JP2013012961A - Image decoder, image decoding method and image decoding program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve quality of a prediction signal.SOLUTION: A decoding string generating section 203 obtains a code string, performs decoding string generation processing in response to a prescribed syntax structure, and generates at least residual motion vector information, residual encoded information and an encoding parameter, which belong to a region being a decoding object from the code string. A motion predicting section 208 performs prescribed motion prediction processing in response to motion information belonging to a decoded region, generates first reference information and generates first prediction motion vector information in a first reference picture specified on the basis of the first reference information. Correction prediction signal generating sections 210 and 211 derive second prediction motion vector information in a second reference picture specified on the basis of second reference information from the first prediction motion vector information, generate motion vector information by adding the derived second prediction motion vector information to the residual motion vector information and generate the prediction signal from the second reference picture in response to the generated motion vector information.

Description

  The present invention relates to an image decoding apparatus, an image decoding method, and an image decoding program for decoding a moving image signal.

  MPEG-4 AVC / H. In moving picture coding represented by H.264 (hereinafter referred to as AVC), motion compensation is used as information compression using correlation in the time direction. Motion compensation uses a locally decoded signal that has already been encoded and decoded as a reference picture for a picture to be encoded (hereinafter referred to as a target picture) that is an image signal to be encoded. This is a technique for detecting a motion amount (hereinafter referred to as motion vector information) between a target picture and a reference picture and generating a prediction signal in units of processing.

  Here, in AVC, the size of a block to be processed (hereinafter referred to as a target block) within a 16 × 16 pixel two-dimensional block (hereinafter referred to as a macroblock) which is a predetermined encoding processing unit is variable. A method of predicting using motion vector information for each target block, a method of selecting a reference picture to be used for prediction from a plurality of reference pictures, and a prediction signal from motion vector information between two reference pictures and the target block In particular, the accuracy of motion compensation is improved and the amount of information is reduced by a method for generating a synthesized prediction signal.

  In addition, in AVC, Patent Document 1, and the like, encoding is performed using predicted motion vector information predicted from motion vector information for a decoded block around a target block in order to suppress an increase in the amount of information due to motion vector information. Thus, it has been proposed to perform motion compensation using a direct mode that does not transmit motion vector information necessary for encoding and decoding a target block, and an improvement technique thereof.

JP 2006-191652 A

  However, in the motion compensation by the direct mode in the conventional moving image coding represented by Patent Document 1 and AVC, the motion of the object included in the moving image is in a fixed direction between pictures of a sufficiently short time interval. It is assumed to move continuously at a constant speed. Therefore, when trying to encode a moving image that has a change in the direction and speed of the movement of an object included in the screen, the direct mode cannot sufficiently follow the movement, and it is difficult to obtain a sufficient prediction signal. is there.

  Further, as a feature of the image signal in the target block, for example, in the case of a signal including a strong edge component, even if the spatial position of the prediction signal acquired from the reference picture is slightly shifted, There is a difficult problem that the residual information obtained by subtracting the prediction signal specified from the image signal becomes large, and the code amount increases. Therefore, when the motion vector information predicted by the direct mode cannot follow the motion and sufficient alignment accuracy is not obtained, it is difficult to obtain a prediction signal having a high spatial accuracy. In this case, there is a difficult problem that the residual information similarly increases, leading to an increase in the amount of codes.

  Further, in the conventional video coding represented by AVC, in order to obtain a prediction signal in which the spatial phase is matched with high accuracy, for example, in the coding of a B picture, prediction is performed from one front or rear reference picture. There is a method of obtaining a signal (one-way prediction) and a method of obtaining a prediction signal by averaging after specifying two reference signals from two or less reference pictures (bi-prediction). These methods identify motion vector information by performing motion estimation using predicted motion vector information predicted from encoded blocks around the target block, and subtract the predicted motion vector information. Residual motion vector information is generated and transmitted after encoding.

  Here, one residual motion vector information is transmitted after unidirectional prediction, and two residual motion vector information is transmitted after bi-prediction. However, in unidirectional prediction, since a prediction signal can be obtained only from one reference picture in one direction, if the required signal is not included in the reference picture, it is not always optimal by motion estimation. It is possible that motion vector information indicating a region that cannot be specified is specified. Therefore, even though one residual motion vector information and auxiliary information (hereinafter referred to as reference index information or simply reference information) for specifying a reference picture to be used are allocated, an amount of information is not necessarily allocated. There is a difficult problem that a sufficient prediction signal may not be obtained. In bi-prediction, since reference signals are obtained from two locations and a prediction signal is synthesized, such a one-way prediction problem can be improved to some extent. It is necessary to transmit the reference index information after encoding, which causes a difficult problem of increasing the code amount.

  Therefore, without increasing the motion vector information to be transmitted and the auxiliary information necessary for specifying the motion vector information, the prediction residual is reduced by improving the quality of the prediction signal in bi-prediction, and the coding efficiency is improved. There is a need to improve.

  The present invention has been made in view of such circumstances, and an object thereof is to provide a technique for improving the quality of a prediction signal.

  In order to solve the above-described problem, an image decoding device according to an aspect of the present invention acquires a code string, performs a decoded string generation process based on a predetermined syntax structure, and performs decoding from the code string. A decoding sequence generation unit (203) that generates at least residual motion vector information, residual encoded information, and encoding parameters belonging to, and a predetermined motion prediction process based on the motion information belonging to the decoded region A motion prediction unit (208) configured to generate first reference information and generate first prediction motion vector information in a first reference picture specified based on the first reference information; Second prediction motion vector information in the second reference picture specified based on the second reference information is derived from the one prediction motion vector information, and the derived second prediction motion vector information A modified prediction signal generator (210, 211) that generates motion vector information in addition to the residual motion vector information, and generates a prediction signal from the second reference picture based on the generated motion vector information; A residual decoding unit (204) that performs a predetermined residual decoding process on the post-difference information and generates decoded residual information, and the modified prediction signal generation unit (210, 211) is included in the decoded residual information. And an addition unit (205) that generates a decoded signal by adding the prediction signal generated by (1).

  The modified prediction signal generation unit (210, 211) generates a first prediction signal from the first reference picture based on the first prediction motion vector information, and generates the second prediction motion vector information. Based on this, a second prediction signal may be generated from the second reference picture, and the first prediction signal and the second prediction signal may be combined to generate a combined prediction signal. The addition unit (205) may generate a decoded signal by adding the combined prediction signal generated by the modified prediction signal generation unit (210, 211) to decoding residual information.

  This image decoding apparatus may specify the encoding mode, the motion prediction mode, and the operation mode from the decoding parameter using the encoding parameter as a decoding parameter. The modified prediction signal generation unit (210, 211) operates based on an operation mode for distinguishing a plurality of operations, and the second corresponding to the operation mode by the derivation process specified by the operation mode. Deriving predicted motion vector information, adding the derived second predicted residual motion vector information to residual motion vector information, generating motion vector information corresponding to the operation mode, and based on the generated motion vector information A second prediction signal corresponding to the operation mode may be generated from the second reference picture.

  This image decoding apparatus may specify the encoding mode, the motion prediction mode, and the operation mode from the decoding parameter using the encoding parameter as a decoding parameter. The motion prediction unit (208) operates based on a motion prediction mode for distinguishing a plurality of motions, and predicted motion vector information corresponding to the motion prediction mode by a motion prediction process specified by the motion prediction mode. Candidates may be identified. The modified prediction signal generation unit (210, 211) specifies prediction motion vector information corresponding to the operation mode from candidates of prediction motion vector information corresponding to the motion prediction mode, and the first prediction motion Vector information may be used.

  This image decoding apparatus may specify the encoding mode, the motion prediction mode, and the operation mode from the decoding parameter using the encoding parameter as a decoding parameter. When the coding mode is an AMVP (Advanced Motion Vector prediction) mode, the motion prediction unit (208) performs the best prediction from prediction motion vector information candidates by AMVP processing according to the AMVP mode of the coding mode. Motion vector information may be determined. The modified prediction signal generation unit (210, 211) may acquire the determined predicted motion vector information according to the AMVP mode of the coding mode, and may use the first predicted motion vector information.

  This image decoding apparatus may specify the encoding mode, the motion prediction mode, and the operation mode from the decoding parameter using the encoding parameter as a decoding parameter. When the coding mode is the merge mode, the motion prediction unit (208) determines the best predicted motion vector information from the prediction motion vector information candidates by the merge process according to the merge mode of the coding mode. May be. The modified prediction signal generation unit (210, 211) may acquire the predicted motion vector information determined according to the merge mode of the coding mode, and use the information as the first motion vector predictor information.

  Another aspect of the present invention is an image decoding method. In this method, a code string is acquired, a decoded string generation process is performed based on a predetermined syntax structure, and at least residual motion vector information belonging to an area to be decoded from the code string, after residual encoding A step of generating information and an encoding parameter, and performing a predetermined motion prediction process based on motion information belonging to the decoded region to generate first reference information, and based on the first reference information Generating first predictive motion vector information in the specified first reference picture; and second information in the second reference picture specified based on second reference information from the first predictive motion vector information. 2 prediction motion vector information is derived, and the derived second motion vector predictor information is added to the residual motion vector information to generate motion vector information. Generating a prediction signal from the second reference picture based on the toll information, performing a predetermined residual decoding process on the residual encoded information, and generating decoded residual information; Adding the prediction signal to decoded residual information to generate a decoded signal.

  Another aspect of the present invention is an image encoding device. This apparatus is an image encoding apparatus that acquires and encodes a moving image sequence, performs predetermined motion prediction processing based on motion information belonging to an encoded region, and generates first reference information In addition, a motion prediction unit (104) that generates first prediction motion vector information in the first reference picture specified based on the first reference information, and a second prediction motion vector information from the first prediction motion vector information, Second prediction motion vector information in the second reference picture specified based on the reference information of the second reference picture is derived, and a predetermined reference area is determined in the region of the second reference picture based on the derived second prediction motion vector information. A motion search is performed to identify motion vector information, a prediction signal is generated from the second reference picture based on the identified motion vector information, and the motion vector information is used to generate the prediction signal. A modified prediction signal generation unit (106, 107) that generates residual motion vector information by subtracting the prediction motion vector information of No. 2, and subtracts the prediction signal generated by the modified prediction signal generation unit from the target signal A subtraction unit (109) that generates difference information, a residual encoding unit (110) that performs predetermined residual encoding processing on the residual information and generates post-residual encoding information, And a code string generation unit (113) for generating a code string of the residual encoded information, the residual motion vector information, and an encoding parameter based on a syntax structure.

  Note that any combination of the above-described components and a representation of the present invention converted between a method, an apparatus, a system, a recording medium, a computer program, and the like are also effective as an aspect of the present invention.

  According to the present invention, the quality of a prediction signal can be improved.

It is a block diagram of the moving image encoder in Embodiment 1 of this invention. It is a block diagram which shows the connection relationship of an encoding control part and each part of the moving image encoder in Embodiment 1 of this invention. It is a basic flowchart of the moving image encoder in Embodiment 1 of this invention. It is a block diagram of the moving image decoding apparatus in Embodiment 1 of this invention. It is a block diagram which shows the connection relation of a decoding control part and each part of the moving image decoding apparatus in Embodiment 1 of this invention. It is a basic flowchart of the moving image decoding apparatus in Embodiment 1 of this invention. FIGS. 7A to 7D are configuration diagrams illustrating details of the modified prediction signal generation unit on the video encoding device side according to Embodiment 2 of the present invention. FIGS. 8A to 8C are configuration diagrams illustrating details of the first modified prediction signal generation unit on the video encoding device side in Embodiment 2 of the present invention. FIGS. 9A to 9C are configuration diagrams illustrating details of the second modified prediction signal generation unit on the moving image encoding device side according to Embodiment 2 of the present invention. It is a flowchart which shows operation | movement of each part of the 1st correction prediction signal production | generation part by the side of a moving image encoder in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 1st correction prediction signal production | generation part by the side of a moving image encoder in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 1st correction prediction signal production | generation part by the side of a moving image encoder in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 2nd correction prediction signal production | generation part by the side of a moving image encoder in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 2nd correction prediction signal production | generation part by the side of a moving image encoder in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 2nd correction prediction signal production | generation part by the side of a moving image encoder in Embodiment 2 of this invention. FIGS. 16A to 16C are configuration diagrams illustrating details of the first modified prediction signal generation unit on the video decoding device side in Embodiment 2 of the present invention. FIGS. 17A to 17C are configuration diagrams illustrating details of the second modified prediction signal generation unit on the video decoding device side in Embodiment 2 of the present invention. It is a flowchart which shows operation | movement of each part of the 1st correction prediction signal generation part by the moving image decoding apparatus side in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 1st correction prediction signal generation part by the moving image decoding apparatus side in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 1st correction prediction signal generation part by the moving image decoding apparatus side in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 2nd correction prediction signal generation part by the moving image decoding apparatus side in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 2nd correction prediction signal generation part by the moving image decoding apparatus side in Embodiment 2 of this invention. It is a flowchart which shows operation | movement of each part of the 2nd correction prediction signal generation part by the moving image decoding apparatus side in Embodiment 2 of this invention. It is a table | surface (the 1) for demonstrating the operation mode in Embodiment 2 of this invention. It is a table | surface (2) for demonstrating the operation mode in Embodiment 2 of this invention. It is a conceptual diagram for showing the relationship between a target prediction block and an encoded adjacent block. It is a conceptual diagram for showing the relationship between a target prediction block and a neighboring block at the same position in another picture with different time. FIGS. 28A and 28B are diagrams in which the encoding modes are roughly classified based on the encoding prediction modes. It is the conceptual diagram which showed an example of the code sequence in Embodiment 4 of this invention. FIGS. 30A and 30B are diagrams illustrating an example of syntax that is a predetermined syntax structure of a code string generated by moving image coding. FIGS. 31A to 31C are diagrams showing an example of code string syntax in Embodiment 4 of the present invention. FIGS. 32A to 32C are diagrams illustrating another example of the syntax of the code string according to Embodiment 4 of the present invention. 33A to 33D are diagrams showing another example of the syntax of the code string in Embodiment 4 of the present invention. 34A to 34D are diagrams showing another example of the syntax of the code string according to Embodiment 4 of the present invention. It is a conceptual diagram of the moving image transmitter and moving image receiver in Embodiment 5 of this invention. It is a conceptual diagram of the information processing apparatus in Embodiment 6 of this invention.

  EMBODIMENT OF THE INVENTION Below, the form for inventing is demonstrated with reference to drawings.

(Embodiment 1) Basic configuration and operation of moving picture coding apparatus and moving picture decoding apparatus.
First, the moving picture coding apparatus according to Embodiment 1 of the present invention will be described. FIG. 1 is a configuration diagram of a moving picture coding apparatus 100 according to Embodiment 1 of the present invention. FIG. 2 is a configuration diagram illustrating a connection relationship between each unit of the video encoding device 100 and the encoding control unit 120 according to Embodiment 1 of the present invention.

  The moving image coding apparatus 100 according to Embodiment 1 includes an input unit 101, a picture storage unit 102, a motion information storage unit 103, a motion prediction unit 104, a reference analysis unit 105, a first modified prediction signal generation unit 106, and a second modification. Prediction signal generation unit 107, mode determination unit 108, subtraction unit 109, residual encoding unit 110, residual decoding unit 111, addition unit 112, code sequence generation unit 113, code sequence storage unit 114, code amount control unit 115, At least an output unit 116 and an encoding control unit 120 are provided. Furthermore, it may be configured to include a unidirectional prediction signal generation unit 117, a bi-prediction signal generation unit 118, and an intra prediction unit 119. In the first embodiment, the description will be made assuming that all of these are included.

  In the first embodiment, for convenience, the operation mode of the entire moving image encoding process is referred to as an encoding operation mode or simply an encoding mode, and the first modified prediction signal generation unit 106 and the second modification prediction signal generation unit 107 The operation mode based on the operation is simply referred to as an operation mode. The operation mode in the motion prediction unit 104 is referred to as a motion prediction mode.

  The input unit 101 obtains a predetermined spatial resolution as a moving image signal from, for example, the recording device 133 or the storage device 135 that uses or does not use the imaging device 131, the transmission device 132, and the recording medium 134 as shown in FIG. Get input image sequence. The input unit 101 uses, as a target signal, a subtraction unit 109, a mode determination unit, and a signal in a target region that is in a spatial region currently being encoded in a target picture to be encoded in the input image sequence. 108, the unidirectional prediction signal generation unit 117, the first modified prediction signal generation unit 106, the second modified prediction signal generation unit 107, and the bi-prediction signal generation unit 118. Here, the target area is a current encoding target by dividing the target picture without any gap by a rectangular area (hereinafter referred to as a block) having a predetermined size, which is a spatial two-dimensional area, such as AVC. The target block. Here, the shape of the target area is a block, but it may be configured to divide without any gap by an arbitrary shape and supply the target signal of the target area in each area, and it is not particularly limited. .

  The picture storage unit 102 stores a reference picture and supplies it to each unit as necessary. Here, in order to easily specify the reference picture to be stored, for example, it is preferable to store the picture after associating it with a picture order count (hereinafter referred to as POC) which is a number for indicating the order of pictures. POC is used in AVC as information for specifying a reference picture. The picture accumulation unit 102 acquires the decoded signal generated by the addition unit 112 and accumulates it as a reference picture. The picture storage unit 102 supplies the requested reference picture to the unidirectional prediction signal generation unit 117, the first modified prediction signal generation unit 106, the second modified prediction signal generation unit 107, and the bi-prediction signal generation unit 118. . When acquiring the decoded signal from the adding unit 112, the picture storage unit 102 sequentially acquires the image signal of the decoded region obtained by local decoding after encoding, and then acquires the image signal of the local decoding region (hereinafter referred to as local decoding). (Picture). The picture storage unit 102 may hold the image signal of the locally decoded picture as a reference picture when the local decoding result of the entire encoding target picture is obtained. The picture storage unit 102 supplies the requested local decoded picture to the intra prediction unit 119. If necessary, the picture storage unit 102 is also requested to the unidirectional prediction signal generation unit 117, the first modified prediction signal generation unit 106, the second modified prediction signal generation unit 107, and the bi-prediction signal generation unit 118. Alternatively, a locally decoded picture may be supplied. When the reference information is supplied when supplying the reference picture or the locally decoded picture to each unit, the picture storage unit 102 specifies the POC from the reference information, and specifies the requested reference picture or locally decoded picture. And supply.

  The motion information storage unit 103 stores motion information. Here, the motion information includes predicted motion vector information, reference information, corrected predicted motion vector information (hereinafter, corrected predicted motion vector information), corrected reference information (hereinafter, corrected reference information), motion vector information, and remaining information. It is assumed that differential motion vector information is included. The motion information accumulation unit 103 collectively manages motion information used for each prediction processing unit (hereinafter, prediction unit) in the target picture and the reference picture. For example, when performing bi-prediction as in AVC, it is preferable that two pieces of motion information can be managed for each prediction unit. Here, for convenience, one is referred to as motion information on the list 0 (hereinafter, L0) side, and the other is referred to as motion information on the list 1 (hereinafter, L1) side. In addition, information relating to a picture that can be referred to is managed separately as a reference list in L0 and L1 as in AVC, for example, so that a reference picture can be specified by a value stored in reference information such as ref_idx_l0 and ref_idx_l1 Good.

  The motion information storage unit 103 manages the motion information in association with the mode information so that the motion information obtained by what prediction mode can be specified. The motion information accumulating unit 103 identifies motion information and a reference picture so that motion information belonging to which prediction unit region (hereinafter referred to as reference prediction region) in which reference picture belongs to. It is managed in association with information. As information for specifying a reference picture, for example, POC, which is a number for indicating the order of pictures as used in AVC, or reference information such as ref_idx may be used. Here, when reference information is used, for example, when reference information having a configuration such as AVC is used, the POC is specified from the reference information as necessary, thereby belonging to the required reference prediction region. It becomes possible to specify motion information. For example, the motion information storage unit 103 is generated by the first modified prediction signal generation unit 106, the second modified prediction signal generation unit 107, the unidirectional prediction signal generation unit 117, and the bi-prediction signal generation unit 118 according to a predetermined operation mode. At least one of the obtained motion information is acquired and accumulated. At this time, the association is made based on the mode information, POC or reference information, and the position information of the prediction area that is the current prediction target (hereinafter, the target prediction area), and then stored.

  The motion information accumulation unit 103 supplies at least one of information included in the specified stored motion information in response to a request from each unit. For example, the motion information accumulating unit 103 specifies desired information in the mode information, the POC, and the decodable area in order to specify the motion information belonging to the necessary encoded and decodable area (hereinafter, decodable area). Position information of the prediction area is acquired from the motion prediction unit 104, and at least the identified motion vector information is supplied to the motion prediction unit 104. If necessary, other stored motion information may be supplied. Here, it should be noted that the decodable area includes an area in the reference picture, that is, a reference prediction area and an encoded area in the target picture. Similarly, the motion information accumulating unit 103 corresponds to the mode determination unit 108, the first modified prediction signal generation unit 106, the second modified prediction signal generation unit 107, the unidirectional prediction signal generation unit 117, and the bi-prediction signal generation unit 118. Also, the mode information, the POC, and the position information of the desired prediction area within the decodable area are acquired from these units, and at least one of the specified already stored motion information is supplied as necessary.

  For example, the motion prediction unit 104 generates predicted motion vector information by a predetermined motion prediction process such as AVC, and the reference analysis unit 105, the unidirectional prediction signal generation unit 117, the bi-directional information together with the reference information used in the generation. This is supplied to the prediction signal generation unit 118. Here, in order to acquire motion information belonging to a decodable region that is considered to have high spatial correlation and temporal correlation with respect to the target region, the motion prediction unit 104 The mode information, the POC, and the position information of the desired prediction area within the decodable area are supplied, and the identified motion information is acquired from the motion information storage unit 103. The motion prediction unit 104 performs motion prediction processing using the acquired motion information. When motion information cannot be obtained, predetermined vector information is supplied as predicted motion vector information. In this case, a zero vector is generally used, but another predetermined vector value may be used. In addition, the predetermined motion prediction process has a better configuration if it includes a plurality of different motion prediction processes and can be used by switching them. In such a configuration, the motion prediction mode for specifying the used motion prediction process is managed by, for example, the encoding control unit 120, and motion prediction mode information is generated from the determined motion prediction mode to generate a code string. You may make it supply to the part 113. FIG.

  The reference analysis unit 105 acquires reference information and predicted motion vector information from the motion prediction unit 104. The reference analysis unit 105 performs a predetermined reference analysis process on the acquired reference information, and generates prediction correction information indicating whether the acquired motion vector predictor information can be corrected. Here, the prediction correction information is “1” when correction is possible, and “0” when it is not possible or necessary. The reference analysis unit 105 supplies the acquired reference information, predicted motion vector information, and the generated prediction correction information to the first corrected prediction signal generation unit 106 and the second correction prediction signal generation unit 107. Here, in the predetermined reference analysis processing, for example, the position of the reference picture to be used is specified from the acquired reference information as the condition for whether or not correction is possible, and the positional relationship with the target picture is analyzed. Then, whether correction is possible may be determined based on whether there is a reference picture closer to the target picture. Further, as another condition, there is not only a reference picture that is closer to the target picture, but also a reference picture registered in the reference list to which the reference information belongs is closer to the target picture. Whether or not the correction can be made may be determined based on whether or not it is correct. Furthermore, information regarding image quality may be added to the above-described conditions. For example, based on information that affects the image quality when the reference picture is encoded, the quality is the best, depending on whether there is a reference picture close to the target picture, and the best quality in the reference list, Whether correction is possible may be determined based on whether there is a reference picture close to the target picture. Here, for example, quantization parameter information, a picture type, or the like may be used as information that affects the image quality at the time of encoding.

  The first modified prediction signal generation unit 106 acquires L0 side reference information, prediction motion vector information, and prediction correction information from the reference analysis unit 105. The first modified prediction signal generation unit 106 acquires the reference information and prediction correction information on the L1 side from the reference analysis unit 105 if it can be acquired. The first modified prediction signal generation unit 106 acquires the target signal from the input unit 101 and the reference picture from the picture storage unit 102. Here, the first modified prediction signal generation unit 106 supplies reference information to the picture storage unit 102 when acquiring a reference picture.

