performs orthogonal transform and quantization based on the quantization parameter D36 input from the code amount control unit 305. With such a configuration, code amount control when entropy encoding is performed using CABAC is performed.

Next, a configuration example of the image encoding device 100 of the present embodiment will be described with reference to FIG.
The image encoding apparatus 100 according to the present embodiment includes a motion vector detection unit 101 that detects a motion vector. The motion vector detection unit 101 generates and outputs a motion vector signal D10 including a difference value between a motion vector prediction value (hereinafter referred to as pmv) generated using the motion vector and peripheral blocks. In addition, macroblock information D11 mainly including a prediction mode, a partition pattern, an orthogonal transform size, an intra-inter flag, a field frame flag, and the like is generated and output. The motion vector signal D10 output from the motion vector detection unit 101 is input to the orthogonal transform quantization units 102 and 103.

  The orthogonal transform quantization unit 102 performs motion compensation of the motion vector signal D10 supplied from the motion vector detection unit 101. Then, after performing orthogonal transform on the prediction residual signal, quantization is performed based on the quantization parameter D16 input from the code amount control unit 106, and an orthogonal transform coefficient D12 is output.

  The binarization unit 103 generates and outputs a binarized symbol sequence D13 from the input motion vector signal D10, macroblock information D11, and orthogonal transform coefficient D12.

  The binarized symbol sequence D13 output from the binarizing unit 103 is input to the arithmetic encoding unit 104. The arithmetic encoding unit 104 performs context calculation of the input binarized symbol string D13, inputs the binarized symbol, the MPS described above, and the occurrence probability to an arithmetic encoder (not shown), Arithmetic coding is performed and a code D14 is output.

  Furthermore, it has a generated code amount estimation unit 105, a code amount control unit 106, and a differential motion vector binary code amount calculation unit (first binary code amount calculation unit) 107. In addition, a macroblock information binarized symbol length calculation unit (second binary code amount calculation unit) 108 and an orthogonal transform coefficient binary symbol length calculation unit (third binary code amount calculation unit) 109 are included. .

  The differential motion vector binary code amount calculation unit 107 calculates a differential motion vector binary symbol length D17 for each MB from the motion vector signal D10 input from the motion vector detection unit 101. When the macroblock information D11 is input, the macroblock information binarized symbol length calculation unit 108 calculates and outputs the macroblock information binarized symbol length D18 for each MB. When the orthogonal transform coefficient D12 output from the orthogonal transform quantization unit 102 is input, the orthogonal transform coefficient binarized symbol length calculation unit 109 calculates an orthogonal transform coefficient binarized symbol length D19 for each MB. Output.

  The generated code amount estimation unit 105 includes the above-described differential motion vector binarized symbol length D17 for each MB, macroblock information binarized symbol length D18 for each MB, and orthogonal transform coefficient binarized symbol length D19 for each MB. When input, the code length after arithmetic coding is estimated. Then, an estimated code length D15 is generated and output. Further, when the estimated code length D15 is output from the generated code amount estimation unit 105, the code amount control unit 106 calculates and outputs a quantization parameter D16 that determines a quantization step value for each encoding target MB.

  Since the operations of the motion vector detection unit 101, the orthogonal transform quantization unit 102, the binarization unit 103, and the arithmetic coding unit 104 are described in Non-Patent Document 1, detailed description thereof is omitted in this embodiment.

  The difference motion vector binary code amount calculation unit 107 calculates a binarized symbol length generated by the binarization of the CABAC method described in Non-Patent Document 1. This calculation process is performed on the difference motion vector (difference motion vector = motion vector−pmv) output from the motion vector detection unit 101. That is, the binarized symbol length when the difference motion vector is binarized by the UEG3 method described in Non-Patent Document 1 is calculated. The UEG3 scheme is Concatenated unary / k-th order Exp-Golomb binarization.

  A similar method is used for binarization of the difference motion vector performed in the binarization unit 103. Since the binarization unit 103 is a binarization process for generating a code, it is necessary to binarize the orthogonal transformation coefficient D12 by the UEG3 method according to the order of Zig-zag scan or Field scan. For this reason, the binarization unit 103 stores the orthogonal transform coefficient D12 in the storage device (not shown) for MB, and then reads out the orthogonal transform coefficient in the order of the Zig-zag scan or the field scan described above, and stores it in the UEG3 method. Convert to value.