  The first modified prediction signal generation unit 106 performs a modified prediction signal generation process based on predetermined information on the L0 side based on each piece of acquired information, and performs reference information, motion vector information, residual motion for each operation mode. Vector information and prediction signals are generated. Further, modified reference information and modified predicted motion vector information are generated according to the operation mode. The first modified prediction signal generation unit 106 supplies the acquired and generated motion information for each operation mode to the motion information storage unit 103. The first modified prediction signal generation unit 106 supplies the reference information or modified reference information used for generation, the generated residual motion vector information, and the prediction signal to the mode determination unit 108 for each operation mode. Details regarding the predetermined modified prediction signal generation processing in the first modified prediction signal generation unit 106 will be described later.

  The second modified prediction signal generation unit 107 acquires L1 side reference information, prediction motion vector information, and prediction correction information from the reference analysis unit 105 if it can be acquired. The second modified prediction signal generation unit 107 acquires L0 side reference information and prediction correction information from the reference analysis unit 105 if it can be acquired. Here, as a difference between the second modified predicted signal generation unit 107 and the first modified predicted signal generation unit 106, it should be noted that the information to be acquired is switched between L0 and L1. In addition, when the information on the L1 side cannot be obtained in the second modified predicted signal generation unit 107, it is preferable not to perform subsequent processing.

  The second modified prediction signal generation unit 107 acquires the target signal from the input unit 101 and the reference picture from the picture storage unit 102. Here, the second modified prediction signal generation unit 107 supplies reference information to the picture storage unit 102 when acquiring a reference picture.

  The second modified prediction signal generation unit 107 performs a modified prediction signal generation process based on information on the predetermined L1 side based on the acquired information, and performs reference information, motion vector information, residual motion for each operation mode. Vector information and prediction signals are generated. Further, modified reference information and modified predicted motion vector information are generated according to the operation mode. The second modified prediction signal generation unit 107 supplies the acquired and generated motion information for each operation mode to the motion information storage unit 103. The second modified prediction signal generation unit 107 supplies the reference information or correction reference information used for generation, the generated residual motion vector information, and the prediction signal to the mode determination unit 108 for each operation mode. Details regarding the predetermined modified prediction signal generation processing in the second modified prediction signal generation unit 107 will be described later.

  For each operation mode, the mode determination unit 108 transmits the reference information or the corrected reference information, the generated residual motion vector information, and the prediction signal to the first corrected prediction signal generation unit 106, the second corrected prediction signal generation unit 107, the one-way Obtained from the prediction signal generation unit 117 and the bi-prediction signal generation unit 118. The mode determination unit 108 acquires a prediction signal from the intra prediction unit 119 if acquisition is possible. Further, the mode determination unit 108 acquires a target signal from the input unit 101.

  The mode determination unit 108 performs predetermined mode determination processing and comparison determination for selecting an operation mode, determines the operation mode, and generates operation mode information. In this mode determination process, the mode determination unit 108 obtains the amount of information required for the acquired information for each operation mode, and compares the information amount for each operation mode. Then, the operation mode with the smallest amount of information is selected and confirmed as that operation mode. In addition, when there are a plurality of predetermined motion prediction processes of the motion prediction unit 104 and can be switched and used, after determining the operation mode for each motion prediction mode, It is preferable that the motion prediction mode to be specified is specified, and the operation mode at that time is finally determined to be the operation mode.

  After determining the mode, the mode determination unit 108 supplies the prediction signal obtained in the determined operation mode to the subtraction unit 109 and the addition unit 112. The mode determination unit 108 supplies the obtained residual motion vector information, reference information or modified reference information, and operation mode information to the code string generation unit 113 according to the determined operation mode.

  The mode determination unit 108 can use the following method as an example of calculating the information amount for each operation mode. The mode determining unit 108 calculates a residual signal by subtracting the prediction signal from the target signal, and outputs a predetermined binary value to the residual signal together with the residual motion vector information, reference information or modified reference information, and operation mode information. After performing the conversion, the amount of the code string obtained by performing the predetermined entropy encoding is set as the information amount for each operation mode. In order to reduce processing, residual motion vector information, reference information or modified reference information, operation mode information, and the like are obtained with a predetermined approximate value, and the residual signal is simply encoded (PCM, DPCM, ADPCM, etc.) or simple frequency conversion (Hadamard transform, Haar transform, etc.) may be used as an approximate value, and the total amount of information may be used for comparison. In order to further reduce the processing, the residual signal itself may be used as an information amount for comparison. The method for obtaining the information amount as described above is merely an example, and the information amount may be obtained by another different method, and is not particularly limited as long as the information amount does not break down. Note that.

  The mode determination unit 108 selects a more likely operation mode based on the predetermined operation mode, the operation mode predicted from the encoded region, or the predetermined look-ahead information, and the one-way prediction signal generation unit 117, the bi-prediction Not required in advance by notifying the encoding control unit 120 not to cause any one of the signal generation unit 118, the intra prediction unit 119, the first modified prediction signal generation unit 106, and the second modified prediction signal generation unit 107 to function. The generation process of the prediction signal or the combined prediction signal in the various operation modes may be suppressed. For example, since the motion information storage unit 103 stores motion modes and motion information that have been encoded and belong to a decodable area, the mode determination unit 108 belongs to an encoded decodable area around the target area. The operation mode and the motion information may be acquired, the operation mode may be selected based on a predetermined operation mode selection process, and a necessary process may be specified and notified to the encoding control unit 120.

  The subtraction unit 109 acquires the target signal from the input unit 101 and the prediction signal from the mode determination unit 108, subtracts the prediction signal from the target signal, generates residual information, and sends the residual information to the residual encoding unit 110. Supply.

  The residual encoding unit 110 acquires residual information from the subtracting unit 109, generates post-residual encoding information by a predetermined residual encoding process, and supplies the residual information to the residual decoding unit 111 and the code string generation unit 113. To do. Here, the predetermined residual encoding process is preferably accompanied by predetermined orthogonal transform and quantization, such as AVC. Further, the residual encoding unit 110 acquires code amount control information from the code amount control unit 115 in order to control the generated code amount, and performs a predetermined residual encoding process according to the code amount control information. The generated code amount of the generated information after residual coding is controlled. In general, by changing the quantization parameter, the generated code amount is changed by controlling the quantization step at the time of quantization.

  The residual decoding unit 111 acquires residual post-encoding information from the residual encoding unit 110, generates decoded residual information by a predetermined residual decoding process, and supplies the decoded residual information to the adding unit 112. Here, the predetermined residual decoding process is preferably accompanied by predetermined inverse quantization and inverse orthogonal transform, such as AVC.

  The adding unit 112 obtains the prediction signal from the mode determination unit 108 and the decoding residual information from the residual decoding unit 111, generates a decoded signal by adding the decoding residual information to the prediction signal, and generates a decoded signal. 102.

  The code string generation unit 113 acquires residual motion vector information, reference information or modified reference information, and operation mode information from the mode determination unit 108, and acquires residual encoded information from the residual encoding unit 110. The code string generation unit 113 acquires, from the encoding control unit 120, an encoding parameter whose operation is determined and that needs to be output. The code string generation unit 113 generates a code string based on the acquired information according to a predetermined syntax structure, and supplies the code string to the code string storage unit 114. The code sequence generation unit 113 may be configured to compress the information amount with a predetermined entropy encoding process when generating the code sequence. Here, the predetermined entropy encoding process may be configured to generate a code string by arithmetic encoding such as CABAC or variable length encoding such as CAVLC, which is an AVC entropy encoding process, for example. Further, different entropy encoding processes such as PCM may be adopted as necessary, or these may be switched and used.

  The code string accumulation unit 114 acquires the code string from the code string generation unit 113 and temporarily accumulates it. The code string storage unit 114 measures the code amount of the temporarily stored code string and supplies the code amount to the code amount control unit 115. The code string accumulation unit 114 supplies the code string temporarily accumulated to the output unit 116.

  The code amount control unit 115 acquires the code amount from the code string storage unit 114, performs code amount control processing based on the acquired code amount, generates code amount control information, and transmits the code amount control information to the residual encoding unit 110. Supply. Here, in the code amount control process, for example, the value of the quantization parameter used in the next encoding is determined based on the acquired code amount so as to be close to the target moving image encoding transmission rate. However, this may be code amount control information.

  The output unit 116 acquires a code string from the code string storage unit 114, and uses the external transmission device 141 and the recording medium 143 connected to the output unit 116, or the recording device 142 or the storage device 144 that does not use the recording medium. Output.

  The encoding control unit 120 controls the operation of each unit included in the moving image encoding apparatus 100 and input / output of information, and manages the entire moving image encoding process. The encoding control unit 120 manages information for controlling and confirming the operation in the moving image encoding apparatus 100 such as the operation mode and the operation order in each unit as a coding parameter in association with each other, and outputs the determined output. The necessary encoding parameters are supplied to the code string generation unit 113. For example, typical coding parameters include a motion prediction mode, an operation mode, reference information, a quantization parameter, an intra prediction mode, and the like. Note that it is included in

  The encoding control unit 120 acquires information on the selected operation mode from the mode determination unit 108, and does not function each unit that is no longer necessary, and performs each operation necessary for realizing the selected operation mode. Alternatively, input / output of operations and information may be controlled to manage the entire moving image encoding process.

  The unidirectional prediction signal generation unit 117 acquires the target signal from the input unit 101 and the reference picture on the L0 side or the L1 side from the picture storage unit 102. Here, the unidirectional prediction signal generation unit 117 supplies the L0 side or L1 side reference information to the picture storage unit 102 when acquiring the reference picture. The unidirectional prediction signal generation unit 117 acquires L0 side or L1 side reference information and prediction motion vector information from the motion prediction unit 104. The unidirectional prediction signal generation unit 117 generates motion vector information, residual motion vector information, and a prediction signal by a predetermined unidirectional prediction signal generation process. Here, the predetermined unidirectional prediction signal generation process specifies motion vector information by a predetermined motion estimation process, for example, AVC, and generates residual motion vector information from the specified motion vector information and predicted motion vector information. Generate. Then, the predetermined unidirectional prediction signal generation process may generate a prediction signal from the reference picture based on the motion vector information specified by the predetermined motion compensation process, for example, as in AVC. The unidirectional prediction signal generation unit 117 supplies the generated motion information to the motion information storage unit 103. The unidirectional prediction signal generation unit 117 supplies the reference information used for generation, the generated residual motion vector information, and the prediction signal to the mode determination unit 108.

  The bi-prediction signal generation unit 118 acquires the target signal from the input unit 101 and the reference pictures on the L0 side and the L1 side from the picture storage unit 102. Here, the bi-prediction signal generation unit 118 supplies the L0 side and L1 side reference information to the picture storage unit 102 when acquiring the reference picture. The bi-prediction signal generation unit 118 acquires L0 side and L1 side reference information and prediction motion vector information from the motion prediction unit 104. The bi-prediction signal generator 118 generates L0 side and L1 side motion vector information, residual motion vector information, and a prediction signal by a predetermined bi-prediction signal generation process. Then, a combined prediction signal is generated by combining the L0 side prediction signal and the L1 side prediction signal. Here, in the predetermined bi-prediction signal generation process, for example, as in AVC, motion vector information is specified by a predetermined motion estimation process on each of the L0 side and the L1 side, and the specified motion vector information and the predicted motion vector are specified. Residual motion vector information is generated from the information. The predetermined bi-prediction signal generation process generates a prediction signal from the reference picture based on the specified motion vector information by the predetermined motion compensation process on each of the L0 side and the L1 side, for example, AVC. Good. The bi-prediction signal generation unit 118 supplies the generated L0 side and L1 side motion information to the motion information storage unit 103. The bi-prediction signal generation unit 118 supplies the L0 side and L1 side reference information used for generation, the generated residual motion vector information, and the combined prediction signal to the mode determination unit 108.

  The intra prediction unit 119 acquires a local decoded picture from the picture storage unit 102, generates a prediction signal in the target region from the local decoded picture based on a predetermined intra prediction process, for example, AVC, and the mode determination unit 108.

  Next, the basic operation of moving picture coding apparatus 100 according to Embodiment 1 will be described below using the flowchart of FIG.

  First, when receiving an encoding start request from a user or the like, the encoding control unit 120 sends an encoding start command to each unit. In response thereto, the input unit 101 acquires an input image sequence, acquires an encoding target picture from the input image sequence (S101), and converts a target signal in the target region into a subtraction unit 109, a mode determination unit 108, and a unidirectional prediction. The signals are supplied to the signal generation unit 117, the first modified prediction signal generation unit 106, the second modified prediction signal generation unit 107, and the bi-prediction signal generation unit 118.

  Thereafter, the encoding control unit 120 sends a motion prediction start command to the motion prediction unit 104. The motion prediction unit 104 receives a command from the encoding control unit 120 and starts a predetermined motion prediction process (S102). In the predetermined motion prediction process, first, the motion prediction unit 104 supplies the motion information storage unit 103 with mode information, POC, and position information of a desired prediction region within the decodable region, and the specified motion Information is acquired from the motion information storage unit 103. Here, various types of information are supplied to the motion information storage unit 103 so as to acquire motion information of at least one area belonging to a decodable area adjacent to the target area, such as AVC, and It is preferable to acquire motion information of the area to be used. Thereafter, the motion prediction unit 104 performs a predetermined motion prediction process such as AVC, for example, using the acquired motion information to generate predicted motion vector information. When motion information cannot be obtained, predetermined vector information is used as predicted motion vector information. Thereafter, the motion prediction unit 104 supplies the predicted motion vector information to the reference analysis unit 105, the unidirectional prediction signal generation unit 117, and the bi-prediction signal generation unit 118 together with the reference information used at the time of generation.

  Thereafter, the encoding control unit 120 sends a reference analysis start command to the reference analysis unit 105. The reference analysis unit 105 receives a command from the encoding control unit 120 and starts a predetermined reference analysis process (S103). In this reference analysis process, first, the reference analysis unit 105 acquires reference information and predicted motion vector information from the motion prediction unit 104. Thereafter, the reference analysis unit 105 performs a predetermined reference analysis process on the acquired reference information to generate prediction correction information. Thereafter, the reference analysis unit 105 supplies the acquired reference information, predicted motion vector information, and the generated prediction correction information to the first corrected prediction signal generation unit 106 and the second correction prediction signal generation unit 107.

  Thereafter, the encoding control unit 120 performs the following operations on the first modified prediction signal generation unit 106, the second modified prediction signal generation unit 107, the unidirectional prediction signal generation unit 117, the bi-prediction signal generation unit 118, and the intra prediction unit 119. Send a prediction signal generation start command. The first modified prediction signal generation unit 106, the second modified prediction signal generation unit 107, the unidirectional prediction signal generation unit 117, the bi-prediction signal generation unit 118, and the intra prediction unit 119 receive an instruction from the encoding control unit 120, A predetermined prediction signal generation process in each unit is started (S104). In this prediction signal generation process, a first modified prediction signal generation process (S104-1), a second modified prediction signal generation process (S104-2), a unidirectional prediction signal generation process (S104-3), and a bi-prediction signal generation process (S104-4) and intra prediction processing (S104-5) are performed. Each unit supplies the acquired and generated motion information for each operation mode to the motion information storage unit 103, and the reference information or the corrected reference information used for generation, the generated residual motion vector information, and the prediction signal for each operation mode. To the mode determination unit 108.

  Here, the first modified predicted signal generation process described above is performed by the first modified predicted signal generation unit 106. First, the first modified prediction signal generation unit 106 acquires L0-side reference information, prediction motion vector information, and prediction correction information from the reference analysis unit 105. The first modified prediction signal generation unit 106 acquires the reference information and prediction correction information on the L1 side from the reference analysis unit 105 if it can be acquired. The first modified prediction signal generation unit 106 acquires the target signal from the input unit 101 and the reference picture from the picture storage unit 102. Here, the first modified prediction signal generation unit 106 supplies reference information to the picture storage unit 102 when acquiring a reference picture. After that, the first modified prediction signal generation unit 106 performs a modified prediction signal generation process based on the information on the predetermined L0 side based on the acquired information, and performs reference information, motion vector information, and remaining information for each operation mode. Difference motion vector information and a prediction signal are generated. Further, modified reference information and modified predicted motion vector information are generated according to the operation mode. The first modified prediction signal generation unit 106 supplies the acquired and generated motion information for each operation mode to the motion information storage unit 103. The first modified prediction signal generation unit 106 supplies the reference information or modified reference information used for generation, the generated residual motion vector information, and the prediction signal to the mode determination unit 108 for each operation mode. Thereby, the first modified prediction signal generation process is completed.

  In addition, the second modified predicted signal generation process described above is performed by the second modified predicted signal generation unit 107. First, the second corrected prediction signal generation unit 107 acquires L1 side reference information, prediction motion vector information, and prediction correction information from the reference analysis unit 105 if it can be acquired. The second modified prediction signal generation unit 107 acquires L0 side reference information and prediction correction information from the reference analysis unit 105 if it can be acquired. The second modified prediction signal generation unit 107 acquires the target signal from the input unit 101 and the reference picture from the picture storage unit 102. Here, the second modified prediction signal generation unit 107 supplies reference information to the picture storage unit 102 when acquiring a reference picture. Thereafter, the second modified predicted signal generation unit 107 performs a modified predicted signal generation process based on the information on the predetermined L1 side based on the acquired information, and performs reference information, motion vector information, and remaining information for each operation mode. Difference motion vector information and a prediction signal are generated. Further, modified reference information and modified predicted motion vector information are generated according to the operation mode. The second modified prediction signal generation unit 107 supplies the acquired and generated motion information for each operation mode to the motion information storage unit 103. The second modified prediction signal generation unit 107 supplies the reference information or correction reference information used for generation, the generated residual motion vector information, and the prediction signal to the mode determination unit 108 for each operation mode. Thereby, the second modified prediction signal generation process is completed.

  Further, the above-described unidirectional prediction signal generation process is performed by the unidirectional prediction signal generation unit 117. First, the unidirectional prediction signal generation unit 117 acquires the target signal from the input unit 101 and the L0 side or L1 side reference picture from the picture storage unit 102. Here, the unidirectional prediction signal generation unit 117 supplies the L0 side or L1 side reference information to the picture storage unit 102 when acquiring the reference picture. The unidirectional prediction signal generation unit 117 acquires L0 side or L1 side reference information and prediction motion vector information from the motion prediction unit 104. Thereafter, the unidirectional prediction signal generation unit 117 generates motion vector information, residual motion vector information, and a prediction signal by a predetermined unidirectional prediction signal generation process. Thereafter, the unidirectional prediction signal generation unit 117 supplies the generated motion information to the motion information storage unit 103. The unidirectional prediction signal generation unit 117 supplies the reference information used for generation, the generated residual motion vector information, and the prediction signal to the mode determination unit 108. Thereby, the one-way predicted signal generation process is completed.

  The bi-predictive signal generation process described above is performed by the bi-predictive signal generation unit 118. First, the bi-prediction signal generation unit 118 acquires the target signal from the input unit 101 and the reference pictures on the L0 side and the L1 side from the picture storage unit 102. Here, the bi-prediction signal generation unit 118 supplies the L0 side and L1 side reference information to the picture storage unit 102 when acquiring the reference picture. The bi-prediction signal generation unit 118 acquires L0 side and L1 side reference information and prediction motion vector information from the motion prediction unit 104. Thereafter, the bi-prediction signal generation unit 118 generates L0 side and L1 side motion vector information, residual motion vector information, and a prediction signal by a predetermined bi-prediction signal generation process, and the L0 side prediction signal and the L1 side prediction are generated. By combining the signals, a combined prediction signal is generated. Thereafter, the bi-prediction signal generation unit 118 supplies the generated L0 side and L1 side motion information to the motion information storage unit 103. The bi-prediction signal generation unit 118 supplies the L0 side and L1 side reference information used for generation, the generated residual motion vector information, and the combined prediction signal to the mode determination unit 108. Thereby, the bi-prediction signal generation process is completed.

  The intra prediction process described above is performed by the intra prediction unit 119. First, the intra prediction unit 119 acquires a locally decoded picture from the picture storage unit 102, generates a prediction signal in the target region from the locally decoded picture based on a predetermined intra prediction process, and supplies the prediction signal to the mode determination unit 108 To do. Thereby, the intra prediction process is completed.

  When the prediction signal generation process as described above is completed, the encoding control unit 120 sends a mode determination start command to the mode determination unit 108. The mode determination unit 108 starts the mode determination process (S105) in response to an instruction from the encoding control unit 120. In this mode determination process, first, the mode determination unit 108 uses, for each operation mode, the reference information or the corrected reference information, the generated residual motion vector information, and the prediction signal as the first corrected prediction signal generation unit 106 and the second correction signal. Obtained from the prediction signal generation unit 107, the unidirectional prediction signal generation unit 117, and the bi-prediction signal generation unit 118, respectively. The mode determination unit 108 acquires a prediction signal from the intra prediction unit 119 if acquisition is possible. The mode determination unit 108 acquires a target signal from the input unit 101. Thereafter, the mode determination unit 108 performs a comparison determination for determining the operation mode. When the operation mode is confirmed (YES in S106), operation mode information is generated. Thereafter, the mode determination unit 108 supplies the prediction signal obtained in the determined operation mode to the subtraction unit 109 and the addition unit 112. The mode determination unit 108 supplies the obtained residual motion vector information, reference information or modified reference information, and operation mode information to the code string generation unit 113 according to the determined operation mode. If the operation mode has not been determined (NO in S106), the process returns to step S102.

  Thereafter, the encoding control unit 120 sends a residual signal generation start command to the subtraction unit 109. The subtraction unit 109 receives a command from the encoding control unit 120 and starts a predetermined residual signal generation process (S107). In this predetermined residual signal generation process, the subtracting unit 109 acquires the target signal from the input unit 101 and the prediction signal from the mode determination unit 108, and generates residual information by subtracting the prediction signal from the target signal. And supplied to the residual encoding unit 110.

  Thereafter, the encoding control unit 120 sends a residual encoding start command to the residual encoding unit 110. Residual encoding unit 110 receives a command from encoding control unit 120 and starts a predetermined residual encoding process (S108). In this predetermined residual encoding process, the residual encoding unit 110 acquires residual information from the subtracting unit 109, generates post-residual encoding information by the predetermined residual encoding process, and performs residual decoding. To the unit 111 and the code string generation unit 113.

  Thereafter, the encoding control unit 120 sends a residual decoding start command to the residual decoding unit 111. Residual decoding section 111 receives a command from encoding control section 120 and starts a predetermined residual decoding process (S109). In this predetermined residual decoding process, the residual decoding unit 111 obtains post-residual encoding information from the residual encoding unit 110, generates decoded residual information by the predetermined residual decoding process, and adds an adding unit. 112.

  Thereafter, the encoding control unit 120 sends a decoded signal generation start command to the adding unit 112. In response to the instruction from the encoding control unit 120, the adding unit 112 starts a predetermined decoded signal generation process (S110). In this predetermined decoded signal generation process, the adding unit 112 acquires the prediction signal from the mode determination unit 108 and the decoding residual information from the residual decoding unit 111, and adds the decoding residual information to the prediction signal, thereby adding the decoded signal. And is supplied to the picture storage unit 102.

  The encoding control unit 120 sends a code string generation start command to the code string generation unit 113. The code string generation unit 113 receives a command from the coding control unit 120 and starts a predetermined code string generation process (S111). In this predetermined code sequence generation process, the code sequence generation unit 113 acquires residual motion vector information, reference information or modified reference information, and operation mode information from the mode determination unit 108 and from the residual encoding unit 110. Obtain post-differential information. The code string generation unit 113 acquires, from the encoding control unit 120, an encoding parameter whose operation is determined and that needs to be output. The code string generation unit 113 generates a code string based on the acquired information according to a predetermined syntax structure, and supplies the code string to the code string storage unit 114.

  Thereafter, the encoding control unit 120 sends a code amount control start command to the code amount control unit 115. The code amount control unit 115 receives a command from the coding control unit 120 and starts a predetermined code amount control process (S112). In the predetermined code amount control process, the code amount control unit 115 generates the code amount control information by acquiring the code amount from the code string storage unit 114 and performing the code amount control process based on the acquired code amount. And then supplied to the residual encoding unit 110.