  On the other hand, the difference motion vector binary code amount calculation unit 107, unlike the binarization unit 103, calculates only the binarized symbol length, so it is necessary to keep the order of Zig-zag scan or Field scan described above. Absent. Therefore, it is not necessary to store the orthogonal transform coefficient of the MB part in the storage device. That is, if the orthogonal transform coefficient D12 is output from the orthogonal transform quantization unit 102, the binarized symbol length can be calculated immediately. In order to calculate the code length, dedicated hardware for realizing the processing described in “9.3.2.3” of Non-Patent Document 1 may be mounted. Further, if the range of the difference motion vector is determined in advance, the code lengths of all the difference motion vector values are recorded in the ROM, and the code length may be immediately calculated using the difference motion vector values as indexes. Good.

  Macroblock information binarized symbol length calculation section 108 calculates the binarized symbol length of Syntax Element (hereinafter referred to as SE) described in Non-Patent Document 1. Elements included in the Syntax Element have the following names: That is, mb_type. mb_skip_flag. sub_mb_type. ref_idx_l0. ref_idx_l1. mb_qp_delta. intra_chroma_pred_mode. prev_intra4x4_pred_mode_flag. prev_intra8x8_pred_mode_flag. rem_intra4x4_pred_mode. rem_intra8x8_pred_mode. mb_field_decoding_flag. coded_block_pattern. It has the name end_of_slice_flag.

  mb_type is an SE indicating the MB type. When the MB to be encoded is intra prediction, the type (I, P, B) of the slice being encoded and transform_size_8x8_flag indicating the orthogonal transform size are input from the motion vector detection unit 101.

  In the case of intra prediction, a signal indicating whether the prediction is 4x4, 8x8, or 16x16 is determined. In addition, the prediction mode of intra 16 × 16 prediction, macroblock information such as Coded Block Pattern Luma, Coeded Block Pattern Chroma, and the like are input, and mb_type is determined. When mb_type is determined, the code length is calculated using the table described in Table 9-27 of Non-Patent Document 1.

  The aforementioned Coded Block Pattern Luma and Coded Block Pattern Chroma are SEs determined from the presence or absence of non-zero orthogonal transform coefficients in the MB. The SE value is determined at the timing when all the orthogonal transform coefficients in the MB are output from the orthogonal transform quantization unit 102 described above.

  mb_skip_flag is an SE that becomes 1 when the MB being encoded is skipped. The determination of skip is made by determining whether or not the values of all orthogonal transform coefficients are 0 and the MVD is 0 when the slice to be encoded is a P slice and a B slice.

  When the MB to be encoded is intra prediction, the binarized symbol length of mb_skip_flag is 0, but in other predictions, the binarized symbol length of mb_skip_flag is a fixed length, so the symbol length is calculated. Is possible.

  sub_mb_type is an SE existing at the time of P slice or B slice, and indicates an MB type of 8 × 8 or less of MB. As the binarized symbol length of sub_mb_type, a signal indicating how many 8 × 8 blocks are divided, the prediction method for each partition, and the vertical and horizontal sizes of each partition are input from the motion vector detection unit 101 as macroblock information D11. To do. Thus, calculation is performed using Table 9-29 described in Non-Patent Document 1.

  ref_idx_l0 and ref_idx_l1 are SEs indicating reference picture numbers. A reference picture number is input as macroblock information D11 from the motion vector detection unit 101, and a value obtained by adding 1 to the absolute value of SE is a binarized symbol length. The reason for this is described in Non-Patent Document 1 and will not be described.

  mb_qp_delta indicates the difference value between the quantization parameters of MBs before and after viewed in the coding order. A quantization parameter value is input from the orthogonal transform quantization unit 102 or the code amount control unit 106, and a difference value (SE value) from the previous value is calculated. Then, it is converted into code Num (described in Table 9-3 of Non-Patent Document 1), and the binarized symbol length is calculated.

  intra_chroma_pred_mode is an SE indicating the intra prediction mode of the color difference signal for MB. The intra prediction mode of the color difference signal is input as the macroblock information D11 from the motion vector detection unit 101, and the binarized symbol length corresponding to the SE value is calculated. Intra_chroma_pred_mode is binarized by the binarization unit 103 using Truncated unary cMax = 3. Since the binarized symbol length by this method is fixed, it is possible to calculate the binarized symbol length of intra_chroma_pred_mode.

  prev_intra4x4_pred_mode_flag is SE which shows the prediction mode in a screen with respect to each of 4x4 block. In the binarization unit 103, the SE is binarized using the Fixed-length cMax = 1 method. Moreover, prev_intra8x8_pred_mode_flag is also binarized by the same binarization method. Since the binarized symbol length for the SE is a fixed value, the binarized symbol length can be calculated.

  rem_intra4x4_pred_mode and rem_intra8x8_pred_mode are SE indicating the intra prediction mode for the 4x4 block or the 8x8 block. The SE is binarized in the binarization unit 103 when the predetermined flag in the macroblock information D11 from the motion vector detection unit 101 is 0. That is, it is binarized when prev_intra4x4_pred_mode_flag or prev_intra8x8_pred_mode_flag is 0.