  Thereafter, the encoding control unit 120 checks whether or not the encoding of the entire encoding target picture has been completed based on the progress of the encoding process (S113). If encoding of the entire encoding target picture has not been completed (NO in S113), the encoding control unit 120 updates the target area to the next target area, and returns to step S102.

  When encoding of the entire encoding target picture is completed (YES in S113), the encoding control unit 120 sends an output start command to the code string storage unit 114 and the output unit 116. The code string accumulation unit 114 receives a command from the encoding control unit 120 and supplies the code string to the output unit 116, and the output unit 116 acquires the code string from the code string accumulation unit 114 and sends it to the output unit 116. The data is output to the recording device 142 or the storage device 144 that uses or does not use the external transmission device 141 and recording medium 143 that are connected (S114).

  Through the steps as described above, the basic operation of the moving picture coding apparatus 100 according to Embodiment 1 is completed.

  Here, the following effects are acquired by the structure and operation | movement of the moving image encoder 100 of Embodiment 1. FIG.

  First, in the video encoding device 100 according to the first embodiment, when at least one prediction motion vector information can be acquired from the motion prediction unit 104, the other prediction motion vector is obtained from this one prediction motion vector information. Information can be derived as modified prediction motion vector information. At the same time, the motion vector information can be determined by correcting the corrected predicted motion vector information to a better position as a reference destination by a predetermined motion search. By synthesizing the prediction signal based on the determined motion vector information and the prediction signal based on the prediction motion vector information first obtained from the motion prediction unit 104, a combined prediction signal can be generated. Also, residual motion vector information can be obtained by subtracting the derived modified prediction motion vector information from the confirmed motion vector information. Further, by transmitting operation mode information for specifying such an operation to the decoding side, a similar combined prediction signal can be obtained on the decoding side. Therefore, the moving picture coding apparatus 100 according to Embodiment 1 can perform bi-prediction similar to the coding side on the decoding side only by transmitting one residual motion vector information. In this regard, conventionally, when performing bi-prediction, it is necessary to transmit residual motion vector information for two lines.

  In addition, when bi-prediction is performed without using residual motion vector information as in the conventional spatial and temporal direct mode, the quality of the prediction signal is reduced, and as a result, the code amount is accurately determined. In some cases, it was not possible to suppress it. In this regard, since the moving picture coding apparatus 100 according to Embodiment 1 can correct the motion vector information at a more probable position by using residual motion vector information for one line, better prediction is possible. The possibility that a signal or a combined prediction signal can be generated can be increased. Therefore, the entire code amount when describing the motion vector information and the prediction signal can be suppressed as compared with the conventional case.

  Further, in the moving picture encoding apparatus 100 according to Embodiment 1, for example, a one-way prediction signal generation unit 117, a bi-prediction signal generation unit 118, and an intra prediction unit 119, which are conventional prediction signal generation units such as AVC, The first modified prediction signal generation unit 106 and the second modification prediction signal generation unit 107, which are new components, are provided. The first modified prediction signal generation unit 106 derives the L1 side modified prediction motion vector information based on the L0 side motion information to generate a prediction signal or a combined prediction signal, and the second modified prediction signal generation unit 107 Based on the L1 side motion information, the L0 side modified prediction motion vector information is derived to generate a prediction signal or a combined prediction signal. According to this configuration, compared with a configuration including only a conventional prediction signal generation unit, the amount of code when encoding a prediction signal or a combined prediction signal, residual motion vector information, and reference information obtained in each operation mode However, there is a possibility that a smaller operation mode can be selected. In this case, it is possible to generate a code string having a code amount smaller than the conventional code amount.

  Further, the moving picture encoding apparatus 100 according to Embodiment 1 selects a more probable operation mode based on a predetermined operation mode, an operation mode predicted from an encoded region, or predetermined look-ahead information, Any of the direction prediction signal generation unit 117, the bi-prediction signal generation unit 118, the intra prediction unit 119, the first modified prediction signal generation unit 106, and the second modified prediction signal generation unit 107 can be prevented from functioning. As a result, the moving picture encoding apparatus 100 according to Embodiment 1 can reduce the amount of calculation in the process until the generation of the prediction signal, and can reduce the time required for encoding.

  Also, in the video encoding device 100 according to Embodiment 1, when the reference information predicted by the motion prediction unit 104 is changed to the reference information by the reference analysis unit 105, the reference destination is changed. If the reference information is used as the corrected reference information, or if the corrected reference information can be uniquely identified on the decoding side based on a predetermined rule or operation mode, the reference information or the corrected reference information itself is encoded and It can be configured not to transmit. Thereby, in the moving image coding apparatus 100 of Embodiment 1, the code amount regarding motion information can further be suppressed and coding efficiency can be improved.

  Next, the moving picture decoding apparatus according to the first embodiment that decodes the code string generated by the moving picture encoding apparatus 100 according to the first embodiment will be described. FIG. 4 is a configuration diagram of the moving picture decoding apparatus 200 according to Embodiment 1 of the present invention. FIG. 5 is a configuration diagram illustrating a connection relationship between each unit of the video decoding device 200 and the decoding control unit 216 according to Embodiment 1 of the present invention.

  The moving picture decoding apparatus 200 according to Embodiment 1 includes an input unit 201, a picture storage unit 202, a decoded sequence generation unit 203, a residual decoding unit 204, an addition unit 205, a mode identification unit 206, a motion information storage unit 207, and motion prediction. Unit 208, reference analysis unit 209, first modified prediction signal generation unit 210, second modified prediction signal generation unit 211, output unit 212, and decoding control unit 216. Furthermore, it may be configured to include a unidirectional prediction signal generation unit 213, a bi-prediction signal generation unit 214, and an intra prediction unit 215. In the first embodiment, the description will be made assuming that all of these are included.

  The input unit 201 acquires a code string from, for example, a recording device 232 or a storage device 234 that uses or does not use the transmission device 231 and the recording medium 233 as illustrated in FIG. 4 and supplies the code sequence to the decoded sequence generation unit 203. To do. Here, the acquired code string is preferably generated by the moving image encoding apparatus 100 according to Embodiment 1, for example. Further, the acquired code string may be generated and transmitted or stored by another having an equivalent function.

  The picture storage unit 202 has a function equivalent to that of the picture storage unit 102 of the video encoding device 100. The picture accumulation unit 202 accumulates pictures that are known not to be used as reference pictures as decoded pictures. The picture storage unit 202 supplies the reference picture or decoded picture to be displayed to the output unit 212 according to the display timing.

  The decoded sequence generation unit 203 acquires a code sequence from the input unit 201, performs a decoded sequence generation process based on a predetermined syntax structure, and converts the code sequence into a region that is currently a decoding target (hereinafter, a decoding target region). At least residual motion vector information, reference information, operation mode information, information after residual encoding, and encoding parameters are generated. Thereafter, the decoded sequence generation unit 203 supplies residual motion vector information, reference information, and operation mode information to the mode identification unit 206, supplies residual encoded information to the residual decoding unit 204, and sets encoding parameters. The decoding parameters are supplied to the decoding control unit 216, respectively. The decoding sequence generation unit 203 may be configured to perform decoding with a predetermined entropy decoding process that is paired with a predetermined entropy encoding process of the video encoding device 100 when generating the decoded sequence. Here, the predetermined entropy decoding process may be configured to generate a decoded string by arithmetic decoding such as CABAC or variable length decoding such as CAVLC, which is an AVC entropy decoding process, for example. Further, different entropy decoding processes such as PCM may be adopted as necessary, or these may be switched and used.

  The residual decoding unit 204 acquires residual post-encoding information from the decoded sequence generation unit 203, generates decoded residual information by a predetermined residual decoding process, and supplies the decoded residual information to the addition unit 205. Here, the predetermined residual decoding process is preferably accompanied by predetermined inverse quantization and inverse orthogonal transform, such as AVC.

  The adding unit 205 obtains a prediction signal from the mode identification unit 206 and decoding residual information from the residual decoding unit 204, generates a decoded signal by adding the decoding residual information to the prediction signal, and generates a decoded signal. 202.

  The mode identification unit 206 acquires residual motion vector information, reference information, and operation mode information belonging to the decoding target region from the decoded sequence generation unit 203. The mode identifying unit 206 identifies the acquired operation mode information, and notifies the decoding control unit 216 to operate each unit according to the identified operation mode. The mode identification unit 206 supplies the acquired motion information to the motion information storage unit 207 together with the operation mode information. The mode identifying unit 206 is selected from the first modified prediction signal generation unit 210, the second modified prediction signal generation unit 211, the unidirectional prediction signal generation unit 213, the bi-prediction signal generation unit 214, and the intra prediction unit 215 according to the operation mode. The acquisition source is specified, and the prediction signal is acquired from the acquisition source. Thereafter, the mode identification unit 206 supplies the acquired prediction signal to the addition unit 205.

  The motion information storage unit 207 has a function equivalent to that of the motion information storage unit 103 of the video encoding device 100. The motion information storage unit 207 acquires the operation mode information and the motion information from the mode identification unit 206, and stores the motion information. The motion information storage unit 207 selects a supply destination according to the operation mode from the first modified prediction signal generation unit 210, the second modified prediction signal generation unit 211, the unidirectional prediction signal generation unit 213, and the bi-prediction signal generation unit 214. The residual motion vector information is supplied to the supply destination.

  The motion prediction unit 208 and the reference analysis unit 209 have functions equivalent to those of the motion prediction unit 104 and the reference analysis unit 105 of the video encoding device 100, and thus description thereof is omitted.

  The first modified prediction signal generation unit 210 acquires L0-side reference information, prediction motion vector information, and prediction correction information from the reference analysis unit 209 according to the operation mode. The first corrected prediction signal generation unit 210 acquires the reference information and prediction correction information on the L1 side from the reference analysis unit 209 if it can be acquired. The first modified prediction signal generation unit 210 acquires a reference picture from the picture storage unit 102. Here, the first modified prediction signal generation unit 210 supplies reference information to the picture storage unit 202 when acquiring a reference picture. The first modified prediction signal generation unit 210 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207.

  The first modified predictive signal generation unit 210 performs a modified predictive signal generation process based on information obtained on the basis of predetermined information on the L0 side, and includes reference information, motion vector information, and predictive signals according to the operation mode. Is generated. Further, modified reference information and modified predicted motion vector information are generated according to the operation mode. The first modified prediction signal generation unit 210 supplies motion information corresponding to the acquired and generated operation mode to the motion information storage unit 207. The first modified prediction signal generation unit 210 supplies the generated prediction signal to the mode identification unit 206. Details regarding the predetermined modified prediction signal generation process in the first modified prediction signal generation unit 210 will be described later.

  The second modified prediction signal generation unit 211 acquires reference information, prediction motion vector information, and prediction correction information on the L1 side from the reference analysis unit 209, if it can be acquired, according to the operation mode. The second modified prediction signal generation unit 211 acquires the reference information and prediction correction information on the L0 side from the reference analysis unit 209 according to the operation mode if it can be acquired. The second modified prediction signal generation unit 211 acquires a reference picture from the picture storage unit 202. Here, the second modified prediction signal generation unit 211 supplies reference information to the picture storage unit 202 when acquiring a reference picture. The second modified prediction signal generation unit 211 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207.

  The second modified prediction signal generation unit 211 performs a modified prediction signal generation process based on information on the predetermined L1 side based on each acquired information, and includes reference information, motion vector information, and a prediction signal corresponding to the operation mode. Is generated. Further, modified reference information and modified predicted motion vector information are generated according to the operation mode. The second modified prediction signal generation unit 211 supplies motion information corresponding to the acquired and generated operation mode to the motion information storage unit 207. The second modified prediction signal generation unit 211 supplies the generated prediction signal to the mode identification unit 206. Details regarding the predetermined modified prediction signal generation process in the second modified prediction signal generation unit 211 will be described later.

  The output unit 212 acquires a reference picture or decoded picture to be displayed from the picture storage unit 202 according to the display timing, and uses an external transmission device 241 and a recording medium 243 connected to the output unit 212. The decoded picture is output by supplying to the recording device 242, the storage device 244, the display device 245, or the like that is used or not used.

  The decoding control unit 216 controls the operation of each unit included in the moving image decoding apparatus 200 and the input / output of information, and manages the entire moving image decoding process. The decoding control unit 216 obtains, from the decoded sequence generation unit 203, decoding parameters associated with each other, which are information for controlling and determining the operation in the video decoding device 200, such as the operation mode and the operation order in each unit. And manages the entire moving picture decoding process. The decoding control unit 216 controls the operation of each unit included in the moving image decoding apparatus 200 and the input / output of information so that each unit executes an operation according to the operation mode identified by the mode identifying unit 206.

  The unidirectional prediction signal generation unit 213 specifies which reference picture to be acquired from the L0 side or the L1 side from the operation mode, and acquires the specified reference picture from the picture storage unit 202. Here, the unidirectional prediction signal generation unit 213 supplies the L0 side or L1 side reference information to the picture storage unit 202 when acquiring a reference picture. The unidirectional prediction signal generation unit 213 acquires the specified L0 side or L1 side reference information and prediction motion vector information from the motion prediction unit 208. The unidirectional prediction signal generation unit 213 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207. The unidirectional prediction signal generation unit 213 generates motion vector information and a prediction signal by a predetermined unidirectional prediction signal generation process. Here, as predetermined unidirectional prediction signal generation processing, for example, as in AVC, motion vector information is generated by adding residual motion vector information to the acquired prediction motion vector information, and is generated by predetermined motion compensation processing. A prediction signal may be generated from the reference picture based on the motion vector information. The unidirectional prediction signal generation unit 213 supplies the generated motion information to the motion information storage unit 207. The unidirectional prediction signal generation unit 213 supplies the generated prediction signal to the mode identification unit 206.

  The bi-prediction signal generation unit 214 acquires the L0 side and L1 side reference pictures from the picture storage unit 202 based on the operation mode. Here, the bi-prediction signal generation unit 214 supplies the L0 side and L1 side reference information to the picture storage unit 202 when acquiring the reference picture. The bi-prediction signal generation unit 214 acquires L0 side and L1 side reference information and prediction motion vector information from the motion prediction unit 208. The bi-prediction signal generation unit 214 acquires, from the motion information accumulation unit 207, residual motion vector information on the L0 side and the L1 side that belong to the decoding target region.

  The bi-prediction signal generation unit 214 generates motion vector information and prediction signals on the L0 side and the L1 side by a predetermined bi-prediction signal generation process, and synthesizes the prediction signal on the L0 side and the prediction signal on the L1 side. Then, a combined prediction signal is generated. Here, the predetermined bi-prediction signal generation process generates motion vector information by adding the residual motion vector information to the acquired predicted motion vector information corresponding to the L0 side and the L1 side, respectively, such as AVC. Based on the generated L0 side and L1 side motion vector information by predetermined motion compensation, L0 side and L1 side prediction signals may be generated from the reference pictures corresponding to the L0 side and L1 side, respectively. . The bi-prediction signal generation unit 214 supplies the generated motion information on the L0 side and the L1 side to the motion information storage unit 207. The bi-prediction signal generation unit 214 supplies the generated L0 side and L1 side prediction signals to the mode identification unit 206.

  The intra prediction unit 215 obtains a local decoded picture from the picture storage unit 202 according to the operation mode, and predicts the decoding within the decoding target region from the local decoded picture based on a predetermined intra prediction process such as AVC, for example. A signal is generated and supplied to the mode identification unit 206.

  Next, the basic operation of the moving picture decoding apparatus 200 according to Embodiment 1 will be described below using the flowchart of FIG.

  First, when receiving a decoding start request from a user or the like, the decoding control unit 216 sends a decoding start command to each unit. In response thereto, the input unit 201 acquires a code string (S201) and supplies the code string to the decoded string generation unit 203.

  Thereafter, the decoding control unit 216 sends a decoded sequence generation start command to the decoded sequence generation unit 203. The decoded sequence generation unit 203 receives a command from the decoding control unit 216 and starts a predetermined decoded sequence generation process (S202). In this predetermined decoded sequence generation process, first, the decoded sequence generation unit 203 acquires a code sequence from the input unit 201, performs a decoded sequence generation process based on a predetermined syntax structure, and at least a residual motion vector from the code sequence. Information, reference information, operation mode information, information after residual encoding, and encoding parameters are generated. Thereafter, the decoded sequence generation unit 203 supplies residual motion vector information, reference information, and operation mode information to the mode identification unit 206, supplies residual encoded information to the residual decoding unit 204, and sets encoding parameters. The decoding parameters are supplied to the decoding control unit 216, respectively.

  Thereafter, the decoding control unit 216 sends a residual decoding start command to the residual decoding unit 204. Residual decoding unit 204 receives a command from decoding control unit 216 and starts a predetermined residual decoding process (S203). In this predetermined residual decoding process, first, the residual decoding unit 204 obtains post-residue encoded information from the decoded sequence generation unit 203, generates decoded residual information by a predetermined residual decoding process, and adds To the unit 205.

  Next, the decoding control unit 216 sends a mode identification process start command to the mode identification unit 206. The mode identification unit 206 receives a command from the decoding control unit 216 and starts a predetermined mode identification process (S204). In this predetermined mode identification process, the mode identification unit 206 acquires residual motion vector information, reference information, and operation mode information belonging to the decoding target region from the decoded sequence generation unit 203. The mode identifying unit 206 identifies the acquired operation mode information, and notifies the decoding control unit 216 to operate each unit according to the identified operation mode. The mode identification unit 206 supplies the acquired motion information to the motion information storage unit 207 together with the operation mode information.

  Next, the decoding control unit 216 sends a motion prediction start command to the motion prediction unit 208. The motion prediction unit 208 receives a command from the decoding control unit 216 and starts a predetermined motion prediction process (S205). In this predetermined motion prediction process, first, the motion prediction unit 208 generates predicted motion vector information by a predetermined motion prediction process such as AVC, for example, and the reference analysis unit 105 together with the reference information used in the generation. The supply destination according to the operation mode is identified from the one-way prediction signal generation unit 117 and the bi-prediction signal generation unit 118, and the supply destination is supplied to the supply destination. When the motion prediction mode can be acquired, predicted motion vector information is generated based on a predetermined motion prediction process specified from the motion prediction mode.

  Thereafter, when the decoding control unit 216 determines from the operation mode that the first modified prediction signal generation process or the second modified prediction signal generation process is necessary in the prediction signal generation process, Send a reference analysis processing start command. The reference analysis unit 209 receives a command from the decoding control unit 216 and starts a predetermined reference analysis process (S206). If it is determined from the operation mode that the first modified predicted signal generation process and the second modified predicted signal generation process are not necessary in the predicted signal generation process, step S206 is skipped. In this predetermined reference analysis process, the reference analysis unit 209 acquires reference information and prediction motion vector information belonging to the decoding target region from the motion prediction unit 208. The reference analysis unit 209 performs predetermined reference analysis processing on the acquired reference information, and generates prediction correction information indicating whether the acquired prediction motion vector information can be corrected according to the operation mode. Thereafter, the reference analysis unit 209 converts the acquired reference information and predicted motion vector information, and the prediction correction information generated according to the operation mode, to the first corrected prediction signal generation unit 210 or the second correction prediction according to the operation mode. The signal is supplied to the signal generator 211.

  Thereafter, the decoding control unit 216 operates from the first modified prediction signal generation unit 210, the second modified prediction signal generation unit 211, the unidirectional prediction signal generation unit 213, the bi-prediction signal generation unit 214, and the intra prediction unit 215. A prediction signal generation start command is sent to what is specified by the mode. When the first modified prediction signal generation unit 210, the second modified prediction signal generation unit 211, the unidirectional prediction signal generation unit 213, the bi-prediction signal generation unit 214, and the intra prediction unit 215 receive a command from the decoding control unit 216, Then, predetermined prediction signal generation processing (S207) in each unit is started. In this predicted signal generation process, a first modified predicted signal generation process (S207-1), a second modified predicted signal generation process (S207-2), a unidirectional predicted signal generation process (S207-3), and a bi-predicted signal generation process. In (S207-4) and the intra prediction process (S207-5), the process specified by the operation mode is performed. Each unit supplies the acquired and generated motion information to the motion information storage unit 207, and supplies the generated prediction signal to the mode identification unit 206.

  Here, the first modified prediction signal generation process described above is performed by the first modified prediction signal generation unit 210. First, the first modified prediction signal generation unit 210 acquires L0-side reference information, prediction motion vector information, and prediction correction information from the reference analysis unit 209 according to the operation mode. The first corrected prediction signal generation unit 210 acquires the reference information and prediction correction information on the L1 side from the reference analysis unit 209 if it can be acquired. The first modified prediction signal generation unit 210 acquires a reference picture from the picture storage unit 102. Here, the first modified prediction signal generation unit 210 supplies reference information to the picture storage unit 202 when acquiring a reference picture. The first modified prediction signal generation unit 210 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207.

  The first modified predictive signal generation unit 210 performs a modified predictive signal generation process based on information obtained on the basis of predetermined information on the L0 side, and includes reference information, motion vector information, and predictive signals according to the operation mode. Is generated. Further, modified reference information and modified predicted motion vector information are generated according to the operation mode. The first modified prediction signal generation unit 210 supplies motion information corresponding to the acquired and generated operation mode to the motion information storage unit 207. The first modified prediction signal generation unit 210 supplies the generated prediction signal to the mode identification unit 206. Thereby, the first modified prediction signal generation process is completed.

  The second modified predicted signal generation process described above is performed by the second modified predicted signal generation unit 211. First, the second modified prediction signal generation unit 211 acquires reference information, prediction motion vector information, and prediction correction information on the L1 side from the reference analysis unit 209, if it can be acquired, according to the operation mode. The second modified prediction signal generation unit 211 acquires the reference information and prediction correction information on the L0 side from the reference analysis unit 209 according to the operation mode if it can be acquired. The second modified prediction signal generation unit 211 acquires a reference picture from the picture storage unit 202. Here, the second modified prediction signal generation unit 211 supplies reference information to the picture storage unit 202 when acquiring a reference picture. The second modified prediction signal generation unit 211 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207.

  The second modified predictive signal generation unit 211 performs a modified predictive signal generation process based on predetermined information on the L1 side based on each acquired information, and reference information, motion vector information, and predictive signal according to the operation mode. Is generated. Further, modified reference information and modified predicted motion vector information are generated according to the operation mode. The second modified prediction signal generation unit 211 supplies motion information corresponding to the acquired and generated operation mode to the motion information storage unit 207. The second modified prediction signal generation unit 211 supplies the generated prediction signal to the mode identification unit 206. Thereby, the second modified prediction signal generation process is completed.

  In addition, the above-described unidirectional prediction signal generation process is performed by the unidirectional prediction signal generation unit 213. First, the unidirectional prediction signal generation unit 213 specifies which reference picture to be acquired from the L0 side or the L1 side from the operation mode, and acquires the specified reference picture from the picture storage unit 202. Here, the unidirectional prediction signal generation unit 213 supplies the L0 side or L1 side reference information to the picture storage unit 202 when acquiring a reference picture. The unidirectional prediction signal generation unit 213 acquires the specified L0 side or L1 side reference information and prediction motion vector information from the motion prediction unit 208. The unidirectional prediction signal generation unit 213 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207. The unidirectional prediction signal generation unit 213 generates motion vector information and a prediction signal by a predetermined unidirectional prediction signal generation process. The unidirectional prediction signal generation unit 213 supplies the generated motion information to the motion information storage unit 207. The unidirectional prediction signal generation unit 213 supplies the generated prediction signal to the mode identification unit 206. Thereby, the one-way predicted signal generation process is completed.

  The bi-predictive signal generation process described above is performed by the bi-predictive signal generation unit 214. First, the bi-prediction signal generation unit 214 acquires the L0 side and L1 side reference pictures from the picture storage unit 202 based on the operation mode. Here, the bi-prediction signal generation unit 214 supplies the L0 side and L1 side reference information to the picture storage unit 202 when acquiring the reference picture. The bi-prediction signal generation unit 214 acquires L0 side and L1 side reference information and prediction motion vector information from the motion prediction unit 208. The bi-prediction signal generation unit 214 acquires, from the motion information accumulation unit 207, residual motion vector information on the L0 side and the L1 side that belong to the decoding target region.

  The bi-prediction signal generation unit 214 generates motion vector information and prediction signals on the L0 side and the L1 side by a predetermined bi-prediction signal generation process, and synthesizes the prediction signal on the L0 side and the prediction signal on the L1 side. Then, a combined prediction signal is generated. The bi-prediction signal generation unit 214 supplies the generated motion information on the L0 side and the L1 side to the motion information storage unit 207. The bi-prediction signal generation unit 214 supplies the generated L0 side and L1 side prediction signals to the mode identification unit 206. Thereby, the bi-prediction signal generation process is completed.

  The intra prediction process described above is performed by the intra prediction unit 215. First, the intra prediction unit 215 obtains a locally decoded picture from the picture storage unit 202, and generates a prediction signal in the decoding target region from the locally decoded picture based on a predetermined intra prediction process, for example, AVC. , And supplied to the mode identification unit 206. Thereby, the intra prediction process is completed.

  When the prediction signal generation process as described above is completed, the decoding control unit 216 sends a decoded signal generation start command to the mode identification unit 206 and the addition unit 205. In response to the instruction from the decoding control unit 216, the mode identifying unit 206 and the adding unit 205 start the decoded signal generation process (S208). First, the mode identifying unit 206 supplies a prediction signal corresponding to the acquired operation mode to the adding unit 205. Thereafter, the adding unit 205 obtains a prediction signal from the mode identification unit 206 and decoding residual information from the residual decoding unit 204, generates a decoded signal by adding the decoding residual information to the prediction signal, The data is supplied to the storage unit 202.

  After that, the decoding control unit 216 confirms whether or not the decoding of the entire decoded picture to be decoded has been completed from the progress of the decoding process (S209). If the decoding of the entire decoded picture has not been completed (NO in S209), the decoding control unit 216 updates the decoding target area to the next decoding target area, and returns to step S202.

  When the decoding of the entire decoded picture is completed (YES in S209), the decoding control unit 216 sends an output start command to the output unit 212. The output unit 212 receives a command from the decoding control unit 216, acquires a reference picture or decoded picture to be displayed from the picture storage unit 202, and transmits an external transmission connected to the output unit 212 as a decoded image. The data is output to the recording device 142, the storage device 144, or the like that uses or does not use the device 241 and the recording medium 243 (S210).

  Through the steps as described above, the basic operation of the moving picture decoding apparatus 200 according to Embodiment 1 is completed.

  Here, the following effects are obtained by the configuration and operation of the moving picture decoding apparatus 200 of the first embodiment.

  First, in the video decoding device 200 according to the first embodiment, the code sequence generated by the video encoding device 100 according to the first embodiment is acquired, and the decoded sequence generation unit 203 acquires at least a remaining part belonging to the decoding target region. Difference motion vector information, reference information, operation mode information, information after residual encoding, and encoding parameters can be generated and stored in the motion information storage unit 207. The motion information storage unit 207 can store the operation mode and motion information belonging to the decoded area. From such a motion information storage unit 207, the motion prediction unit 208 acquires motion information belonging to the decoding target region and motion information belonging to the region around the decoding target region according to the operation mode, and performs predetermined processing according to the operation mode. With this motion prediction process, at least one piece of predicted motion vector information can be acquired. Thereby, according to the operation mode, the other predicted motion vector information can be derived as the corrected predicted motion vector information from the one predicted motion vector information. At the same time, motion vector information can be generated by adding residual motion vector information to the corrected predicted motion vector information. By synthesizing the prediction signal based on the determined motion vector information and the prediction signal based on the prediction motion vector information first obtained from the motion prediction unit 208, a combined prediction signal can be generated. As described above, in the moving picture decoding apparatus 200 according to the first embodiment, the motion vector information corresponding to the operation mode and the prediction signal or the same as the moving picture coding apparatus 100 according to the first embodiment are obtained from the acquired code string. A synthesized prediction signal can be obtained, and the code string generated by the moving picture coding apparatus 100 according to Embodiment 1 can be correctly decoded.

  In addition, in moving picture decoding apparatus 200 according to Embodiment 1, according to the operation mode, a configuration for deriving as corrected predicted motion vector information, a configuration for generating motion vector information by adding residual motion vector information, and a prediction signal or A configuration for generating a combined prediction signal is provided. Thereby, the other prediction motion vector information and motion vector information which cannot be acquired conventionally can be obtained. As a result, when bi-prediction is performed without using the residual motion vector information as in the conventional spatial and temporal direct mode, the quality of the prediction signal is reduced, resulting in the code amount being There was a case where it increased. In this regard, the moving picture decoding apparatus 200 according to Embodiment 1 decodes one residual motion vector information from a code string having a code amount equal to or less than the code amount according to the conventional method, and a motion vector at a more probable position. Since the information can be modified to be information, a better prediction signal or a combined prediction signal can be generated from a code amount equal to or less than the code amount according to the conventional method.

  Further, in the moving picture decoding apparatus 200 according to Embodiment 1, for example, a one-way prediction signal generation unit 213, a bi-prediction signal generation unit 214, an intra prediction unit 215, which are conventional prediction signal generation units such as AVC, A first modified prediction signal generation unit 210 and a second modification prediction signal generation unit 211, which are new components, are provided. The first modified prediction signal generation unit 210 derives the L1 side modified prediction motion vector information based on the L0 side motion information and generates a prediction signal or a combined prediction signal. The second modified prediction signal generation unit 211 derives the L0 side corrected prediction motion vector information based on the L1 side motion information and generates a prediction signal or a combined prediction signal. According to this configuration, it is possible to correctly decode the code string generated by the moving picture coding apparatus 100 according to Embodiment 1 according to the operation mode. In this way, a code string encoded with a code amount that could not be achieved with the configuration including only the conventional prediction signal generation unit is obtained, and each unit necessary for decoding is made to function according to the operation mode. The prediction signal or the combined prediction signal corresponding to the operation mode can be generated and correctly decoded.

(Embodiment 2) Detailed configuration and operation of moving picture coding apparatus and moving picture decoding apparatus.
Next, the first modified prediction signal generation unit 106 and the second modified prediction signal generation unit 107 in the moving image coding device 100 of the first embodiment are configured in more detail as the moving image coding device of the second embodiment. It shows what was realized by. Detailed configurations of the first modified predicted signal generation unit 106 and the second modified predicted signal generation unit 107 in the second embodiment will be described with reference to FIGS. 7A and 7B.

  The first modified prediction signal generation unit 106 in FIG. 7A is, for example, a first L0 one-side prediction signal generation unit 106-1, a second L0 one-side prediction signal generation unit 106-2, and a first L0-L1 one-side prediction signal generation unit 106. -3, the second L0-L1 one-side prediction signal generation unit 106-4, the first L0-L1 bi-prediction signal generation unit 106-5, the second L0-L1 bi-prediction signal generation unit 106-6, and the third L0-L1 bi-prediction signal generation Of the unit 106-7 and the fourth L0-L1 bi-prediction signal generation unit 106-8, it may be configured by at least one, two or more arbitrary combinations, or all. Here, the discussion will be made assuming that everything is used. Here, L0-L1 attached to the names of the parts indicates that the L1 side motion vector information is derived based on the L0 side reference information or the corrected reference information and the predicted motion vector information. ing.

  Similarly, the second modified prediction signal generation unit 107 in FIG. 7B includes, for example, a first L1 one-side prediction signal generation unit 107-1, a second L1 one-side prediction signal generation unit 107-2, and a first L1-L0 one-side prediction signal. Generation unit 107-3, second L1-L0 one-side prediction signal generation unit 107-4, first L1-L0 bi-prediction signal generation unit 107-5, second L1-L0 bi-prediction signal generation unit 107-6, and third L1-L0 bi-prediction unit 107-6. Of the prediction signal generation unit 107-7 and the fourth L1-L0 bi-prediction signal generation unit 107-8, at least one, any combination of two or more, or all may be used. Here, the discussion will be made assuming that everything is used. Here, L1-L0 given to the names of the parts indicates that the L0 side motion vector information is derived based on the L1 side reference information or the corrected reference information and the predicted motion vector information. ing.

  FIGS. 8A to 8C and FIGS. 9A to 9C are configuration diagrams showing details of each part in FIGS. 7A and 7B. FIG. 8A is a configuration diagram of the first L0 one-side prediction signal generation unit 106-1 and the second L0 one-side prediction signal generation unit 106-2. FIG. 8B is a configuration diagram of the first L0-L1 one-side prediction signal generation unit 106-3 and the second L0-L1 one-side prediction signal generation unit 106-4. FIG. 8C shows the first L0-L1 bi-prediction signal generation unit 106-5, the second L0-L1 bi-prediction signal generation unit 106-6, the third L0-L1 bi-prediction signal generation unit 106-7, and the fourth L0-L1. It is a block diagram of the bi prediction signal production | generation part 106-8.

  Similarly, FIG. 9A is a configuration diagram of the first L1 one-side prediction signal generation unit 107-1 and the second L1 one-side prediction signal generation unit 107-2. FIG. 9B is a configuration diagram of the first L1-L0 one-side prediction signal generation unit 107-3 and the second L1-L0 one-side prediction signal generation unit 107-4. FIG. 9C illustrates the first L1-L0 bi-predictive signal generation unit 107-5, the second L1-L0 bi-predictive signal generation unit 107-6, the third L1-L0 bi-predictive signal generation unit 107-7, and the fourth L1-L0. It is a block diagram of the bi prediction signal production | generation part 107-8.

  The configuration of FIG. 8A is a configuration for generating L0 side motion vector information and generating an L0 side prediction signal based on the L0 side motion information. 8A includes an L0 side motion information acquisition unit 801, a reference destination change unit 802, an L0 side motion derivation unit 803, a reference picture acquisition unit 804, a target picture acquisition unit 805, an L0 side motion search unit 806, mvd. A generation unit 807 and a one-side prediction signal generation unit 808 are included.

  The L0 side motion information acquisition unit 801 performs a predetermined L0 side motion information acquisition process. In this predetermined L0 side motion information acquisition process, the L0 side motion information acquisition unit 801 receives, from the reference analysis unit 105, L0 side reference information, predicted motion vector information, and prediction correction as the L0 side motion information belonging to the target region. Get information. The L0 side motion information acquisition unit 801 supplies the acquired L0 side motion information to the reference destination changing unit 802.

  The reference destination changing unit 802 performs a predetermined reference destination changing process. In the predetermined reference destination changing process, the reference destination changing unit 802 acquires the L0 side motion information from the L0 side motion information acquiring unit 801. When the predicted correction information on the L0 side is “1”, the reference destination changing unit 802 corrects the reference information by a predetermined reference destination changing process, and generates corrected reference information. Here, it is preferable that the predetermined reference destination changing process does not function according to the operation mode, and the reference information is changed to the corrected reference information. For example, the first L0 one-side predicted signal generation unit 106-1 may use the reference information as the corrected reference information without causing the predetermined reference destination changing process to function. As the operation mode, the first L0 one-side predicted signal generation unit 106-1 does not correct the L0-side predicted motion vector information with the L0-side corrected reference information, but generates L0-side motion vector information and generates an L0-side predicted signal. Operate in the first L0 one-sided prediction mode. On the other hand, the second L0 one-side predicted signal generation unit 106-2 may function the predetermined reference destination changing process. The second L0 one-side predicted signal generation unit 106-2 corrects the L0-side predicted motion vector information with the L0-side corrected reference information as the operation mode, generates the L0-side motion vector information, and generates the L0-side predicted signal. It operates in the second L0 one-sided prediction mode as generated. The reference destination changing unit 802 supplies the acquired and generated motion information to the L0 side motion deriving unit 803 and the reference picture acquiring unit 804.

The L0 side motion deriving unit 803 performs a predetermined L0 side motion deriving process. In the predetermined L0 side motion deriving process, the L0 side motion deriving unit 803 acquires the L0 side motion information from the reference destination changing unit 802. Thereafter, the L0 side motion deriving unit 803 is identified from the reference information and the modified reference information on the L0 side, the time interval td_l0 between the target picture and the reference picture identified from the reference information, and the target picture and the modified reference information. A time interval m_td_l0 with respect to the reference picture is obtained. When the predicted motion vector information is mvp_10 and the corrected predicted motion vector information is m_mvp_10,
m_mvp_l0 = mvp_l0 × m_td_l0 / td_l0
... (Formula 1)
By doing so, the L0 side modified prediction motion vector information is derived. The L0 side motion deriving unit 803 supplies the acquired and generated motion information to the L0 side motion searching unit 806.

  The reference picture acquisition unit 804 performs a predetermined reference picture acquisition process. In this predetermined reference picture acquisition process, the reference picture acquisition unit 804 acquires corrected reference information as motion information from the reference destination change unit 802. The reference picture acquisition unit 804 supplies the L0 side modified reference information to the picture storage unit 102, and acquires the identified reference picture. The reference picture acquisition unit 804 supplies the acquired reference picture to the L0 side motion search unit 806 and the one side prediction signal generation unit 808.

  The target picture acquisition unit 805 performs a predetermined target picture acquisition process. In the predetermined target picture acquisition process, the target picture acquisition unit 805 acquires a target signal from the input unit 101 and supplies the target signal to the L0 side motion search unit 806.

  The L0 side motion search unit 806 performs a predetermined L0 side motion search process. In the predetermined L0 side motion search processing, the L0 side motion search unit 806 receives the L0 side motion information from the L0 side motion deriving unit 803, the reference picture from the reference picture acquisition unit 804, and the target signal from the target picture acquisition unit 805, respectively. get. The L0 side motion search unit 806 performs a predetermined motion search process such as AVC, for example, by using the acquired corrected predicted motion vector information, and generates motion vector information. The L0 side motion search unit 806 supplies the acquired and generated motion information to the mvd generation unit 807.

  The mvd generation unit 807 performs predetermined mvd generation processing. In this predetermined mvd generation process, the mvd generation unit 807 acquires motion information from the L0 side motion search unit 806, subtracts the L0 side corrected predicted motion vector information from the acquired L0 side motion vector information, and outputs the L0 side. Residual motion vector information is generated. The mvd generation unit 807 supplies the acquired and generated motion information to the one-side prediction signal generation unit 808.

  The one-side prediction signal generation unit 808 performs a predetermined one-side prediction signal generation process. In the predetermined one-side prediction signal generation process, the one-side prediction signal generation unit 808 acquires motion information from the mvd generation unit 807 and a reference picture from the reference picture acquisition unit 804. The one-side prediction signal generation unit 808 generates a prediction signal by a predetermined motion compensation process such as AVC based on the acquired motion vector information. The one-side prediction signal generation unit 808 supplies operation mode information, motion information, and a prediction signal for specifying the operation mode to the mode determination unit 108. The one-side prediction signal generation unit 808 supplies the acquired and generated motion information to the motion information storage unit 103 together with the operation mode information.

  The configuration of FIG. 8B is a configuration for generating L1 side motion vector information and generating an L1 side prediction signal based on the L0 side motion information. 8B includes an L0 side motion information acquisition unit 821, an L1 side reference information acquisition unit 822, a reference destination change unit 823, an L1 side motion derivation unit 824, a reference picture acquisition unit 825, a target picture acquisition unit 826, An L1-side motion search unit 827, an mvd generation unit 828, and a one-side prediction signal generation unit 829 are included.

  The L0 side motion information acquisition unit 821 performs a predetermined L0 side motion information acquisition process. In this predetermined L0 side motion information acquisition process, the L0 side motion information acquisition unit 821 receives, from the reference analysis unit 105, L0 side reference information, predicted motion vector information, and prediction correction as the L0 side motion information belonging to the target region. Get information. The L0 side motion information acquisition unit 821 supplies the acquired L0 side motion information to the reference destination changing unit 823.

  The L1 side reference information acquisition unit 822 performs a predetermined L1 side reference information acquisition process. In this predetermined L1 side reference information acquisition process, the L1 side reference information acquisition unit 822 acquires reference information and prediction correction information on the L1 side from the reference analysis unit 105, if possible, and a reference destination change unit 823. To supply.

  The reference destination changing unit 823 performs a predetermined reference destination changing process. In this predetermined reference destination changing process, the reference destination changing unit 823 acquires the L0 side motion information from the L0 side motion information acquisition unit 821 and the L1 side reference information acquisition unit 822 from the L1 side reference information and the prediction correction information. . When the L1 side prediction correction information is “1”, the reference destination changing unit 823 corrects the reference information by a predetermined reference destination changing process, and generates corrected reference information. Here, it is preferable that the predetermined reference destination changing process does not function according to the operation mode, and the reference information is changed to the corrected reference information. For example, the first L0-L1 one-sided prediction signal generation unit 106-3 may not use this predetermined reference destination changing process, and may make the reference information the corrected reference information. The first L0-L1 one-side predicted signal generation unit 106-3 corrects the L0-side predicted motion vector information based on the reference information instead of the L1-side corrected reference information as the operation mode, and the L1-side motion vector information. It operates in a first L0-L1 one-sided prediction mode that generates a prediction signal on the L1 side. Further, the second L0-L1 one-side predicted signal generation unit 106-4 may function the predetermined reference destination changing process. The second L0-L1 one-side predicted signal generation unit 106-4 corrects the L0-side predicted motion vector information based on the L1-side corrected reference information as the operation mode, generates L1-side motion vector information, and generates L1 Operate in a second L0-L1 one-sided prediction mode that generates a side prediction signal. The reference destination changing unit 823 supplies the acquired and generated L0 side and L1 side motion information to the L1 side motion deriving unit 824 and the reference picture acquiring unit 825.

The L1 side motion deriving unit 824 performs a predetermined L1 side motion deriving process. In this predetermined L1 side motion deriving process, the L1 side motion deriving unit 824 acquires the L0 side and L1 side motion information from the reference destination changing unit 823. After that, when operating in the first L0-L1 one-sided prediction mode as the operation mode, the L1 side motion deriving unit 824 determines, from the reference information on the L0 side, the reference picture specified from the target picture and the reference information on the L0 side. The time interval td_l1 between the current picture and the reference picture specified from the reference information on the L1 side is obtained from the time interval td_l0 between them and the reference information on the L1 side. When the predicted motion vector information is mvp_l0 and the corrected predicted motion vector information is m_mvp_l1,
m_mvp_l1 = mvp_l0 × td_l1 / td_l0
... (Formula 2)
Thus, the L1 side modified predicted motion vector information is derived from the L0 side predicted motion vector information.

Further, when operating in the second L0-L1 one-sided prediction mode as the operation mode, the L1 side motion deriving unit 824 determines whether the reference picture specified from the target picture and the reference information on the L0 side is obtained from the reference information on the L0 side. The time interval m_td_l1 between the current picture and the reference picture specified from the modified reference information on the L1 side is obtained from the time interval td_l0 between them and the reference information on the L1 side. When the predicted motion vector information is mvp_l0 and the corrected predicted motion vector information is m_mvp_l1,
m_mvp_l1 = mvp_l0 × m_td_l1 / td_l0
... (Formula 3)
Thus, the L1 side corrected predicted motion vector information is derived from the L0 side predicted motion vector information based on the L1 side corrected reference information. Here, when the reference information on the L1 side and the corrected reference information on the L1 side are the same, the corrected predicted motion vector information on the L1 side may be derived by the above (Equation 3). By doing in this way, it becomes possible to unify a process. The L1 side motion deriving unit 824 supplies the acquired and generated L0 side and L1 side motion information to the L1 side motion search unit 827.

  The reference picture acquisition unit 825 performs a predetermined reference picture acquisition process. In this predetermined reference picture acquisition process, the reference picture acquisition unit 825 acquires the corrected reference information from the reference destination change unit 823 as the motion information on the L1 side. The reference picture acquisition unit 825 supplies the L1 side modified reference information to the picture storage unit 102 and acquires the identified reference picture. The reference picture acquisition unit 825 supplies the acquired reference picture to the L1 side motion search unit 827 and the one side prediction signal generation unit 829.

  The target picture acquisition unit 826 performs a predetermined target picture acquisition process. In the predetermined target picture acquisition process, the target picture acquisition unit 826 acquires a target signal from the input unit 101 and supplies the target signal to the L1 side motion search unit 827.

  The L1 side motion search unit 827 performs a predetermined L1 side motion search process. In this predetermined L1 side motion search process, the L1 side motion search unit 827 receives the motion information on the L0 side and the L1 side from the L1 side motion derivation unit 824, the reference picture from the reference picture acquisition unit 825, and the reference picture from the target picture acquisition unit 826. Each target signal is acquired. The L1 side motion search unit 827 performs a predetermined motion search process such as AVC, for example, using the acquired L1 side corrected predicted motion vector information, and generates L1 side motion vector information. The L1 side motion search unit 827 supplies the acquired and generated L0 side and L1 side motion information to the mvd generation unit 828.

  The mvd generation unit 828 performs predetermined mvd generation processing. In this predetermined mvd generation process, the mvd generation unit 828 acquires the L0 side and L1 side motion information from the L1 side motion search unit 827, and the L1 side modified predicted motion vector information from the acquired L1 side motion vector information. Is subtracted to generate residual motion vector information on the L1 side. The mvd generation unit 828 supplies the acquired and generated L0 side and L1 side motion information to the one-side prediction signal generation unit 829.

  The one-side prediction signal generation unit 829 performs a predetermined one-side prediction signal generation process. In the predetermined one-side prediction signal generation processing, the one-side prediction signal generation unit 829 acquires the L0 side and L1 side motion information from the mvd generation unit 828 and the reference picture from the reference picture acquisition unit 825. The one-side prediction signal generation unit 829 generates a prediction signal by a predetermined motion compensation process such as AVC, based on the acquired L1 side motion vector information. The one-side prediction signal generation unit 829 supplies the mode determination unit 108 with operation mode information, motion information, and a prediction signal for specifying the operation mode. The one-side prediction signal generation unit 829 supplies the acquired and generated L0 side and L1 side motion information to the motion information storage unit 103 together with the operation mode information.

  The configuration of FIG. 8C is based on the L0 side motion information, generates L1 side motion vector information, generates a L0 side prediction signal and a L1 side prediction signal, and then becomes a bi-prediction signal. This is a configuration for generating a combined prediction signal. 8C includes an L0 side motion information acquisition unit 841, an L1 side reference information acquisition unit 842, a reference destination change unit 843, an L0 side motion derivation unit 844, an L1 side motion derivation unit 845, and a reference picture acquisition unit 846. A target picture acquisition unit 847, an L1 side motion search unit 848, an mvd generation unit 849, and a bi-prediction signal generation unit 850.

  The L0 side motion information acquisition unit 841 performs a predetermined L0 side motion information acquisition process. In this predetermined L0 side motion information acquisition processing, the L0 side motion information acquisition unit 841 receives, from the reference analysis unit 105, L0 side reference information, predicted motion vector information, and prediction correction as the L0 side motion information belonging to the target region. Get information. The L0 side motion information acquisition unit 841 supplies the acquired L0 side motion information to the reference destination changing unit 843.

  The L1 side reference information acquisition unit 842 performs a predetermined L1 side reference information acquisition process. In this predetermined L1 side reference information acquisition process, the L1 side reference information acquisition unit 842 acquires L1 side reference information and prediction correction information from the reference analysis unit 105, if possible, and a reference destination change unit 843. To supply.

  The reference destination changing unit 843 performs a predetermined reference destination changing process. In this predetermined reference destination changing process, the reference destination changing unit 843 acquires the L0 side motion information from the L0 side motion information acquisition unit 821 and the L1 side reference information and prediction correction information from the L1 side reference information acquisition unit 822. . When the prediction correction information on the L0 side is “1”, the reference destination changing unit 843 corrects the reference information on the L0 side by a predetermined reference destination changing process, and generates correction reference information on the L0 side.

  Similarly, when the prediction correction information on the L1 side is “1”, the reference destination changing unit 843 corrects the reference information on the L1 side by a predetermined reference destination changing process, and changes the reference reference information on the L1 side. Generate. Here, it is preferable that the predetermined reference destination changing process does not function according to the operation mode, and the reference information is changed to the corrected reference information. For example, the first L0-L1 bi-prediction signal generation unit 106-5 may use the reference information as corrected reference information without causing the predetermined reference destination changing process to function. The first L0-L1 bi-prediction signal generator 106-5 uses the reference information on the L0 side and the reference information on the L1 side as the operation mode, and applies the reference information on the L1 side to the predicted motion vector information on the L0 side. Thus, the corrected predicted motion vector information on the L1 side is derived, and the prediction signal on the L1 side is generated from the reference information and the predicted motion vector information on the L0 side, and the reference information and the corrected predicted motion vector on the L1 side are generated. It operates in the first L0-L1 bi-prediction mode in which a prediction signal on the L1 side is generated from information and then a combined prediction signal is generated.