  That is, when the flag is 1, binarization symbol length does not occur because binarization is not performed. However, since it is possible to know the values of the 16 flags from the motion vector detection unit 101, the binarized symbol length for rem_intra4x4_pred_mode and rem_intra8x8_pred_mode can also be calculated. Note that rem_intra4x4_pred_mode and rem_intra8x8_pred_mode are binarized by the binarization unit 103 using the Fixed-length cMax = 7 method. Therefore, the binarized symbol length of the SE is a fixed length.

  mb_field_decoding_flag is an SE that becomes 1 when an MB pair that considers MBs adjacent in the vertical direction as one pair is in the field mode. The binarization unit 103 binarizes only when encoding in the MB-AFF mode.

  Input from the motion vector detection unit 101 a signal indicating whether or not the MB-AFF mode is included, and a signal indicating whether or not the previous MB was skipped in the encoding order and the address of the current MB included in the macroblock information D11 To do. Then, it is determined whether or not binarization is necessary, and when it is necessary to binarize, a binarized symbol length is calculated. Since this SE is binarized by the Fixed-length cMax = 1 method, since the binarized symbol length is fixed, it is only necessary to determine whether or not to binarize.

  coded_block_pattern is SE determined from the presence or absence of non-zero orthogonal transform coefficient in MB. Since the binarized symbol length of the SE is fixed regardless of the value of the SE, the binarized symbol length can be calculated.

  end_of_slice_flag is an SE that appears at the end of the binarized symbol string of the MB and becomes 1 when the encoding of the last MB of the slice is finished. Since the binarized symbol length of the SE is a fixed length, a fixed binarized symbol length may be output for each MB.

  The orthogonal transform coefficient binarized symbol length calculation unit 109 receives the orthogonal transform coefficient D12 orthogonally transformed and quantized by the orthogonal transform quantization unit 102. Orthogonal transform coefficient binarized symbol length calculation section 109 calculates the binarized symbol length of coded_block_flag which is SE. Also, the binarized symbol length of significant_coeff_flag is calculated. Also, the binarized symbol length of last_significant_coeff_flag is calculated. Also, the binarized symbol length of coeff_abs_level_minus1 and coeff_sign_flag is calculated.

  coded_block_flag is an SE that takes a value of 1 if there is a non-zero coefficient in the 4x4 block. The binarized symbol length of the SE can be calculated at the timing when all the coefficients in the MB are output from the orthogonal transform quantization unit 102. For example, in 1 MB, there are 24 4 × 4 blocks including luminance and color difference (when chroma_format is 4: 2: 0). If non-zero coefficients exist in 10 blocks out of 24 blocks, the binarized symbol length of coded_block_flag is 10 bits.

  significant_coeff_flag is an SE that takes a value of 1 if one transform coefficient is non-zero. last_significant_coeff_flag is an SE that is 1 when it is the last non-zero transform coefficient of a 4 × 4 block including one transform coefficient. Since the binarization unit 103 needs to output the binarized symbols after binarization in the order conforming to the standard, it is necessary to temporarily store the conversion coefficient for 1 MB in the storage device. The number binarized symbol length calculation unit 109 only needs to calculate the binarized symbol length. Therefore, it can be calculated at the timing when all the coefficients in the MB are output from the orthogonal transform quantization unit 102.

  coeff_abs_level_minus1 is SE which takes the value which subtracted 1 from the absolute value of the orthogonal transformation coefficient D12 output from the orthogonal transformation quantization part 102, and is binarized by UEGk system. The SE can also be calculated at the timing when all the coefficients in the MB are output from the orthogonal transform quantization unit 102. The binarized symbol length may be calculated by preparing a ROM or the like that holds the binarized symbol length corresponding to the SE value in advance.

  coeff_sign_flag is an SE indicating the code of the orthogonal transform coefficient D12 output from the orthogonal transform quantization unit 102. When this orthogonal transform coefficient D12 is non-zero, a 1-bit binary symbol length is generated. The binarized symbol length of the SE can also be calculated at the timing when all the coefficients in the MB are output from the orthogonal transform quantization unit 102.

Next, the operation procedure of the above-described image encoding device 100 of the present embodiment will be described with reference to the flowchart of FIG.
First, in step S601, a motion vector is detected by the motion vector detection unit 101, and a motion vector signal D10 and macroblock information D11 are generated and output as detection results.