  The second L0-L1 bi-prediction signal generation unit 106-6 may function the predetermined reference destination changing process. The second L0-L1 bi-prediction signal generation unit 106-6 uses the reference information on the L0 side and the corrected reference information on the L1 side as the operation mode, and the corrected reference information on the L1 side with respect to the predicted motion vector information on the L0 side. Is used to derive the corrected predicted motion vector information on the L1 side, generate the L0 side prediction signal from the reference information and the predicted motion vector information on the L0 side, and the corrected reference information and the correction on the L1 side An operation is performed in the second L0-L1 bi-prediction mode in which a prediction signal on the L1 side is generated from the prediction motion vector information and then a combined prediction signal is generated.

  In addition, the third L0-L1 bi-prediction signal generation unit 106-7 may function the predetermined reference destination changing process. The third L0-L1 bi-prediction signal generation unit 106-7 uses the modified reference information on the L0 side and the reference information on the L1 side as the operation mode, and the modified reference information on the L0 side with respect to the predicted motion vector information on the L0 side. L0 side corrected motion vector information is derived by applying and corrected, and further, L1 side reference information is applied to the L0 side predicted motion vector information and corrected. Predicted motion vector information is generated, and an L0 side prediction signal is generated from the L0 side corrected reference information and the corrected predicted motion vector information, and an L1 side prediction signal is generated from the L1 side reference information and the corrected predicted motion vector information. Then, it operates in the third L0-L1 bi-prediction mode that generates a composite prediction signal.

  Further, the fourth L0-L1 bi-predictive signal generation unit 106-8 may function the predetermined reference destination changing process. The fourth L0-L1 bi-prediction signal generation unit 106-8 uses the modified reference information on the L0 side and the modified reference information on the L1 side as the operation mode, and the modified reference on the L0 side with respect to the predicted motion vector information on the L0 side. The corrected predicted motion vector information on the L0 side is derived by applying and correcting the information, and further, the corrected reference vector information on the L1 side is applied to the predicted motion vector information on the L0 side and corrected. And the L0 side prediction signal and the corrected prediction motion vector information are generated from the L0 side correction reference information and the corrected prediction motion vector information, and the L1 side prediction signal is corrected from the L1 side correction reference information and the corrected prediction motion vector information. It operates in the 4L0-L1 bi-prediction mode that generates a combined prediction signal after generating a signal.

  The reference destination changing unit 843 stores the acquired and generated L0 side motion information in the L0 side motion deriving unit 844, and acquires and generates the acquired L0 side and L1 side motion information in the L1 side motion deriving unit 845 and the reference picture acquiring unit 846. To supply.

  The L0 side motion deriving unit 844 performs a predetermined L0 side motion deriving process. In the predetermined L0 side motion deriving process, the L0 side motion deriving unit 844 acquires the L0 side motion information from the reference destination changing unit 843. After that, when operating in the first L0-L1 bi-prediction mode or the second L0-L1 bi-prediction mode as the operation mode, the L0 side motion deriving unit 844 does not cause the predetermined L0 side motion deriving process to function. The predicted motion vector information may be modified predicted motion vector information.

Further, when operating in the third L0-L1 bi-prediction mode or the fourth L0-L1 bi-prediction mode as the operation mode, the L0 side motion deriving unit 844 uses the reference information on the L0 side and the reference information on the L0 side from the reference information on the L0 side. The time interval m_td_10 between the target picture and the reference picture specified from the modified reference information on the L0 side is obtained from the time interval td_10 between the reference picture identified from the reference picture and the modified reference information on the L0 side. When the L0 side predicted motion vector information is mvp_10 and the L0 side predicted motion vector information is m_mvp_10,
m_mvp_l0 = mvp_l0 × m_td_l0 / td_l0
... (Formula 4)
By doing so, the L0-side corrected predicted motion vector information is derived from the L0-side predicted motion vector information. Here, when the reference information on the L0 side and the corrected reference information on the L0 side are the same, the corrected predicted motion vector information on the L0 side may be derived by the above (Equation 4). By doing in this way, it becomes possible to unify a process. The L0 side motion deriving unit 844 supplies the acquired and generated L0 side motion information to the bi-prediction signal generation unit 850.

The L1 side motion deriving unit 845 performs a predetermined L1 side motion deriving process. In the predetermined L1 side motion deriving process, the L1 side motion deriving unit 845 acquires the L0 side and L1 side motion information from the reference destination changing unit 843. After that, when operating in the first L0-L1 bi-prediction mode or the third L0-L1 bi-prediction mode as the operation mode, the L1 side motion deriving unit 845 uses the target picture and the L0 side reference information from the L0 side reference information. The time interval td_l1 between the target picture and the reference picture specified from the reference information on the L1 side is obtained from the time interval td_10 between the reference picture specified from the reference picture and the reference information on the L1 side. When the predicted motion vector information on the L0 side is mvp_l0 and the modified predicted motion vector information on the L1 side is m_mvp_l1,
m_mvp_l1 = mvp_l0 × td_l1 / td_l0
... (Formula 5)
Thus, the L1 side modified predicted motion vector information is derived from the L0 side predicted motion vector information.

In addition, when operating in the second L0-L1 bi-prediction mode or the fourth L0-L1 bi-prediction mode as the operation mode, the L1 side motion deriving unit 845 uses the target picture and the L0 side reference information from the L0 side reference information. The time interval m_td_l1 between the target picture and the reference picture specified from the modified reference information on the L1 side is obtained from the time interval td_10 between the reference picture specified from the reference picture and the reference information on the L1 side. When the predicted motion vector information is mvp_l0 and the corrected predicted motion vector information is m_mvp_l1,
m_mvp_l1 = mvp_l0 × m_td_l1 / td_l0
... (Formula 6)
Thus, the L1 side corrected predicted motion vector information is derived from the L0 side predicted motion vector information based on the L1 side corrected reference information. Here, when the reference information on the L1 side and the corrected reference information on the L1 side are the same, the corrected predicted motion vector information on the L1 side may be derived by the above (Equation 6). By doing in this way, it becomes possible to unify a process. The L1 side motion deriving unit 845 supplies the acquired and generated L0 side and L1 side motion information to the L1 side motion search unit 848.

  The reference picture acquisition unit 846 performs a predetermined reference picture acquisition process. In this predetermined reference picture acquisition process, the reference picture acquisition unit 846 acquires the corrected reference information from the reference destination change unit 843 as the motion information on the L0 side and the L1 side. The reference picture acquisition unit 846 supplies the L0 side and L1 side modified reference information to the picture storage unit 102, and acquires the identified L0 side and L1 side reference pictures. The reference picture acquisition unit 846 supplies the acquired L1 side reference picture to the L1 side motion search unit 827. The reference picture acquisition unit 846 supplies the acquired L0 side and L1 side reference pictures to the bi-prediction signal generation unit 850.

  The target picture acquisition unit 847 performs a predetermined target picture acquisition process. In the predetermined target picture acquisition process, the target picture acquisition unit 847 acquires the target signal from the input unit 101 and supplies the target signal to the L1 side motion search unit 848.

  The L1 side motion search unit 848 performs a predetermined L1 side motion search process. In this predetermined L1 side motion search process, the L1 side motion search unit 848 receives the L0 side and L1 side motion information from the L1 side motion derivation unit 845, the reference picture from the reference picture acquisition unit 846, and the target picture acquisition unit 847. Each target signal is acquired. The L1 side motion search unit 848 performs predetermined motion search processing such as AVC, for example, using the acquired L1 side corrected predicted motion vector information, and generates L1 side motion vector information. The L1 side motion search unit 848 supplies the acquired and generated motion information on the L0 side and the L1 side to the mvd generation unit 849.

  The mvd generation unit 849 performs a predetermined mvd generation process. In this predetermined mvd generation process, the mvd generation unit 849 acquires the L0 side and L1 side motion information from the L1 side motion search unit 848, and the L1 side corrected predicted motion vector information from the acquired L1 side motion vector information. Is subtracted to generate residual motion vector information on the L1 side. The mvd generation unit 849 supplies the acquired and generated motion information on the L0 side and the L1 side to the bi-prediction signal generation unit 850.

  The bi-prediction signal generation unit 850 performs a predetermined bi-prediction signal generation process. In this predetermined bi-prediction signal generation process, the bi-prediction signal generation unit 850 acquires the L0 side and L1 side motion information from the mvd generation unit 828, and the L0 side and L1 side reference pictures from the reference picture acquisition unit 825. The bi-prediction signal generation unit 850 generates a prediction signal on the L0 side by a predetermined motion compensation process such as AVC based on the acquired corrected prediction motion vector information on the L0 side. Further, the bi-prediction signal generation unit 850 generates an L1 side prediction signal by a predetermined motion compensation process such as AVC, based on the acquired L1 side motion vector information. Thereafter, the bi-prediction signal generation unit 850 generates a combined prediction signal by combining the generated L0-side prediction signal and the L1-side prediction signal, thereby obtaining a bi-prediction signal. The bi-prediction signal generation unit 850 supplies the mode determination unit 108 with the operation mode information for specifying the operation mode, the acquired and generated L0 side and L1 side motion information, and the combined prediction signal. Also, the bi-prediction signal generation unit 850 supplies the acquired and generated motion information on the L0 side and the L1 side to the motion information storage unit 103 together with the operation mode information.

  In the configuration of FIG. 8A, the L0 side motion vector information is generated based on the L0 side motion information, and the L0 side prediction signal is generated. On the other hand, in the configuration of FIG. 9A, the handling of the L0 side and the L1 side is reversed, and the configuration for generating the prediction signal on the L1 side is based on the motion information on the L1 side. It is different from the configuration of (A). The details of each part in the configuration of FIG. 9A are omitted because the functions on the L0 side and the L1 side are considered to be reversed to have the same function as the configuration of FIG. 8A. .

  In the configuration of FIG. 8B, the L1 side motion vector information is generated based on the L0 side motion information, and the L1 side prediction signal is generated. On the other hand, in the configuration of FIG. 9B, the L0 side motion information is generated based on the L1 side motion information by reversing the handling of the L0 side and the L1 side, and the L0 side prediction signal is generated. This is different from the configuration shown in FIG. The details of each part in the configuration of FIG. 9B are omitted because the functions on the L0 side and the L1 side are considered to be reversed to have the same function as the configuration of FIG. 8B. .

  In the configuration of FIG. 8C, the L1 side motion vector information is generated based on the L0 side motion information, the L0 side prediction signal and the L1 side prediction signal are generated, and then the bi-prediction signal is obtained. The configuration is for generating a synthesized prediction signal. On the other hand, the configuration of FIG. 9C reverses the handling of the L0 side and the L1 side, generates the L0 side motion vector information based on the L1 side motion information, and generates the L0 side prediction signal and the L1 side. 8C is different from the configuration of FIG. 8C in that a configuration for generating a combined prediction signal to be a bi-prediction signal after generating a prediction signal on the side is generated. The details of each part in the configuration in FIG. 9C are omitted because the functions on the L0 side and the L1 side are considered to be reversed and the functions are the same as those in the configuration in FIG. .

  Next, the operation in the configuration of FIG. 8A will be described with reference to the flowchart of FIG.

  First, upon receiving a prediction signal generation start command, the L0 side motion information acquisition unit 801 receives, from the reference analysis unit 105, L0 side reference information, prediction motion vector information, and prediction correction information as L0 side motion information belonging to the target region. Is acquired (S301). Thereafter, the L0 side motion information acquisition unit 801 supplies the acquired L0 side motion information to the reference destination changing unit 802.

  The reference destination changing unit 802 acquires the L0 side motion information from the L0 side motion information acquisition unit 801. When the predicted correction information on the L0 side is “1”, the reference destination changing unit 802 corrects the reference information by a predetermined reference destination changing process, and generates corrected reference information. In this way, the reference destination on the L0 side is changed (S302). Here, in the case of the first L0 one-sided prediction mode as the operation mode, the reference destination changing process is not allowed to function, and the reference information may be the corrected reference information. Further, in the case of the second L0 one-side prediction mode as the operation mode, the reference destination changing process is preferably caused to function. Thereafter, the reference destination changing unit 802 supplies the acquired and generated motion information to the L0 side motion deriving unit 803 and the reference picture acquiring unit 804.

  Next, the L0 side motion deriving unit 803 acquires the L0 side motion information from the reference destination changing unit 802. After that, the L0 side motion deriving unit 803 performs a predetermined L0 side motion deriving process to derive L0 side modified predicted motion vector information (S303). Thereafter, the L0 side motion deriving unit 803 supplies the acquired and generated motion information to the L0 side motion searching unit 806.

  Thereafter, the reference picture acquisition unit 804 acquires corrected reference information as motion information from the reference destination change unit 802. The reference picture acquisition unit 804 supplies the L0-side modified reference information to the picture storage unit 102, and acquires the identified reference picture (S304). The reference picture acquisition unit 804 supplies the acquired reference picture to the L0 side motion search unit 806 and the one side prediction signal generation unit 808.

  Thereafter, the target picture acquisition unit 805 acquires a target signal from the input unit 101 (S305) and supplies the target signal to the L0 side motion search unit 806.

  Thereafter, the L0 side motion search unit 806 acquires the L0 side motion information from the L0 side motion deriving unit 803, the reference picture from the reference picture acquisition unit 804, and the target signal from the target picture acquisition unit 805, respectively. The L0 side motion search unit 806 performs the L0 side motion search by performing a predetermined motion search process such as AVC, for example, by using the obtained corrected predicted motion vector information, and determines the motion vector information ( S306). Thereafter, the L0 side motion search unit 806 supplies the acquired and generated motion information to the mvd generation unit 807.

  Next, the mvd generation unit 807 obtains motion information from the L0 side motion search unit 806, subtracts the L0 side modified predicted motion vector information from the obtained L0 side motion vector information, and obtains an L0 side residual motion vector. Information is generated (S307). The mvd generation unit 807 supplies the acquired and generated motion information to the one-side prediction signal generation unit 808.

  Thereafter, the one-side prediction signal generation unit 808 acquires motion information from the mvd generation unit 807 and a reference picture from the reference picture acquisition unit 804. The one-side prediction signal generation unit 808 generates a prediction signal on the L0 side by a predetermined motion compensation process such as AVC based on the acquired motion vector information (S308). Thereafter, the one-side prediction signal generation unit 808 supplies operation mode information, motion information, and a prediction signal for specifying the operation mode to the mode determination unit 108. In addition, the one-side prediction signal generation unit 808 supplies the acquired and generated motion information to the motion information storage unit 103 together with the operation mode information.

  Through the above steps, the operation in the configuration of FIG.

  Next, the operation in the configuration of FIG. 8B will be described with reference to the flowchart of FIG.

  First, upon receiving a prediction signal generation start command, the L0 side motion information acquisition unit 821 receives L0 side reference information, prediction motion vector information, and prediction correction information from the reference analysis unit 105 as L0 side motion information belonging to the target region. Is acquired (S401). Thereafter, the L0 side motion information acquisition unit 821 supplies the acquired L0 side motion information to the reference destination changing unit 823.

  The L1 side reference information acquisition unit 822 acquires the reference information and prediction correction information on the L1 side from the reference analysis unit 105 if it can be acquired (S402), and supplies them to the reference destination change unit 823.

  Thereafter, the reference destination changing unit 823 acquires the L0 side motion information from the L0 side motion information acquisition unit 821, and the L1 side reference information and prediction correction information from the L1 side reference information acquisition unit 822. When the L1 side prediction correction information is “1”, the reference destination changing unit 823 corrects the reference information by a predetermined reference destination changing process, and generates corrected reference information. In this way, the reference destination on the L1 side is changed (S403). Here, in the case of the first L0-L1 one-sided prediction mode as the operation mode, the reference destination changing process is not allowed to function, and the reference information may be the corrected reference information. Further, in the case of the second L0-L1 one-side prediction mode as the operation mode, the reference destination changing process is preferably functioned. Thereafter, the reference destination changing unit 823 supplies the acquired and generated L0 side and L1 side motion information to the L1 side motion deriving unit 824 and the reference picture acquiring unit 825.

  Thereafter, the L1 side motion deriving unit 824 acquires the L0 side and L1 side motion information from the reference destination changing unit 823. Thereafter, the L1 side motion deriving unit 824 performs a predetermined L1 side motion deriving process, and derives L1 side modified predicted motion vector information from the L0 side (S404). Thereafter, the L1 side motion deriving unit 824 supplies the acquired and generated L0 side and L1 side motion information to the L1 side motion search unit 827.

  Next, the reference picture acquisition unit 825 acquires the corrected reference information as the motion information on the L1 side from the reference destination changing unit 823. The reference picture acquisition unit 825 supplies the L1 side modified reference information to the picture storage unit 102, and acquires the identified reference picture (S405). Thereafter, the reference picture acquisition unit 825 supplies the acquired reference picture to the L1 side motion search unit 827 and the one side prediction signal generation unit 829.

  Thereafter, the target picture acquisition unit 826 acquires the target signal from the input unit 101 (S406) and supplies the target signal to the L1 side motion search unit 827.

  Next, the L1 side motion search unit 827 acquires the L0 side and L1 side motion information from the L1 side motion derivation unit 824, the reference picture from the reference picture acquisition unit 825, and the target signal from the target picture acquisition unit 826, respectively. After that, the L1 side motion search unit 827 performs a predetermined motion search process such as AVC by using the acquired L1 side corrected predicted motion vector information to determine the L1 side motion vector information (S407). ). Thereafter, the L1 side motion search unit 827 supplies the acquired and generated motion information on the L0 side and the L1 side to the mvd generation unit 828.

  After that, the mvd generation unit 828 acquires the L0 side and L1 side motion information from the L1 side motion search unit 827, subtracts the L1 side corrected predicted motion vector information from the acquired L1 side motion vector information, Residual motion vector information is generated (S408). Thereafter, the mvd generation unit 828 supplies the acquired and generated movement information on the L0 side and the L1 side to the one-side prediction signal generation unit 829.

  Thereafter, the one-side prediction signal generation unit 829 acquires the L0 side and L1 side motion information from the mvd generation unit 828 and the reference picture from the reference picture acquisition unit 825. Thereafter, the one-side prediction signal generation unit 829 generates a prediction signal on the L1 side by a predetermined motion compensation process such as AVC based on the acquired L1-side motion vector information (S409). The one-side prediction signal generation unit 829 supplies the mode determination unit 108 with operation mode information, motion information, and a prediction signal for specifying the operation mode. The one-side prediction signal generation unit 829 supplies the acquired and generated L0 side and L1 side motion information to the motion information storage unit 103 together with the operation mode information.

  Through the above steps, the operation in the configuration of FIG. 8B is completed.

  Next, the operation in the configuration of FIG. 8C will be described with reference to the flowchart of FIG.

  First, upon receiving a prediction signal generation start command, the L0 side motion information acquisition unit 841 receives L0 side reference information, prediction motion vector information, and prediction correction information from the reference analysis unit 105 as L0 side motion information belonging to the target region. Is acquired (S501). Thereafter, the L0 side motion information acquisition unit 841 supplies the acquired L0 side motion information to the reference destination changing unit 843.

  After that, the L1 side reference information acquisition unit 842 acquires the reference information and prediction correction information on the L1 side from the reference analysis unit 105 if it can be acquired (S502), and supplies them to the reference destination change unit 843.

  Next, the reference destination changing unit 843 acquires the L0 side motion information from the L0 side motion information acquisition unit 821, and the L1 side reference information and prediction correction information from the L1 side reference information acquisition unit 822. When the prediction correction information on the L0 side is “1”, the reference destination changing unit 843 corrects the reference information on the L0 side by a predetermined reference destination changing process, and generates correction reference information on the L0 side. Similarly, when the prediction correction information on the L1 side is “1”, the reference destination changing unit 843 corrects the reference information on the L1 side by a predetermined reference destination changing process, and changes the reference reference information on the L1 side. Generate. In this way, the reference destination on the L0 side and / or the L1 side is changed (S503). Here, in the case of the first L0-L1 bi-prediction mode as the operation mode, the predetermined reference destination changing process is not allowed to function, and the reference information may be the corrected reference information. Further, in the case of the second L0-L1 bi-prediction mode, the third L0-L1 bi-prediction mode, and the fourth L0-L1 bi-prediction mode as the operation mode, the predetermined reference destination changing process is preferably made to function. After that, the reference destination changing unit 843 obtains and generates the L0 side motion deriving unit 844 for the L0 side motion information, and acquires and generates the L0 side and L1 side motion information for the L1 side motion deriving unit 845 and the reference picture obtaining unit. 846.

  Next, the L0 side motion deriving unit 844 acquires the L0 side motion information from the reference destination changing unit 843. Thereafter, the L0 side motion deriving unit 844 performs a predetermined L0 side motion deriving process, and derives L0 side corrected predicted motion vector information from the L0 side (S504). Thereafter, the L0 side motion deriving unit 844 supplies the acquired and generated L0 side motion information to the bi-prediction signal generation unit 850.

  In addition, the L1 side motion deriving unit 845 acquires the L0 side and L1 side motion information from the reference destination changing unit 843. Thereafter, the L1 side motion deriving unit 845 performs a predetermined L1 side motion deriving process, and derives L1 side corrected predicted motion vector information from the L0 side (S505). Thereafter, the L1 side motion deriving unit 845 supplies the acquired and generated L0 side and L1 side motion information to the L1 side motion search unit 848.

  The reference picture acquisition unit 846 acquires the modified reference information as the motion information on the L0 side and the L1 side from the reference destination changing unit 843. Thereafter, the reference picture acquisition unit 846 supplies the L0 side and L1 side modified reference information to the picture storage unit 102, and acquires the specified L0 side and L1 side reference pictures (S506). Thereafter, the reference picture acquisition unit 846 supplies the acquired L1 side reference picture to the L1 side motion search unit 827. Also, the reference picture acquisition unit 846 supplies the acquired L0 side and L1 side reference pictures to the bi-prediction signal generation unit 850.

  Further, the target picture acquisition unit 847 acquires the target signal from the input unit 101 (S507) and supplies the target signal to the L1 side motion search unit 848.

  Next, the L1 side motion search unit 848 acquires the L0 side and L1 side motion information from the L1 side motion deriving unit 845, the reference picture from the reference picture acquisition unit 846, and the target signal from the target picture acquisition unit 847, respectively. After that, the L1 side motion search unit 848 performs the L1 side motion search by a predetermined motion search process such as AVC, for example, using the acquired L1 side corrected prediction motion vector information, and determines the motion vector information. (S508). Thereafter, the L1 side motion search unit 848 supplies the acquired and generated L0 side and L1 side motion information to the mvd generation unit 849.

  Next, the mvd generation unit 849 acquires the L0 side and L1 side motion information from the L1 side motion search unit 848, subtracts the L1 side corrected predicted motion vector information from the acquired L1 side motion vector information, and outputs the L1 side motion information. Side residual motion vector information is generated (S509). Thereafter, the mvd generation unit 849 supplies the acquired and generated motion information on the L0 side and the L1 side to the bi-prediction signal generation unit 850.

  Next, the bi-prediction signal generation unit 850 acquires the L0 side and L1 side motion information from the mvd generation unit 828, and the L0 side and L1 side reference pictures from the reference picture acquisition unit 825. Thereafter, the bi-prediction signal generation unit 850 generates a prediction signal on the L0 side by a predetermined motion compensation process such as AVC based on the acquired corrected prediction motion vector information on the L0 side (S510).

  Further, the bi-prediction signal generation unit 850 generates an L1 side prediction signal by a predetermined motion compensation process such as AVC based on the acquired L1 side motion vector information (S511).

  Thereafter, the bi-prediction signal generation unit 850 generates a combined prediction signal by combining the generated L0-side prediction signal and the L1-side prediction signal (S512), thereby obtaining a bi-prediction signal. Thereafter, the bi-prediction signal generation unit 850 supplies the mode determination unit 108 with the operation mode information for specifying the operation mode, the acquired and generated motion information on the L0 side and the L1 side, and the combined prediction signal. The bi-prediction signal generation unit 850 supplies the acquired and generated motion information on the L0 side and the L1 side to the motion information storage unit 103 together with the operation mode information.

  Through the above steps, the operation in the configuration of FIG. 8C is completed.

  The operation in the configuration of FIG. 9A is shown by the flowchart of FIG. Here, the operation in the configuration of FIG. 9A is equivalent to the operation in FIG. 10 showing the operation in the configuration of FIG. 8A with the handling of the L0 side and the L1 side reversed. The description is omitted.

  The operation in the configuration of FIG. 9B is shown by the flowchart of FIG. Here, the operation in the configuration in FIG. 9B is equivalent to the operation in FIG. 11 showing the operation in the configuration in FIG. 8B with the handling of the L0 side and the L1 side reversed. The description is omitted.

  The operation in the configuration of FIG. 9C is shown by the flowchart of FIG. Here, the operation in the configuration of FIG. 9C is equivalent to the operation in FIG. 12 showing the operation in the configuration of FIG. 8C with the handling of the L0 side and the L1 side reversed. The description is omitted.

  Here, the following effects are obtained by the configuration and operation of the moving picture coding apparatus 100 of the second embodiment.

  First, the configuration of the first modified prediction signal generation unit 106 in the video encoding device 100 according to the second embodiment is configured to generate the first L0 one-side prediction signal in FIG. 7A as shown in FIG. Unit 106-1 and the second L0 one-side prediction signal generation unit 106-2. Thereby, based on the L0 side motion information, the L0 side motion vector information can be generated, and the L0 side prediction signal can be generated. Thereby, in the moving picture coding apparatus 100 of Embodiment 2, even when only the motion information on the L0 side can be used, the reference destination of the reference information on the L0 side can be changed, so that more operations can be performed. Motion vector information and prediction signals according to modes can be obtained. Therefore, it becomes possible to identify an operation mode that can be encoded with a smaller code amount from the results of many operation modes, and generate a code string by encoding in the operation mode. Further, by having a configuration like the moving image encoding apparatus 100 of the second embodiment, the code amount is equal to or less than that required when encoding residual motion vector information, reference information, and a prediction signal by a conventional method. By selecting and encoding the result according to the operation mode, a code string can be generated with a code amount equal to or less than the code amount according to the conventional method.

  In addition, the configuration of the second modified prediction signal generation unit 107 in the moving picture coding apparatus 100 according to Embodiment 2 is configured to generate the first L1 one-side prediction signal in FIG. 7B as shown in FIG. 9A. Unit 107-1 and second L1 one-side predicted signal generation unit 107-2. Thereby, based on the motion information on the L1 side, motion vector information on the L1 side can be generated, and a prediction signal on the L1 side can be generated. Therefore, similarly to the effect of the configuration shown in FIG. 8A, a code string can be generated with a code amount equal to or less than the code amount according to the conventional method.

  In addition, the configuration of the first modified prediction signal generation unit 106 in the video encoding device 100 according to the second embodiment is the first L0-L1 one-sided prediction in FIG. 7A as shown in FIG. 8B. The signal generation unit 106-3 and the second L0-L1 one-side prediction signal generation unit 106-4 are included. As a result, based on the L0 side motion information, the L1 side modified prediction motion vector information can be derived, the L1 side motion vector information can be generated, and the L1 side prediction signal can be generated. Further, with respect to the reference information used when deriving the L1 side corrected predicted motion vector information, modified reference information is generated by changing the reference destination based on a predetermined method, and the L1 side corrected predicted motion vector is generated. By having an operation mode used for derivation of information, motion vector information and prediction signals in more operation modes can be obtained. Therefore, it becomes possible to identify an operation mode that can be encoded with a smaller code amount from the results of many operation modes, and generate a code string by encoding in the operation mode. Further, by having a configuration like the moving image encoding apparatus 100 of the second embodiment, the code amount is equal to or less than that required when encoding residual motion vector information, reference information, and a prediction signal by a conventional method. By selecting and encoding the result according to the operation mode, a code string can be generated with a code amount equal to or less than the code amount according to the conventional method.

  In addition, the configuration of the second modified prediction signal generation unit 107 in the video encoding device 100 according to Embodiment 2 is the first L1-L0 one-sided prediction in FIG. 7B as shown in FIG. 9B. The signal generation unit 107-3 and the second L1-L0 one-side prediction signal generation unit 107-4 are included. As a result, based on the L1 side motion information, the L0 side modified prediction motion vector information can be derived, the L0 side motion vector information can be generated, and the L0 side prediction signal can be generated. Therefore, similarly to the effect of the configuration shown in FIG. 8B, a code string can be generated with a code amount equal to or less than the code amount according to the conventional method.

  In addition, the configuration of the first modified prediction signal generation unit 106 in the video encoding device 100 according to the second embodiment is the first L0-L1 bi-prediction in FIG. 7A as shown in FIG. 8C. A signal generation unit 106-5, a second L0-L1 bi-prediction signal generation unit 106-6, a third L0-L1 bi-prediction signal generation unit 106-7, and a fourth L0-L1 bi-prediction signal generation unit 106-8. . Thus, based on the L0 side motion information, the L1 side motion vector information is generated, the L0 side prediction signal and the L1 side prediction signal are generated, and then the combined prediction signal that becomes the bi-prediction signal is generated. Will be able to.

(1) a first L0-L1 bi-prediction mode that is an operation mode in which the reference destination is not changed on both the L0 side and the L1 side;
(2) The reference destination is not changed on the L0 side, and the reference is modified by changing the reference destination based on a predetermined method with respect to the reference information used when the corrected motion vector information on the L1 side is derived. A second L0-L1 bi-prediction mode, which is an operation mode that generates information and is used for deriving the L1 side modified prediction motion vector information;
(3) For the reference information used when deriving the L0-side corrected predicted motion vector information, corrected reference information is generated by changing the reference destination based on a predetermined method, and the L0-side corrected predicted motion A third L0-L1 bi-prediction mode, which is an operation mode that is used for deriving vector information and does not change the reference destination on the L1 side,
(4) In an operation mode in which modified reference information on the L0 side and the L1 side is generated by changing the reference destination on both the L0 side and the L1 side, and is used to derive the modified predicted motion vector information on the L0 side and the L1 side. A fourth L0-L1 bi-prediction mode,
Etc., motion vector information and prediction signals in more operation modes can be obtained. Therefore, it becomes possible to identify an operation mode that can be encoded with a smaller code amount from the results of many operation modes, and generate a code string by encoding in the operation mode. Further, by having a configuration like the moving image encoding apparatus 100 of the second embodiment, the code amount is equal to or less than that required when encoding residual motion vector information, reference information, and a prediction signal by a conventional method. By selecting and encoding the result according to the operation mode, a code string can be generated with a code amount equal to or less than the code amount according to the conventional method.

  In the second L0-L1 bi-prediction mode, the accuracy of the L0 side predicted motion vector information and the reference information obtained from the motion prediction unit 104 is good, and the quality of the L0 side prediction signal obtained by these is good, but the motion prediction unit Based on the reference information on the L1 side obtained from 104, it functions effectively when the accuracy of the modified prediction motion vector information derived from the prediction motion vector information on the L0 side is not very good. In such a case, by changing the reference destination on the L1 side, switching to a reference picture that is likely to obtain a better prediction signal according to the derived corrected prediction motion vector information, and performing a motion search to correct the prediction motion The amount of residual motion vector information, reference information or modified reference information obtained by correcting vector information and identifying motion vector information that can obtain a better prediction signal, and the amount of information of the combined prediction signal are less than the amount of information by the conventional method This is an effective operation mode.

  In the third L0-L1 bi-prediction mode, the quality of the reference picture specified from the reference information on the L0 side obtained from the motion prediction unit 104 is not so good, but based on the reference information on the L1 side obtained from the motion prediction unit 104. Thus, the accuracy of the modified predicted motion vector information derived from the L0 side predicted motion vector information functions effectively when the accuracy is relatively good. In such a case, by changing the reference destination on the L0 side, switching to a reference picture from which a better prediction signal is likely to be obtained, deriving the L0 side modified prediction motion vector information, and better prediction on the L0 side By obtaining a signal, residual motion vector information, reference information or modified reference information obtained by improving the quality of a combined prediction signal obtained by combining with a prediction signal on the L1 side, and the information amount of the combined prediction signal are conventionally This is an effective operation mode when the amount of information is less than or equal to that of the above method.

  In the fourth L0-L1 bi-prediction mode, the quality of the reference picture specified from the L0 side and L1 side reference information obtained from the motion prediction unit 104 is not very good, and the accuracy of the prediction motion vector information on the L0 side is not so good. Since it is not good, it functions effectively when the accuracy of the modified predicted motion vector information on the L1 side is not very good. In such a case, by changing the reference destination on the L0 side, switching to a reference picture from which a better prediction signal is likely to be obtained, deriving the L0 side modified prediction motion vector information, and better prediction on the L0 side Get a signal. In addition, by changing the reference destination on the L1 side and deriving the L1 side modified prediction motion vector information from the L0 side prediction motion vector information, the reference picture that is likely to obtain a better prediction signal is switched and a motion search is performed. To correct the L1 side corrected predicted motion vector information, and specify motion vector information from which a better prediction signal can be obtained, thereby obtaining a better L1 side prediction signal. In this way, residual motion vector information, reference information or modified reference information and synthesis obtained by improving the quality of the synthesized prediction signal obtained by synthesizing the obtained L0 side prediction signal and the L1 side prediction signal. This is an effective operation mode when the information amount of the prediction signal is equal to or less than the information amount according to the conventional method.

  In addition, by having the first to fourth L0-L1 bi-prediction modes as described above, it is possible to change the reference destination or to correct one of the corrected predicted motion vector information by motion search, so that the conventional space The quality of the prediction signal on the L0 side and the prediction signal on the L1 side is changed with a small amount of information, which is difficult in the direct mode in time and time, and the quality of the composite prediction signal is higher than the quality of the composite prediction signal by the conventional method. Can be controlled.

  Further, by having the first to fourth L0-L1 bi-prediction modes as described above, two motion vector information is determined from two prediction motion vector information as in the bi-prediction according to the conventional method, Since it is not necessary to encode the two residual motion vector information, a more efficient code string can be generated for the motion information.

  In addition, the configuration of the second modified prediction signal generation unit 107 in the moving picture coding apparatus 100 according to Embodiment 2 is the first L1-L0 bi-prediction of FIG. 7B as shown in FIG. 9C. A signal generation unit 107-5, a second L1-L0 bi-prediction signal generation unit 107-6, a third L1-L0 bi-prediction signal generation unit 107-7, and a fourth L1-L0 bi-prediction signal generation unit 107-8. . As a result, based on the L1 side motion information, L0 side motion vector information is generated, a L0 side prediction signal and an L1 side prediction signal are generated, and then a combined prediction signal to be a bi-prediction signal is generated. Will be able to. Therefore, similarly to the effect of the configuration shown in FIG. 8C, the code string can be generated with a code amount equal to or less than the code amount according to the conventional method.

  Next, as the moving picture decoding apparatus according to the second embodiment, the first modified prediction signal generation unit 210 and the second modified prediction signal generation unit 211 in the moving picture decoding apparatus 200 according to the first embodiment are realized with a more detailed configuration. I will show you what I did. Detailed configurations of the first modified predicted signal generation unit 210 and the second modified predicted signal generation unit 211 in Embodiment 2 will be described with reference to FIGS. 7C and 7D.

  The first modified prediction signal generation unit 210 in FIG. 7C is, for example, a first L0 one-side prediction signal generation unit 210-1, a second L0 one-side prediction signal generation unit 210-2, and a first L0-L1 one-side prediction signal generation unit 210. -3, the second L0-L1 one-side prediction signal generation unit 210-4, the first L0-L1 bi-prediction signal generation unit 210-5, the second L0-L1 bi-prediction signal generation unit 210-6, and the third L0-L1 bi-prediction signal generation Of the unit 210-7 and the fourth L0-L1 bi-prediction signal generation unit 210-8, at least one, any combination of two or more, or all may be configured. Here, the discussion will be made assuming that everything is used. Here, L0-L1 attached to the names of the parts indicates that the L1 side motion vector information is derived based on the L0 side reference information or the corrected reference information and the predicted motion vector information. ing.

  Similarly, the second modified prediction signal generation unit 211 in FIG. 7D is, for example, a first L1 one-side prediction signal generation unit 211-1, a second L1 one-side prediction signal generation unit 211-2, and a first L1-L0 one-side prediction signal. Generator 211-3, second L1-L0 one-side prediction signal generation unit 211-4, first L1-L0 bi-prediction signal generation unit 211-5, second L1-L0 bi-prediction signal generation unit 211-6, and third L1-L0 bi-prediction Of the prediction signal generation unit 211-7 and the fourth L1-L0 bi-prediction signal generation unit 211-8, at least one, any combination of two or more, or all may be used. Here, the discussion will be made assuming that everything is used. Here, L1-L0 given to the names of the parts indicates that the L0 side motion vector information is derived based on the L1 side reference information or the corrected reference information and the predicted motion vector information. ing.

  16 (A) to 16 (C) and FIGS. 17 (A) to 17 (C) are configuration diagrams showing details of the respective parts in FIGS. 7 (C) and 7 (D). FIG. 16A is a configuration diagram of the first L0 one-side prediction signal generation unit 210-1 and the second L0 one-side prediction signal generation unit 210-2. FIG. 16B is a configuration diagram of the first L0-L1 one-side prediction signal generation unit 210-3 and the second L0-L1 one-side prediction signal generation unit 210-4. FIG. 16C illustrates a first L0-L1 bi-prediction signal generation unit 210-5, a second L0-L1 bi-prediction signal generation unit 210-6, a third L0-L1 bi-prediction signal generation unit 210-7, and a fourth L0-L1. It is a block diagram of the bi prediction signal production | generation part 210-8.

  Similarly, FIG. 17A is a configuration diagram of the first L1 one-side prediction signal generation unit 211-1 and the second L1 one-side prediction signal generation unit 211-2. FIG. 17B is a configuration diagram of the first L1-L0 one-side prediction signal generation unit 211-3 and the second L1-L0 one-side prediction signal generation unit 211-4. FIG. 17C shows the first L1-L0 bi-prediction signal generation unit 211-5, the second L1-L0 bi-prediction signal generation unit 211-6, the third L1-L0 bi-prediction signal generation unit 211-7, and the fourth L1-L0. It is a block diagram of the bi prediction signal production | generation part 211-8.

  The configuration of FIG. 16A is a configuration for generating L0 side motion vector information and generating a L0 side prediction signal on the decoding side based on the L0 side motion information. The configuration of FIG. 16A includes an L0 side motion information acquisition unit 1601, a reference destination change unit 1602, an L0 side motion derivation unit 1603, an mvd addition unit 1604, a reference picture acquisition unit 1605, and a one-side prediction signal generation unit 1606.

  The L0 side motion information acquisition unit 1601 performs a predetermined L0 side motion information acquisition process. In this predetermined L0 side motion information acquisition processing, the L0 side motion information acquisition unit 1601 receives, from the reference analysis unit 209, L0 side reference information, predicted motion vector information, and prediction as the L0 side motion information belonging to the decoding target region. Get correction information. The L0 side motion information acquisition unit 1601 supplies the acquired L0 side motion information to the reference destination change unit 1602.

  The reference destination changing unit 1602 performs a predetermined reference destination changing process. In the predetermined reference destination changing process, the reference destination changing unit 1602 acquires the L0 side motion information from the L0 side motion information acquiring unit 1601. When the prediction correction information on the L0 side is “1”, the reference destination changing unit 1602 corrects the reference information by a predetermined reference destination changing process, and generates corrected reference information. Here, it is preferable that the predetermined reference destination changing process does not function according to the operation mode, and the reference information is changed to the corrected reference information. For example, the first L0 one-side predicted signal generation unit 210-1 may not use this predetermined reference destination changing process, and may make the reference information the corrected reference information. The first L0 one-side prediction signal generation unit 210-1 generates the L0-side motion vector information without correcting the L0-side prediction motion vector information with the L0-side correction reference information as the operation mode, and generates the L0-side prediction signal. Operate in the first L0 one-sided prediction mode. Further, the second L0 one-side predicted signal generation unit 210-2 may function the predetermined reference destination changing process. The second L0 one-side predicted signal generation unit 210-2 corrects the L0-side predicted motion vector information with the L0-side corrected reference information as the operation mode, generates the L0-side motion vector information, and generates the L0-side predicted signal. It operates in the second L0 one-sided prediction mode as generated. The reference destination changing unit 1602 supplies the acquired and generated motion information to the L0 side motion deriving unit 1603 and the reference picture acquiring unit 1605.

  The L0 side motion deriving unit 1603 performs a predetermined L0 side motion deriving process. In this predetermined L0 side motion deriving process, the L0 side motion deriving unit 1603 acquires the L0 side motion information from the reference destination changing unit 1602. Thereafter, similarly to the L0 side motion deriving unit 803 on the encoding side, the L0 side modified prediction motion vector information is derived based on the above (Equation 1). The L0 side motion deriving unit 1603 supplies the acquired and generated motion information to the mvd adding unit 1604.

  The mvd addition unit 1604 performs a predetermined mvd addition process. In this predetermined mvd addition process, the mvd addition unit 1604 acquires the L0 side motion information from the L0 side motion deriving unit 1603. The mvd addition unit 1604 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207. The mvd addition unit 1604 generates the L0 side motion vector information by adding the acquired residual motion vector information to the acquired L0 side modified prediction motion vector information. The mvd addition unit 1604 supplies the acquired and generated L0 side motion information to the one-side prediction signal generation unit 1606.

  The reference picture acquisition unit 1605 performs a predetermined reference picture acquisition process. In this predetermined reference picture acquisition process, the reference picture acquisition unit 1605 acquires the corrected reference information as motion information from the reference destination change unit 1602. The reference picture acquisition unit 1605 supplies the L0-side modified reference information to the picture storage unit 202, and acquires the identified reference picture. The reference picture acquisition unit 1605 supplies the acquired reference picture to the one-side prediction signal generation unit 1606.

  The one-side prediction signal generation unit 1606 performs a predetermined one-side prediction signal generation process. In the predetermined one-side prediction signal generation process, the one-side prediction signal generation unit 1606 acquires motion information from the mvd addition unit 1604 and a reference picture from the reference picture acquisition unit 1605. The one-side prediction signal generation unit 1606 generates a prediction signal by a predetermined motion compensation process such as AVC based on the acquired motion vector information. The one-side prediction signal generation unit 1606 supplies operation mode information, motion information, and a prediction signal for specifying the operation mode to the mode identification unit 206. The one-side prediction signal generation unit 1606 supplies the acquired and generated L0 side motion information to the motion information storage unit 207 together with the operation mode information.

  The configuration of FIG. 16B is a configuration for generating L1 side motion vector information and generating an L1 side prediction signal on the decoding side based on the L0 side motion information. 16B includes an L0 side motion information acquisition unit 1621, an L1 side reference information acquisition unit 1622, a reference destination change unit 1623, an L1 side motion derivation unit 1624, an mvd addition unit 1625, a reference picture acquisition unit 1626, and one side. A prediction signal generation unit 1627 is included.

  The L0 side motion information acquisition unit 1621 performs a predetermined L0 side motion information acquisition process. In this predetermined L0 side motion information acquisition process, the L0 side motion information acquisition unit 1621 receives, from the reference analysis unit 209, the L0 side reference information, the predicted motion vector information, and the prediction as the L0 side motion information belonging to the decoding target area. Get correction information. The L0 side motion information acquisition unit 1621 supplies the acquired L0 side motion information to the reference destination changing unit 1623.

  The L1 side reference information acquisition unit 1622 performs a predetermined L1 side reference information acquisition process. In this predetermined L1 side reference information acquisition process, the L1 side reference information acquisition unit 1622 acquires reference information and prediction correction information on the L1 side from the reference analysis unit 209, if possible, and a reference destination change unit 1623. To supply.

  The reference destination changing unit 1623 performs a predetermined reference destination changing process. In the predetermined reference destination changing process, the reference destination changing unit 1623 acquires the L0 side motion information from the L0 side motion information acquisition unit 1621 and the L1 side reference information acquisition unit 1622 from the L1 side reference information and the prediction correction information. . When the predicted correction information on the L1 side is “1”, the reference destination changing unit 1623 corrects the reference information by a predetermined reference destination changing process, and generates corrected reference information. Here, it is preferable that the predetermined reference destination changing process does not function according to the operation mode, and the reference information is changed to the corrected reference information. For example, the first L0-L1 one-side predicted signal generation unit 210-3 may not use this predetermined reference destination changing process, and may make the reference information the corrected reference information. The first L0-L1 one-side predicted signal generation unit 210-3 corrects the L0-side predicted motion vector information based on the reference information instead of the L1-side corrected reference information as the operation mode, and converts the L1-side motion vector information. It operates in a first L0-L1 one-sided prediction mode that generates a prediction signal on the L1 side. Also, the second L0-L1 one-side prediction signal generation unit 210-4 may function this predetermined reference destination change process. The second L0-L1 one-side predicted signal generation unit 210-4 corrects the L0-side predicted motion vector information based on the L1-side corrected reference information as the operation mode, generates L1-side motion vector information, and generates L1 Operate in a second L0-L1 one-sided prediction mode that generates a side prediction signal. The reference destination changing unit 1623 supplies the acquired and generated L0 side and L1 side motion information to the L1 side motion deriving unit 1624 and the reference picture acquiring unit 1626.

  The L1 side motion deriving unit 1624 performs a predetermined L1 side motion deriving process. In the predetermined L1 side motion deriving process, the L1 side motion deriving unit 1624 acquires the L0 side and L1 side motion information from the reference destination changing unit 1623. Thereafter, similarly to the L1 side motion deriving unit 824 on the encoding side, the L1 side modified prediction motion vector information is derived based on the above (Equation 2) or (Equation 3) according to the operation mode. The L1 side motion deriving unit 1624 supplies the acquired and generated L0 side and L1 side motion information to the mvd addition unit 1625.

  The mvd addition unit 1625 performs a predetermined mvd addition process. In the predetermined mvd addition process, the mvd addition unit 1625 acquires the L0 side and L1 side motion information from the L1 side motion deriving unit 1624. Also, the mvd addition unit 1625 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207. The mvd addition unit 1625 generates the L1 side motion vector information by adding the acquired residual motion vector information to the acquired L1 side modified prediction motion vector information. The mvd addition unit 1625 supplies the acquired and generated motion information on the L0 side and the L1 side to the one-side prediction signal generation unit 1627.

  The reference picture acquisition unit 1626 performs a predetermined reference picture acquisition process. In the predetermined reference picture acquisition process, the reference picture acquisition unit 1626 acquires the motion information on the L0 side and the L1 side from the reference destination changing unit 1623. The reference picture acquisition unit 1626 supplies the acquired modified reference information on the L1 side to the picture storage unit 202, and acquires the identified reference picture. The reference picture acquisition unit 1626 supplies the acquired L1 side reference picture to the one-side prediction signal generation unit 1606.

  The one-side prediction signal generation unit 1627 performs a predetermined one-side prediction signal generation process. In the predetermined one-side prediction signal generation processing, the one-side prediction signal generation unit 1627 acquires the L0 side and L1 side motion information from the mvd addition unit 1625 and the L1 side reference picture from the reference picture acquisition unit 1626. The one-side prediction signal generation unit 1627 generates a prediction signal by a predetermined motion compensation process such as AVC, based on the acquired L1 side motion vector information. The one-side prediction signal generation unit 1627 supplies operation mode information for specifying the operation mode, motion information on the L0 side and L1 side, and a prediction signal to the mode identification unit 206. The one-side prediction signal generation unit 1627 supplies the acquired and generated L0 side and L1 side motion information to the motion information storage unit 207 together with the operation mode information.