  Next, in step S602, the orthogonal transform quantization unit 102 performs orthogonal transform and quantization to generate an orthogonal transform coefficient D12, and a binarization unit 103 and an orthogonal transform coefficient binarized symbol length calculation unit 109. Output to.

Next, in step S603, the differential motion vector binary code amount calculation unit 107 calculates a differential motion vector binary symbol length D17 for each MB.
Next, in step S604, the macroblock information binarized symbol length calculation unit 108 calculates the macroblock information binarized symbol length D18 for each MB. The macroblock information binarized symbol length D18 for each MB is calculated as the binarized symbol length of the signal excluding the differential motion vector and the orthogonal transform coefficient constituting the binarized symbol sequence.
Next, in step S605, the orthogonal transform coefficient binarized symbol length calculation unit 109 calculates an orthogonal transform coefficient binarized symbol length D19 for each MB.

  Next, in step S606, the generated code amount estimation unit 105 performs processing for estimating the code length after arithmetic coding. This process is performed when the above-described differential motion vector binarized symbol length D17 for each MB, macroblock information binarized symbol length D18 for each MB, and orthogonal transform coefficient binarized symbol length D19 for each MB are input. This is performed by estimating the code length after arithmetic coding.

  Next, in step S607, the code amount control unit 106 calculates a quantization parameter D16 that determines a quantization step value for each encoding target MB, and controls the quantization operation performed by the orthogonal transform quantization unit 102. . Thereby, the quantization parameter calculated based on the binary data code amount of MB (binary data code amount × α) can be used in the quantization of the next MB in the encoding order.

Next, in step S608, the binarization unit 103 performs binarization processing of the difference motion vector. Since this process is a binarization process for generating a code as described above, the orthogonal transform coefficient D12 is binarized by the UEG3 method in the order of Zig-zag scan or Field scan.
Next, in step S609, the arithmetic encoding unit 104 performs binary arithmetic encoding and outputs a code D14.

  Next, in step S610, it is determined whether or not to end the process. As a result of the determination, if the process is not terminated, the process returns to step S610 to repeat the above-described process. If all the processes are ended as a result of the determination in step S610, end processing is performed.

  As described above, the image coding apparatus 100 according to the present embodiment includes the difference motion vector binary code amount calculation unit 107 to the orthogonal transform coefficient binary symbol length calculation unit 109 (first to third binary code amounts). A calculation unit). Therefore, the binarized symbol length of the MB can be calculated at the timing of time t2 in FIG. As a result, the binarized symbol length can be returned to the code amount control unit 305 earlier than the conventional configuration by 1 MB time, and more accurate code amount control can be realized.

(Other embodiments according to the present invention)
Each means constituting the image coding apparatus according to the above-described embodiment of the present invention can be realized by operating a program stored in a RAM or ROM of a computer. This program and a computer-readable recording medium recording the program are included in the present invention.

  In addition, the present invention can be implemented as, for example, a system, apparatus, method, program, storage medium, or the like. Specifically, the present invention may be applied to a system including a plurality of devices. The present invention may be applied to an apparatus composed of a single device.

  In the present invention, a software program (in the embodiment, a program corresponding to the flowchart shown in FIG. 6) for executing each step in the above-described image encoding method is directly or remotely supplied to a system or apparatus. In addition, this includes a case where the system or the computer of the apparatus is also achieved by reading and executing the supplied program code.

  Accordingly, since the functions of the present invention are implemented by computer, the program code installed in the computer also implements the present invention. In other words, the present invention includes a computer program itself for realizing the functional processing of the present invention.

  In that case, as long as it has the function of a program, it may be in the form of object code, a program executed by an interpreter, script data supplied to the OS, and the like.

  Various recording media can be used as a recording medium for supplying the program. For example, floppy (registered trademark) disk, hard disk, optical disk, magneto-optical disk, MO, CD-ROM, CD-R, CD-RW, magnetic tape, nonvolatile memory card, ROM, DVD (DVD-ROM, DVD- R).

  As another program supply method, a browser on a client computer is used to connect to an Internet home page. The computer program itself of the present invention or a compressed file including an automatic installation function can be downloaded from the homepage by downloading it to a recording medium such as a hard disk.

  It can also be realized by dividing the program code constituting the program of the present invention into a plurality of files and downloading each file from a different homepage. That is, a WWW server that allows a plurality of users to download a program file for realizing the functional processing of the present invention on a computer is also included in the present invention.