  In the configuration of FIG. 16C, the decoding side generates motion vector information on the L1 side based on the motion information on the L0 side, generates a prediction signal on the L0 side, and a prediction signal on the L1 side. This is a configuration for generating a combined prediction signal that becomes a prediction signal. 16C includes an L0 side motion information acquisition unit 1641, an L1 side reference information acquisition unit 1642, a reference destination change unit 1643, an L0 side motion derivation unit 1644, an L1 side motion derivation unit 1645, an mvd addition unit 1646, A reference picture acquisition unit 1647 and a bi-prediction signal generation unit 1648 are included.

  The L0 side motion information acquisition unit 1641 performs a predetermined L0 side motion information acquisition process. In this predetermined L0 side motion information acquisition process, the L0 side motion information acquisition unit 1641 receives, from the reference analysis unit 209, the L0 side reference information, the predicted motion vector information, and the prediction as the L0 side motion information belonging to the decoding target region. Get correction information. The L0 side motion information acquisition unit 1641 supplies the acquired L0 side motion information to the reference destination changing unit 1643.

  The L1 side reference information acquisition unit 1642 performs a predetermined L1 side reference information acquisition process. In this predetermined L1 side reference information acquisition process, the L1 side reference information acquisition unit 1642 acquires reference information and prediction correction information on the L1 side from the reference analysis unit 209, if possible, and a reference destination change unit 1643. To supply.

  The reference destination changing unit 1643 performs a predetermined reference destination changing process. In this predetermined reference destination changing process, the reference destination changing unit 1643 acquires the L0 side motion information from the L0 side motion information acquisition unit 1641 and the L1 side reference information acquisition unit 1642 from the L1 side reference information and the prediction correction information. . When the prediction correction information on the L0 side is “1”, the reference destination changing unit 1643 corrects the reference information on the L0 side by a predetermined reference destination changing process, and generates correction reference information on the L0 side. Similarly, when the L1 side prediction correction information is “1”, the reference destination changing unit 1643 corrects the L1 side reference information by a predetermined reference destination change process, and the L1 side correction reference information is changed. Generate. Here, it is preferable that the predetermined reference destination changing process does not function according to the operation mode, and the reference information is changed to the corrected reference information. For example, as the operation mode, the first L0-L1 bi-prediction signal generation unit 210-5 operating in the first L0-L1 bi-prediction mode does not function the predetermined reference destination change process and changes the reference information to the corrected reference information. good. Further, as operation modes, a second L0-L1 bi-prediction signal generation unit 210-6 that operates in the second L0-L1 bi-prediction mode, and a third L0-L1 bi-prediction signal generation unit 210- that operates in the third L0-L1 bi-prediction mode. The fourth L0-L1 bi-prediction signal generator 210-8 operating in the seventh and fourth L0-L1 bi-prediction modes may function as the predetermined reference destination changing process. The reference destination changing unit 1643 sends the acquired and generated L0 side motion information to the L0 side motion deriving unit 1644, and the acquired and generated L0 side and L1 side motion information to the L1 side motion deriving unit 1645 and the reference picture acquiring unit 1647. Supply.

  The L0 side motion deriving unit 1644 performs a predetermined L0 side motion deriving process. In this predetermined L0 side motion deriving process, the L0 side motion deriving unit 1644 acquires the L0 side motion information from the reference destination changing unit 1643. Thereafter, similarly to the L0 side motion deriving unit 844 on the encoding side, the L0 side modified prediction motion vector information is derived based on the above (Equation 4) according to the operation mode. The L0 side motion deriving unit 1644 supplies the acquired and generated L0 side motion information to the bi-prediction signal generation unit 1648.

  The L1 side motion deriving unit 1645 performs a predetermined L1 side motion deriving process. In the predetermined L1 side motion deriving process, the L1 side motion deriving unit 1645 acquires the L0 side and L1 side motion information from the reference destination changing unit 1643. Thereafter, similarly to the L1 side motion deriving unit 845 on the encoding side, the L1 side corrected predicted motion vector information is derived based on the above (Equation 5) or (Equation 6) according to the operation mode. The L1 side motion deriving unit 1645 supplies the acquired and generated L0 side and L1 side motion information to the mvd addition unit 1646.

  The mvd addition unit 1646 performs a predetermined mvd addition process. In this predetermined mvd addition process, the mvd addition unit 1646 acquires the L0 side and L1 side motion information from the L1 side motion deriving unit 1645. The mvd addition unit 1646 obtains residual motion vector information belonging to the decoding target area from the motion information accumulation unit 207. The mvd addition unit 1646 generates the L1 side motion vector information by adding the acquired residual motion vector information to the acquired L1 side modified prediction motion vector information. The mvd addition unit 1646 supplies the acquired and generated L0 side and L1 side motion information to the bi-prediction signal generation unit 1648.

  The reference picture acquisition unit 1647 performs a predetermined reference picture acquisition process. In this predetermined reference picture acquisition process, the reference picture acquisition unit 1647 acquires the corrected reference information from the reference destination change unit 1643 as the motion information on the L0 side and the L1 side. The reference picture acquisition unit 1647 supplies the L0 side and L1 side modified reference information to the picture storage unit 202, and acquires the identified L0 side and L1 side reference pictures. The reference picture acquisition unit 1647 supplies the acquired L0 side and L1 side reference pictures to the bi-prediction signal generation unit 1648.

  The bi-prediction signal generation unit 1648 performs a predetermined bi-prediction signal generation process. In this predetermined bi-prediction signal generation process, the bi-prediction signal generation unit 1648 includes the L0 side motion deriving unit 1644 to the L0 side motion information, the mvd addition unit 1646 to the L0 side and L1 side motion information, and the reference picture acquisition unit. From 1647, reference pictures on the L0 side and the L1 side are acquired. The bi-prediction signal generation unit 1648 generates a prediction signal on the L0 side by a predetermined motion compensation process such as AVC based on the acquired corrected prediction motion vector information on the L0 side. The bi-prediction signal generation unit 1648 generates a prediction signal on the L1 side by a predetermined motion compensation process such as AVC based on the acquired L1 side motion vector information. After that, the bi-prediction signal generation unit 1648 generates a bi-prediction signal by generating a combined prediction signal obtained by synthesizing the generated L0-side prediction signal and the L1-side prediction signal. The bi-prediction signal generation unit 1648 supplies the mode identification unit 206 with the operation mode information for specifying the operation mode, the acquired and generated motion information on the L0 side and the L1 side, and the combined prediction signal. The bi-prediction signal generation unit 1648 supplies the acquired and generated L0 side and L1 side motion information to the motion information storage unit 207 together with the operation mode information.

  In the configuration of FIG. 16A, the L0 side motion vector information is generated based on the L0 side motion information, and the L0 side prediction signal is generated. On the other hand, in the configuration of FIG. 17A, the handling of the L0 side and the L1 side is reversed, and the configuration for generating the prediction signal on the L1 side is based on the motion information on the L1 side. Different from (A). The details of each part in the configuration of FIG. 17A are omitted because the functions on the L0 side and the L1 side are considered to be reversed and the function is equivalent to that of the configuration of FIG. .

  In the configuration of FIG. 16B, the L1 side motion vector information is generated based on the L0 side motion information, and the L1 side prediction signal is generated. On the other hand, in the configuration of FIG. 17B, the L0 side motion information is generated based on the L1 side motion information by reversing the handling of the L0 side and the L1 side, and the L0 side prediction signal is generated. This is different from the configuration shown in FIG. The details of each part in the configuration of FIG. 17B are omitted because they are considered to have the same function as the configuration of FIG. .

  In the configuration shown in FIG. 16C, the L1 side motion vector information is generated based on the L0 side motion information, the L0 side prediction signal and the L1 side prediction signal are generated, and then the bi-prediction signal is obtained. The configuration is for generating a synthesized prediction signal. On the other hand, in the configuration of FIG. 17C, the L0 side motion information is generated based on the L1 side motion information by reversing the handling of the L0 side and the L1 side. The configuration for generating the combined prediction signal that becomes the bi-prediction signal after generating the prediction signal on the side is different from the configuration of FIG. The details of each part in the configuration of FIG. 17C are omitted because the functions on the L0 side and the L1 side are considered to be reversed and the function is equivalent to that of the configuration of FIG. .

  Next, operation in the configuration of FIG. 16A will be described with reference to a flowchart of FIG.

  First, upon receiving a prediction signal generation start command, the L0 side motion information acquisition unit 1601 receives, from the reference analysis unit 209, L0 side reference information, prediction motion vector information, and prediction as the L0 side motion information belonging to the decoding target region. Correction information is acquired (S901). Thereafter, the L0 side motion information acquisition unit 1601 supplies the acquired L0 side motion information to the reference destination changing unit 1602.

  Thereafter, the reference destination changing unit 1602 acquires the L0 side motion information from the L0 side motion information acquisition unit 1601. Thereafter, the reference destination changing unit 1602 changes the reference destination on the L0 side by correcting the reference information and generating the corrected reference information by a predetermined reference destination changing process according to the operation mode (S902). Thereafter, the reference destination changing unit 1602 supplies the acquired and generated motion information to the L0 side motion deriving unit 1603 and the reference picture acquiring unit 1605.

  Next, the L0 side motion deriving unit 1603 acquires the L0 side motion information from the reference destination changing unit 1602. Thereafter, the L0 side motion deriving unit 1603 performs a predetermined L0 side motion deriving process, and derives L0 side modified prediction motion vector information (S903). Thereafter, the L0 side motion deriving unit 1603 supplies the acquired and generated motion information to the mvd adding unit 1604.

  After that, the mvd addition unit 1604 acquires the L0 side motion information from the L0 side motion deriving unit 1603. The mvd addition unit 1604 acquires residual motion vector information belonging to the decoding target region from the motion information storage unit 207. Thereafter, the mvd addition unit 1604 adds the residual motion vector information to the acquired corrected predicted motion vector information on the L0 side by a predetermined mvd addition process to generate L0 side motion vector information (S904). Thereafter, the mvd addition unit 1604 supplies the acquired and generated L0 side motion information to the one-side prediction signal generation unit 1606.

  The reference picture acquisition unit 1605 acquires corrected reference information as motion information from the reference destination change unit 1602. After that, the reference picture acquisition unit 1605 supplies the L0 side modified reference information to the picture storage unit 202, and acquires the identified reference picture (S905). Thereafter, the reference picture acquisition unit 1605 supplies the acquired reference picture to the one-side prediction signal generation unit 1606.

  Next, the one-side prediction signal generation unit 1606 acquires motion information from the mvd addition unit 1604 and a reference picture from the reference picture acquisition unit 1605. Thereafter, the one-side prediction signal generation unit 1606 generates a L0-side prediction signal by performing a predetermined one-side prediction signal generation process (S906). Thereafter, the one-side prediction signal generation unit 1606 supplies the mode identification unit 206 with operation mode information, motion information, and a prediction signal for specifying the operation mode. The one-side prediction signal generation unit 1606 supplies the acquired and generated L0 side motion information to the motion information storage unit 207 together with the operation mode information.

  Through the above steps, the operation in the configuration of FIG.

  Next, the operation in the configuration of FIG. 16B will be described with reference to the flowchart of FIG.

  First, upon receiving a prediction signal generation start command, the L0 side motion information acquisition unit 1621 receives, from the reference analysis unit 209, L0 side reference information, prediction motion vector information, and prediction as the L0 side motion information belonging to the decoding target region. Correction information is acquired (S1001). Thereafter, the L0 side motion information acquisition unit 1621 supplies the acquired L0 side motion information to the reference destination changing unit 1623.

  The L1 side reference information acquisition unit 1622 acquires the reference information and prediction correction information on the L1 side from the reference analysis unit 209 if it can be acquired (S1002), and supplies them to the reference destination change unit 1623.

  Next, the reference destination changing unit 1623 acquires the L0 side motion information from the L0 side motion information acquisition unit 1621 and the L1 side reference information and prediction correction information from the L1 side reference information acquisition unit 1622. Thereafter, the reference destination changing unit 1623 modifies the reference information by a predetermined reference destination changing process, and generates the corrected reference information, thereby changing the reference destination on the L1 side (S1003). Thereafter, the reference destination changing unit 1623 supplies the acquired and generated L0 side and L1 side motion information to the L1 side motion deriving unit 1624 and the reference picture acquiring unit 1626.

  Next, the L1 side motion deriving unit 1624 acquires the L0 side and L1 side motion information from the reference destination changing unit 1623. Thereafter, the L1 side motion deriving unit 1624 performs the predetermined L1 side motion deriving process to derive the L1 side corrected predicted motion vector information from the L0 side (S1004). Thereafter, the L1 side motion deriving unit 1624 supplies the acquired and generated L0 side and L1 side motion information to the mvd addition unit 1625.

  Next, the mvd addition unit 1625 acquires the L0 side and L1 side motion information from the L1 side motion deriving unit 1624. The mvd addition unit 1625 acquires residual motion vector information belonging to the decoding target region from the motion information accumulation unit 207. Thereafter, the mvd adding unit 1625 adds the residual motion vector information to the acquired L1 side corrected predicted motion vector information through a predetermined mvd addition process, and generates L1 side motion vector information (S1005). Thereafter, the mvd addition unit 1625 supplies the acquired and generated motion information on the L0 side and the L1 side to the one-side prediction signal generation unit 1627.

  The reference picture acquisition unit 1626 acquires the motion information on the L0 side and the L1 side from the reference destination changing unit 1623. Thereafter, the reference picture acquisition unit 1626 supplies the acquired L1-side corrected reference information to the picture storage unit 202, and acquires the identified reference picture (S1006). Thereafter, the reference picture acquisition unit 1626 supplies the acquired L1 side reference picture to the one-side prediction signal generation unit 1606.

  Next, the one-side prediction signal generation unit 1627 acquires the L0 side and L1 side motion information from the mvd addition unit 1625 and the L1 side reference picture from the reference picture acquisition unit 1626. Thereafter, the one-side prediction signal generation unit 1627 generates a prediction signal on the L1 side by performing a predetermined one-side prediction signal generation process based on the acquired L1 side motion vector information (S1007). Thereafter, the one-side prediction signal generation unit 1627 supplies the mode identification unit 206 with the operation mode information for specifying the operation mode, the L0 side and L1 side motion information, and the prediction signal. The one-side prediction signal generation unit 1627 supplies the acquired and generated L0 side and L1 side motion information to the motion information storage unit 207 together with the operation mode information.

  Through the above steps, the operation in the configuration of FIG.

  Next, the operation in the configuration of FIG. 16C will be described with reference to the flowchart of FIG.

  First, upon receiving a prediction signal generation start command, the L0 side motion information acquisition unit 1641 receives, from the reference analysis unit 209, L0 side reference information, prediction motion vector information, and prediction as the L0 side motion information belonging to the decoding target region. Correction information is acquired (S1101). Thereafter, the L0 side motion information acquisition unit 1641 supplies the acquired L0 side motion information to the reference destination changing unit 1643.

  The L1 side reference information acquisition unit 1642 acquires reference information and prediction correction information on the L1 side from the reference analysis unit 209 if it can be acquired (S1102), and supplies them to the reference destination change unit 1643.

  Next, the reference destination changing unit 1643 acquires the L0 side motion information from the L0 side motion information acquisition unit 1641 and the L1 side reference information and prediction correction information from the L1 side reference information acquisition unit 1642. Thereafter, the reference destination changing unit 1643 corrects the reference information on the L0 side and generates the corrected reference information on the L0 side and corrects the reference information on the L1 side by a predetermined reference destination changing process according to the operation mode. Then, the reference reference on the L0 side and the L1 side is changed by generating the correction reference information on the L1 side (S1103). After that, the reference destination changing unit 1643 acquires the acquired and generated L0 side motion information in the L0 side motion deriving unit 1644, and acquires and generates the acquired L0 side and L1 side motion information in the L1 side motion deriving unit 1645 and the reference picture acquisition. Supplied to the unit 1647.

  Next, the L0 side motion deriving unit 1644 acquires the L0 side motion information from the reference destination changing unit 1643. Thereafter, the L0 side motion deriving unit 1644 derives the L0 side corrected predicted motion vector information by performing a predetermined L0 side motion deriving process according to the operation mode (S1104). Thereafter, the L0 side motion deriving unit 1644 supplies the acquired and generated L0 side motion information to the bi-prediction signal generation unit 1648.

  The L1 side motion deriving unit 1645 acquires the motion information on the L0 side and the L1 side from the reference destination changing unit 1643. Thereafter, the L1 side motion deriving unit 1645 derives the L1 side corrected predicted motion vector information from the L0 side by a predetermined L1 side motion deriving process (S1105). Thereafter, the L1 side motion deriving unit 1645 supplies the acquired and generated L0 side and L1 side motion information to the mvd addition unit 1646.

  Next, the mvd addition unit 1646 acquires L0 side and L1 side motion information from the L1 side motion deriving unit 1645. The mvd addition unit 1646 obtains residual motion vector information belonging to the decoding target area from the motion information accumulation unit 207. Thereafter, the mvd addition unit 1646 adds the residual motion vector information to the acquired L1 side corrected predicted motion vector information through a predetermined mvd addition process to generate L1 side motion vector information (S1106). Thereafter, the mvd addition unit 1646 supplies the acquired and generated L0 side and L1 side motion information to the bi-prediction signal generation unit 1648.

  The reference picture acquisition unit 1647 acquires the corrected reference information from the reference destination changing unit 1643 as the motion information on the L0 side and the L1 side. Thereafter, the reference picture acquisition unit 1647 supplies the L0 side and L1 side modified reference information to the picture storage unit 202, and acquires the specified L0 side and L1 side reference pictures (S1107). Thereafter, the reference picture acquisition unit 1647 supplies the acquired L0 side and L1 side reference pictures to the bi-prediction signal generation unit 1648.

  Next, the bi-prediction signal generation unit 1648 receives the L0 side motion information from the L0 side motion deriving unit 1644, the mvd addition unit 1646 from the L0 side and the L1 side motion information, and the reference picture acquisition unit 1647 from the L0 side and the L1 side. Get a reference picture. Thereafter, the bi-prediction signal generation unit 1648 generates a prediction signal on the L0 side by performing a predetermined bi-prediction signal generation process based on the acquired corrected prediction motion vector information on the L0 side (S1108).

  In addition, the bi-prediction signal generation unit 1648 generates a prediction signal on the L1 side by performing a predetermined bi-prediction signal generation process based on the acquired motion vector information on the L1 side (S1109).

  Thereafter, the bi-prediction signal generation unit 1648 combines the generated L0 side prediction signal and the L1 side prediction signal to generate a combined prediction signal (S1110). Thereafter, the bi-prediction signal generation unit 1648 supplies the mode identification unit 206 with the operation mode information for specifying the operation mode, the acquired and generated L0 side and L1 side motion information, and the combined prediction signal. The bi-prediction signal generation unit 1648 supplies the acquired and generated L0 side and L1 side motion information to the motion information storage unit 207 together with the operation mode information.

  Through the above steps, the operation in the configuration of FIG.

  The operation in the configuration of FIG. 17A is shown by the flowchart of FIG. Here, the operation in the configuration of FIG. 17A is equivalent to the operation in FIG. 18 showing the operation in the configuration of FIG. 16A with the handling of the L0 side and the L1 side reversed. Therefore, the description is omitted.

  The operation in the configuration of FIG. 17B is shown by the flowchart of FIG. Here, the operation in the configuration of FIG. 17B is equivalent to the operation in FIG. 19 showing the operation in the configuration of FIG. 16B with the handling of the L0 side and the L1 side reversed. Therefore, the description is omitted.

  The operation in the configuration of FIG. 17C is shown by the flowchart of FIG. Here, the operation in the configuration of FIG. 17C is equivalent to the operation in FIG. 20 showing the operation in the configuration of FIG. 16C with the handling of the L0 side and the L1 side reversed. Therefore, the description is omitted.

  Here, the following effects are obtained by the configuration and operation of the moving picture decoding apparatus 200 according to the second embodiment.

  First, in the moving picture decoding apparatus 200 of Embodiment 2, the configuration of the first modified prediction signal generation unit 210 is the first L0 one-side prediction signal generation unit in FIG. 7C as shown in FIG. 210-1 and the 2nd L0 one-side prediction signal generation part 210-2 are included. Thereby, based on the L0 side motion information, the L0 side motion vector information can be generated, and the L0 side prediction signal can be generated. Therefore, in the moving image coding apparatus 100 of Embodiment 2, the first modified prediction signal generation unit 106 uses the first L0 one-side prediction signal generation unit 106 in FIG. 7A as shown in FIG. 8A. −1 and a second L0 one-side prediction signal generation unit 106-2 are used to obtain a code string encoded based on an operation mode, and from the acquired code string, motion vector information corresponding to the above-described operation mode, In addition, a prediction signal or a combined prediction signal can be obtained, and a code string generated based on the above operation mode can be correctly decoded.

  Further, in the moving picture decoding apparatus 200 of Embodiment 2, the configuration of the second modified prediction signal generation unit 211 is the first L1 one-side prediction signal generation unit in FIG. 7D as shown in FIG. 211-1 and the second L1 one-side predicted signal generator 211-2. Thereby, based on the motion information on the L1 side, motion vector information on the L1 side can be generated, and a prediction signal on the L1 side can be generated. Therefore, in the moving image coding apparatus 100 of Embodiment 2, the second modified prediction signal generation unit 107 in the first L1 one-side prediction signal generation unit 107 in FIG. 7B as shown in FIG. 9A. -1, a code sequence encoded based on an operation mode that uses the configuration of the second L1 one-side prediction signal generation unit 107-2 is acquired, and a motion vector corresponding to the above-described operation mode is acquired from the acquired code sequence Information, a prediction signal, or a combined prediction signal can be obtained, and a code string generated based on the above-described operation mode can be correctly decoded.

  Also, the configuration of the first modified prediction signal generation unit 210 in the video decoding device 200 of Embodiment 2 is the first L0-L1 one-side prediction signal of FIG. 7C as shown in FIG. A generation unit 210-3 and a second L0-L1 one-side prediction signal generation unit 210-4 are included. As a result, based on the L0 side motion information, the L1 side modified prediction motion vector information can be derived, the L1 side motion vector information can be generated, and the L1 side prediction signal can be generated. Therefore, in the moving image encoding apparatus 100 of Embodiment 2, the first modified prediction signal generation unit 106 generates the first L0-L1 one-side prediction signal in FIG. 7A as shown in FIG. 8B. Unit 106-3 and the second L0-L1 one-side prediction signal generation unit 106-4 are used to obtain a coded sequence based on the operation mode, and from the obtained code sequence to the above-described operation mode. Corresponding motion vector information and a prediction signal or a combined prediction signal can be obtained, and a code string generated based on the above-described operation mode can be correctly decoded.

  Further, in the moving picture decoding apparatus 200 according to Embodiment 2, the configuration of the second modified prediction signal generation unit 211 is the first L1-L0 one-side prediction signal in FIG. 7D as shown in FIG. The generation unit 211-3 and the second L1-L0 one-side prediction signal generation unit 211-4 are included. As a result, based on the L1 side motion information, the L0 side modified prediction motion vector information can be derived, the L0 side motion vector information can be generated, and the L0 side prediction signal can be generated. Therefore, in the moving image coding apparatus 100 of Embodiment 2, the first modified prediction signal generation unit 106 generates the first L1-L0 one-side prediction signal of FIG. 7B as shown in FIG. 9B. Unit 107-3 and the second L1-L0 one-side prediction signal generation unit 107-4 are used to obtain a code sequence encoded based on an operation mode, and from the acquired code sequence to the above-described operation mode. Corresponding motion vector information and a prediction signal or a combined prediction signal can be obtained, and a code string generated based on the above-described operation mode can be correctly decoded.

  Further, the configuration of the first modified prediction signal generation unit 210 in the moving picture decoding apparatus 200 according to Embodiment 2 is the first L0-L1 bi-prediction signal shown in FIG. 7C as shown in FIG. A generation unit 210-5, a second L0-L1 bi-prediction signal generation unit 210-6, a third L0-L1 bi-prediction signal generation unit 210-7, and a fourth L0-L1 bi-prediction signal generation unit 210-8 are included. Thus, based on the L0 side motion information, the L1 side motion vector information is generated, the L0 side prediction signal and the L1 side prediction signal are generated, and then the combined prediction signal that becomes the bi-prediction signal is generated. Will be able to. Therefore, in the moving image encoding apparatus 100 of Embodiment 2, the first modified prediction signal generation unit 106 generates the first L0-L1 bi-prediction signal of FIG. 7A as shown in FIG. 8C. Unit 106-5, second L0-L1 bi-prediction signal generation unit 106-6, third L0-L1 bi-prediction signal generation unit 106-7, and fourth L0-L1 bi-prediction signal generation unit 106-8 A code sequence encoded based on the operation mode is acquired, and motion vector information corresponding to the above-described operation mode and a prediction signal or a combined prediction signal can be obtained from the acquired code sequence. The code string generated based on this can be correctly decoded.