  In addition, the program of the present invention is encrypted, stored in a storage medium such as a CD-ROM, distributed to users, and key information for decryption is downloaded from a homepage via the Internet to users who have cleared predetermined conditions. Let It is also possible to execute the encrypted program by using the key information and install the program on a computer.

  In addition to the functions of the above-described embodiments being realized by the computer executing the read program, the OS running on the computer may perform part or all of the actual processing. The functions of the above-described embodiments can be realized.

  Further, the program read from the recording medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer. Thereafter, the CPU of the function expansion board or function expansion unit performs part or all of the actual processing based on the instructions of the program, and the functions of the above-described embodiments are realized by the processing.

It is a block diagram which shows embodiment of this invention and shows the structure of an image coding apparatus. H. 2 is a block diagram showing a configuration of a CABAC system in H.264. It is a block diagram which shows the structure of the conventional image coding apparatus. It is a figure which shows the processing content for every time of the conventional image coding apparatus. It is a figure which shows embodiment of this invention and shows the processing content for every time of an image coding apparatus. It is a flowchart which shows embodiment of this invention and demonstrates the operation | movement procedure of an image coding apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Image coding apparatus 101 Motion vector detection part 102 Orthogonal transformation quantization part 103 Binarization part 104 Arithmetic coding part 105 Generated code amount estimation part 106 Code quantity control part 107 Difference motion vector binarization symbol length calculation part 108 Macro Block information binarized symbol length calculation unit 109 Orthogonal transform coefficient binarized symbol length calculation unit 201 Binarization unit 202 Context calculation unit 203 Context memory 204 Binary arithmetic coding unit

Claims (8)

In an image encoding device comprising quantization means, code amount control means, binarization means, and binary arithmetic encoding means,
First binary code amount calculating means for calculating a binary symbol length of a differential motion vector from a motion vector signal output from the motion vector detecting means;
Third binary code amount calculation means for calculating the binarized symbol length of the orthogonal transform coefficient output from the quantization means;
A second binary code amount for calculating a binary symbol length of a signal excluding the differential motion vector and the orthogonal transform coefficient constituting the binary symbol sequence from the macroblock information output from the motion vector detecting means A calculation means separately from the binarization means,
The code amount control means determines a quantization parameter based on the binarized symbol length calculated by the first to third binary code amount calculation means, and outputs the quantization parameter to the quantization means. An image encoding device.   Generated code amount estimating means for estimating the code length of the binary arithmetic coding means from the binarized symbol length calculated by the first to third binary code amount calculating means and outputting the estimated code length to the code amount control means The image encoding apparatus according to claim 1, further comprising:   The image coding apparatus according to claim 1 or 2, wherein the binarization means and the binary arithmetic coding means are CABAC systems. In an image encoding method comprising a quantization step, a code amount control step, a binarization step, and a binary arithmetic encoding step,
A first binary code amount calculating step of calculating a binary symbol length of a difference motion vector from a motion vector signal output in the motion vector detection step;
A third binary code amount calculating step of calculating a binarized symbol length of the orthogonal transform coefficient output in the quantization step;
A second binary code amount for calculating a binarized symbol length of a signal excluding the differential motion vector and the orthogonal transform coefficient constituting the binarized symbol sequence from the macroblock information output in the motion vector detecting step A calculation step separately from the binarization step,
In the code amount control step, a quantization parameter is determined based on the binarized symbol length calculated in the first to third binary code amount calculation steps, and is output to the quantization step. An image encoding method.   Generated code amount estimation step of estimating the code length of the binary arithmetic encoding step from the binarized symbol length calculated in the first to third binary code amount calculation steps and outputting the estimated code length to the code amount control step The image encoding method according to claim 4, further comprising:   6. The image encoding method according to claim 4, wherein the binarization step and the binary arithmetic encoding step are a CABAC method. In a program for causing a computer to execute an image encoding method including a quantization step, a code amount control step, a binarization step, and a binary arithmetic encoding step,
A first binary code amount calculating step of calculating a binary symbol length of a difference motion vector from a motion vector signal output in the motion vector detection step;
A third binary code amount calculating step of calculating a binarized symbol length of the orthogonal transform coefficient output in the quantization step;
A second binary code amount for calculating a binary symbol length of a signal excluding the differential motion vector and the orthogonal transform coefficient constituting the binary symbol sequence from the macroblock information output in the motion vector detection step A calculation step,
In the code amount control step, image coding for determining a quantization parameter based on the binarized symbol length calculated in the first to third binary code amount calculation steps and outputting the quantization parameter to the quantization step A program for causing a computer to execute the method.   A computer-readable storage medium storing the program according to claim 7.

Images