(1) a first L0-L1 bi-prediction mode that is an operation mode in which the reference destination is not changed on both the L0 side and the L1 side;
(2) The reference destination is not changed on the L0 side, and the reference is modified by changing the reference destination based on a predetermined method with respect to the reference information used when the corrected motion vector information on the L1 side is derived. A second L0-L1 bi-prediction mode, which is an operation mode that generates information and is used for deriving the L1 side modified prediction motion vector information;
(3) For the reference information used when deriving the L0-side corrected predicted motion vector information, corrected reference information is generated by changing the reference destination based on a predetermined method, and the L0-side corrected predicted motion A third L0-L1 bi-prediction mode, which is an operation mode that is used for deriving vector information and does not change the reference destination on the L1 side,
(4) In an operation mode in which modified reference information on the L0 side and the L1 side is generated by changing the reference destination on both the L0 side and the L1 side, and is used to derive the modified predicted motion vector information on the L0 side and the L1 side. A fourth L0-L1 bi-prediction mode,
Etc., the code string generated based on the above-described operation mode can be correctly decoded.

  In addition, the moving image encoding apparatus 100 according to the second embodiment has the first to fourth L0-L1 bi-prediction modes as described above, so that the reference destination can be changed or one of the prediction motion vector information can be corrected. By using motion search, the quality of the L0 side prediction signal and the L1 side prediction signal, which was difficult in the conventional spatial and temporal direct mode, is changed with a small amount of information, and the combined prediction signal Can be controlled to be equal to or higher than the quality of the synthesized prediction signal by the conventional method. The moving image decoding apparatus 200 according to the second embodiment has the first to fourth L0-L1 bi-prediction modes as described above, so that the code string generated in this way is acquired and decoded correctly. Will be able to.

  In addition, the moving image coding apparatus 100 according to the second embodiment has the first to fourth L0-L1 bi-prediction modes as described above, so that two predictions can be performed as in the bi-prediction according to the conventional method. Since the two motion vector information is determined from the motion vector information and it is not necessary to encode the two residual motion vector information, a more efficient code string can be generated for the motion information. The moving image decoding apparatus 200 according to the second embodiment has the first to fourth L0-L1 bi-prediction modes as described above, so that the code string generated in this way is acquired and decoded correctly. Will be able to.

  Further, in the moving picture decoding apparatus 200 according to Embodiment 2, the configuration of the second modified prediction signal generation unit 211 is the first L1-L0 bi-prediction signal in FIG. 7D as shown in FIG. A generation unit 211-5, a second L1-L0 bi-prediction signal generation unit 211-6, a third L1-L0 bi-prediction signal generation unit 211-7, and a fourth L1-L0 bi-prediction signal generation unit 211-8 are included. As a result, based on the L1 side motion information, L0 side motion vector information is generated, a L0 side prediction signal and an L1 side prediction signal are generated, and then a combined prediction signal to be a bi-prediction signal is generated. Will be able to. Therefore, the second modified prediction signal generation unit 107 in the moving picture coding apparatus 100 of Embodiment 2 generates the first L1-L0 bi-prediction signal of FIG. 7B as shown in FIG. 9C. Unit 107-5, the second L1-L0 bi-prediction signal generation unit 107-6, the third L1-L0 bi-prediction signal generation unit 107-7, and the fourth L1-L0 bi-prediction signal generation unit 107-8 are used. A code sequence encoded based on the operation mode is acquired, and motion vector information corresponding to the above-described operation mode and a prediction signal or a combined prediction signal can be obtained from the acquired code sequence. The code string generated based on this can be correctly decoded.

  Furthermore, the operation modes in the first and second modified prediction signal generation units of the moving image encoding apparatus 100 and the moving image decoding apparatus 200 according to Embodiment 2 are other than the operation modes described above, for example, FIG. 24 and FIG. 25. The encoding control unit 120 and the decoding control unit 216 are configured to control each unit according to the operation mode, having at least one of the operation modes associated with the modified prediction mode index as shown in FIG. It doesn't matter. Moreover, you may comprise so that this correction | amendment prediction mode index may be encoded and decoded as operation mode information.

  In the modified prediction mode index shown in FIGS. 24 and 25, for example, in the case of the first L0-L1 bi-prediction mode, it corresponds to the modified prediction mode index 20 in FIG. Further, in the case of the second L0-L1 bi-prediction mode, it corresponds to the modified prediction mode index 22. Further, in the case of the third L0-L1 bi-prediction mode, it corresponds to the modified prediction mode index 21. Further, in the case of the fourth L0-L1 bi-prediction mode, it corresponds to the modified prediction mode index 23.

  The operation mode corresponding to the modified prediction mode index shown in FIG. 24 is preferably used in a configuration such as that shown in FIGS. 8C, 9C, 16C, and 17C. . The operation modes corresponding to the modified prediction mode index shown in FIG. 25 are, for example, FIG. 8 (A), FIG. 8 (B), FIG. 9 (A), FIG. 9 (B), FIG. 16B, 17A, and 17B may be used.

(Embodiment 3) Association with merge / AMVP.
Two new motion vector information prediction techniques have been studied in the standardization work of moving picture coding in ISO / IEC and ITU-T.

  The first technique is, for example, motion vector information belonging to an encoded adjacent block around the target prediction block as shown in FIG. 26, or an encoded reference picture as shown in FIG. Alternatively, for motion vector information belonging to the peripheral block of the anchor block in the same spatial position as the prediction target block in another picture that is a non-reference picture and has a different time from the target picture, This is a technique for evaluating the amount of generated code when applied as a candidate and specifying predicted motion vector information.

  FIG. 26 and FIG. 27 show an example of adjacent blocks that are candidates for motion vector predictor information. FIG. 26 is an example of an adjacent block in the same picture, and FIG. 27 is an example of a peripheral block at the same position of another picture having a different time. The motion vector information of these blocks is used as a candidate for predicted motion vector information, and predicted motion vector information that minimizes the generated code amount of the difference value between the motion vector information of the block to be processed and the predicted motion vector information candidate is selected. . Residual motion vector information, which is a difference value from the selected predicted motion vector information, and additional information related to the adjacent block corresponding to the predicted motion vector information are encoded and transmitted if necessary. As a motion prediction mode for specifying the operation of the first technique, it is referred to as an advanced motion vector prediction (hereinafter, AMVP) mode for convenience.

  For example, as disclosed in Japanese Patent Laid-Open No. 10-276439, the second technique is that if the motion information of the processing target block and the surrounding already-encoded adjacent blocks are the same, the processing target block is The motion information of the adjacent block is used for encoding without encoding its own motion information. Thus, this is a technique for reducing the amount of code of motion information by encoding additional information that designates an adjacent block having motion information to be referenced. Such a method is called “merge” and has attracted attention as a method for reducing the code amount of motion information. As a motion prediction mode for specifying the operation of the second technique, it is called a merge (or merge) mode for convenience.

  FIG. 28A is a diagram in which encoding modes are roughly classified based on encoding prediction modes, including the above-described two techniques. In the case of inter-image prediction (Inter mode), first, it is divided into a merge mode and a non-merge mode depending on whether or not the merge method is used. The non-merge mode is a mode in which motion detection such as block matching is actually performed, which can be called a motion detection mode. The merge mode is further divided into a skip mode and a non-skip mode. In the skip mode, the prediction signal predicted by the encoding information of the adjacent block to be referred to is a decoded picture. Therefore, the residual information is 0 and encoding transmission of the residual information is not required, and the reference destination of the encoding information is set. In this mode, only the index, which is reference information to be expressed, is encoded and transmitted. The coding information in the skip mode is the same as that in the merge mode. Here, the case where the residual information is encoded and transmitted is limited to the merge mode.

  FIG. 28B shows a table in which signals to be encoded in the Inter mode are classified. Unlike the other modes, the skip mode does not require the coded transmission of the image residual signal. Therefore, the skip mode is determined before the determination in the branch order of the prediction mode in FIG. In this way, the code amount of the flag used for mode determination can be suppressed.

  In the moving picture coding apparatus 100 and the moving picture decoding apparatus 200 according to the third embodiment, the two new techniques as described above are applied to the coding-side motion prediction unit 104 and the decoding in the first and second embodiments. The configuration introduced into the side motion prediction unit 208 can be adopted. Also, the two new techniques as described above and other motion prediction processes can be configured to be able to encode and decode by switching the operation according to the motion prediction mode.

  For example, the operation of the moving picture coding apparatus according to the third embodiment shown in FIG. 3 can be configured to use prediction motion vector information candidates specified by the AMVP mode. In step S102, the motion prediction unit 104 applies the AMVP mode as the predetermined motion prediction process, and sequentially supplies at least one motion vector predictor candidate identified by the AMVP process to the reference analysis unit 105. After that, the process proceeds to the steps shown in FIG. 3, and in step S105, the mode determination unit 108 sets the operation mode generated based on the candidate motion vector information in the AVMP process of the motion prediction unit 104. In response, the reference information or the corrected reference information, the generated residual motion vector information, and the prediction signal are acquired, and the target signal is acquired from the input unit 101.

  Thereafter, the mode determination unit 108 performs a mode determination process and a comparison determination for determining the operation mode. Here, when the information amount of the currently acquired reference information or corrected reference information, the generated residual motion vector information, and the prediction signal is larger than the information amount of the other operation modes, the operation mode has not yet been determined. If there is no (NO in S106), the process returns to step S102. By returning to step S102, the motion prediction unit 104 supplies another reference motion vector information candidate identified by the AMVP process to the reference analysis unit 105. Such processing is repeated, and in step S106, the mode determination unit 108 identifies an operation mode that minimizes the amount of information. Subsequent steps can be configured to perform the same operation as in FIG. 3 and complete the basic operation in the moving picture coding apparatus 100 of the third embodiment.

  Similarly, for example, the operation of the moving picture coding apparatus 100 according to the third embodiment shown in FIG. 3 can be configured to use the predicted motion vector information candidates specified by the merge mode. In step S <b> 102, the motion prediction unit 104 applies the merge mode as the predetermined motion prediction process, and sequentially supplies at least one predicted motion vector information candidate specified by the merge process to the reference analysis unit 105. Thereafter, the process proceeds to the steps shown in FIG. 3. In step S105, the mode determination unit 108 responds to the operation mode generated based on the motion vector predictor 104 that is a candidate for predicted motion vector information in the merge process. Further, reference information or modified reference information, generated residual motion vector information, and a prediction signal are acquired, and a target signal is acquired from the input unit 101.

  Thereafter, the mode determination unit 108 performs a mode determination process and a comparison determination for determining the operation mode. Here, when the information amount of the currently acquired reference information or corrected reference information, the generated residual motion vector information, and the prediction signal is larger than the information amount of the other operation modes, the operation mode has not yet been determined. If there is no (NO in S106), the process returns to step S102. By returning to step S <b> 102, the motion prediction unit 104 supplies another prediction motion vector information candidate specified by the merge process to the reference analysis unit 105. Such processing is repeated, and in step S106, the mode determination unit 108 identifies an operation mode that minimizes the amount of information. Subsequent steps can be configured to perform the same operation as in FIG. 3 and complete the basic operation in the moving picture coding apparatus 100 of the third embodiment.

  The following effects are obtained by the configuration and operation of the moving picture coding apparatus 100 and the moving picture decoding apparatus 200 according to the third embodiment.

  First, by applying the AMVP mode to the configuration and operation of the moving picture coding apparatus 100 and the moving picture decoding apparatus 200 according to the third embodiment, the first modified prediction signal of the moving picture coding apparatus 100 according to the third embodiment. In the generation unit 106 and / or the second modified prediction signal generation unit 107, based on at least one prediction motion vector information candidate carefully selected in the AMVP mode, the obtained information on the list side of the prediction motion vector information candidates is used as a reference. The modified prediction signal generation process can be performed. As a result, candidates that can be used as predicted motion vector information are increased by the AMVP mode, so that the corrected predicted signal generation process is performed using the predicted motion vector information supplied in a predetermined motion prediction process that is not in the AMVP mode. What is different from the corrected predicted motion vector information obtained in this way can be obtained. Therefore, different motion vector information, residual motion vector information, and prediction signals can be obtained than in the case of not being in the AMVP mode. Among such different information, the motion vector is determined by specifying a motion vector predictor information candidate that minimizes the amount of information of the reference information or the corrected reference information, the generated residual motion vector information, and the predicted signal. Will be able to. That is, the best result can be selected and encoded from the results of more various operation modes.

  In addition, since at least one candidate of predicted motion vector information carefully selected in the AMVP mode is obtained, a corrected predicted signal generation process can be performed as more likely predicted motion vector information. Therefore, the accuracy of the generated modified predicted motion vector information can be increased. As a result, a motion search is performed based on the corrected predicted motion vector information to determine the motion vector information, and the information amount of the residual motion vector information obtained by subtracting the corrected predicted motion vector information from the motion vector information is reduced. Will be able to. In addition, since the accuracy of the modified prediction motion vector information is increased, the processing amount of the motion search itself can be suppressed without significantly reducing the quality of the prediction signal.

  In addition, since more reliable predicted motion vector information can be obtained by the AMVP mode, the quality of the prediction signal on the list side of the obtained predicted motion vector information can be improved. Then, the quality of the combined prediction signal obtained by combining the prediction signal on the list side of the obtained prediction motion vector information and the prediction signal on the other list side can be improved. Here, the prediction signal on the other list side is obtained based on the motion vector information by deriving the modified prediction motion vector information on the other list side from the prediction motion vector information, specifying the motion vector information It is the prediction signal on the other list side. Accordingly, since the quality of the combined prediction signal is improved as compared with the case where the AMVP mode is not used, the amount of information of the obtained residual signal can be further reduced.

  In addition, by applying the merge mode, at least the first modified prediction signal generation unit 106 and / or the second modified prediction signal generation unit 107 of the video encoding device 100 according to Embodiment 3 is carefully selected according to the merge mode. Based on a single prediction motion vector information candidate, a modified prediction signal generation process can be performed with reference to information on the acquired list of prediction motion vector information candidates. As a result, candidates that can be used as predicted motion vector information increase in the merge mode, so that the corrected predicted signal generation process is performed using the predicted motion vector information supplied in a predetermined motion prediction process that is not in the merge mode. What is different from the corrected predicted motion vector information obtained in this way can be obtained. Therefore, different motion vector information, residual motion vector information, and prediction signals can be obtained than in the case where the mode is not the merge mode. Among such different information, the motion vector is determined by specifying a motion vector predictor information candidate that minimizes the amount of information of the reference information or the corrected reference information, the generated residual motion vector information, and the predicted signal. Will be able to. That is, the best result can be selected and encoded from the results of more various operation modes.

  Moreover, since at least one candidate of predicted motion vector information carefully selected in the merge mode is obtained, a corrected predicted signal generation process can be performed as more likely predicted motion vector information. Therefore, the accuracy of the generated modified predicted motion vector information can be increased. As a result, a motion search is performed based on the corrected predicted motion vector information to determine the motion vector information, and the information amount of the residual motion vector information obtained by subtracting the corrected predicted motion vector information from the motion vector information is reduced. Will be able to. In addition, since the accuracy of the modified prediction motion vector information is increased, the processing amount of the motion search itself can be suppressed without significantly reducing the quality of the prediction signal.

  Further, since more predictive motion vector information can be obtained by the merge mode, the quality of the prediction signal on the list side of the obtained motion vector predictor information can be improved. Then, the quality of the combined prediction signal obtained by combining the prediction signal on the list side of the obtained prediction motion vector information and the prediction signal on the other list side can be improved. Here, the prediction signal on the other list side is obtained based on the motion vector information by deriving the modified prediction motion vector information on the other list side from the prediction motion vector information, specifying the motion vector information It is the prediction signal on the other list side. Therefore, since the quality of the synthesized prediction signal is improved as compared with the case where the mode is not the merge mode, the information amount of the obtained residual signal can be further reduced.

  Furthermore, the reference information or the corrected reference information, which is an obtained result, is generated by using the AMVP mode, the merge mode, or another predetermined motion prediction mode while switching according to the information amount of the obtained result. The operation mode in which the residual motion vector information and the amount of information of the prediction signal are minimized can be specified, and the encoding efficiency can be further improved.

(Embodiment 4) Description of code string and syntax.
The moving picture encoding apparatus 100 according to Embodiment 4 can be configured to generate and output a code string as shown in FIG. 29, for example. Also, the moving picture decoding apparatus 200 according to Embodiment 4 can be configured to acquire such a code string and correctly decode it. The fourth embodiment can be configured to generate the code string as shown in FIG. 29 with the components as in the first to third embodiments.

  This code string is preferably configured as follows. FIG. 29 is a conceptual diagram for explaining the configuration of a code string generated by the fourth embodiment. For example, the generated code string is roughly composed of header information 2901 and data information 2902. The header information 2901 can be configured to include a bit string 2903 as a flag for indicating that the header information 2901 has been encoded by the moving image encoding apparatus 100 according to the fourth embodiment. In addition, a bit string as a use identification flag is included in the header information in a finer processing unit so that it can be determined whether or not the encoding by the moving image encoding apparatus 100 of Embodiment 4 has been used in a certain predetermined unit. Can be included.

  For example, in order to control the use of coding by the moving picture coding apparatus 100 of the fourth embodiment in units of pictures or slices, the header information 2904 including the header information 2904 and the data information 2905 includes A bit string 2906 as a flag for controlling use of encoding by the moving picture encoding apparatus 100 according to the fourth aspect may be included. Further, in order to control the use of coding by the moving picture coding apparatus 100 according to the fourth embodiment in units of smaller processing units, such as macroblocks and prediction blocks, macroblocks including header information 2907 and data information 2908 The prediction block header information 2907 may include a bit string 2909 as a flag for controlling the use of coding by the moving picture coding apparatus 100 according to the fourth embodiment. In addition, the encoded bit string may be configured such that mode information for specifying the operation mode of the fourth embodiment is stored as a bit string for such header information. Further, for such header information, a coded bit string is configured so that parameter information for controlling the operation and initial state of each part of the moving picture coding apparatus 100 according to Embodiment 4 is stored as a bit string. May be.

  Next, a syntax that is a predetermined syntax structure of the code string will be described. The code string shown in FIG. 29 may be configured to include a syntax as shown in FIGS. 30 (A) and 30 (B), for example. FIGS. 30A and 30B are a part of syntax that is going to be defined by HEVC, which is a moving image coding under standardization work in ISO / IEC and ITU-T. This is a syntax for indicating the syntax of header information and data information required for each prediction block (hereinafter referred to as PU).

  In the syntaxes of FIGS. 30A and 30B, first, skip_flag included in the syntax of a processing unit larger than the unit of PU is referred to. Here, the processing unit larger than the unit of PU, that is, the syntax higher than the PU includes a unit of coding block (coding unit or CU), a unit of coding tree (coding tree), and a unit of slice (slice). , A slice header unit, a picture parameter set unit, and a sequence (sequence parameter set) unit may be used. The unit may be positioned higher than other units. Absent. With regard to skip_flag, the discussion proceeds assuming that it is defined in the syntax of the CU.

  When skip_flag is true “1” and the total number of adjacent block candidates to be referred to exceeds NumMergeCand, the syntax element merge_idx of the index of the merge list that is a list of adjacent block candidates that refer to the encoding information is set. Is done. When there is one NumMergeCand, it becomes an adjacent block to be referred to, so that a candidate for an adjacent block to be referred to is determined without transmitting merge_idx. Next, when the prediction mode of the prediction block is intra-picture prediction (Intra mode), syntax elements necessary for the Intra mode are installed, but the details are omitted. Next, when the prediction mode of the prediction block is inter-picture prediction (Inter mode), a flag merge_flag indicating whether or not the mode is the merge mode is set. When merge_flag is true “1” and the total number NumMergeCand of adjacent block candidates to be referred to exceeds one, the merge mode is applied, and the index of the merge list index that is a list of adjacent block candidates that reference the encoding information A syntax element merge_idx is set. When there is one NumMergeCand, it becomes an adjacent block to be referred to, so that a candidate for an adjacent block to be referred to is determined without transmitting merge_idx. When merge_flag is false “0”, the merge mode is not applied, normal motion vector detection is performed, and encoded information such as a motion vector and a reference picture number is transmitted.

  For each prediction block, based on the prediction direction LX (X = 0 or 1), a syntax element ref_idx_lX representing a reference picture number obtained by motion vector detection, a motion vector, and a prediction motion vector of neighboring blocks around the prediction block The syntax element mvd_lX [] [] [i] of the difference motion vector is established. Here, X represents a prediction direction with 0 or 1, i represents a differential motion vector component, i = 0 represents an x component, and i = 1 represents a y component. Next, when the total number NumMvpCand (LX) of motion vector predictor candidates exceeds one, an MVP list index syntax element mvp_idx_lX, which is a list of motion vector predictor candidates to be referred to, is installed. NumMvpCand (LX) represents a function for calculating the total number of motion vector predictor candidates for the prediction block in the prediction direction LX (X = 0 or 1), and the total number of motion vector predictor candidates according to the motion vector prediction method. When there is one NumMvpCand (LX, i), the index mvp_idx_lX is not set and is not encoded, and if the total number of motion vector predictor candidates is 1, one of them becomes a motion vector predictor, so mvp_idx_lX is transmitted The prediction motion vector candidate to be referred to is determined at least, and based on the syntax as described above, the HEVC stores encoding information related to the PU.

  In contrast to the HEVC syntax as shown in FIGS. 30A and 30B, the moving picture coding apparatus 100 and the moving picture decoding apparatus 200 according to the fourth embodiment use, for example, FIGS. 31A to 31C. 34A to 34D, a code string having a syntax as shown in FIGS. 34A to 34D can be generated and decoded.

  FIGS. 31A to 31C illustrate an example of syntax in the PU when the encoding mode is the merge mode and the encoding is performed according to the operation mode of the fourth embodiment. The syntax of most of FIGS. 31A to 31C is preferably the same as that of FIGS. In Embodiment 4, after merge_idx is installed, lp_bi_mode_flag installed in the syntax higher than PU is referred. This lp_bi_mode_flag is a flag indicating whether or not encoding has been performed by the moving image encoding apparatus 100 according to the fourth embodiment. If lp_bi_mode_flag is true “1”, then lp_bi_mode_idx is installed. This lp_bi_mode_idx is a modified prediction mode index for specifying an operation mode in the first and second modified prediction signal generation units of the moving picture coding apparatus 100 and the moving picture decoding apparatus 200 according to Embodiment 4.

  After that, it is determined whether or not modification reference information is required from the operation mode specified by lp_bi_mode_idx, and is stored as LpBiModList as internal information. From this LpBiModList, lp_bi_ref_idx0 is set when the necessary modification reference information is on the L0 side or on both the L0 side and the L1 side. Further, if the required modification reference information is L1 side or both L0 side and L1 side from this LpBiModList, lp_bi_ref_idx1 is set next. The lp_bi_ref_idx0 and lp_bi_ref_idx1 are indexes for specifying the modified reference information.

  Thereafter, mvd_lp for specifying one residual motion vector information used in the first and second modified prediction signal generation units of the video encoding device 100 and the video decoding device 200 according to the fourth embodiment. [] [] [0] and mvd_lp [] [] [1] are arranged. When lp_bi_mode_flag is false “0”, since the operation mode of the fourth embodiment is not used, the encoding information necessary for the operation mode of the fourth embodiment is not arranged.

  FIGS. 32A to 32C show another example of syntax in the PU when the encoding mode is the merge mode and the encoding is performed according to the operation mode of the fourth embodiment. In FIGS. 32A to 32C, while the merge mode continues in units of PUs, when using the operation mode of the fourth embodiment, only the lp_bi_mode_idx of the PU in which the merge mode is started is used. The syntax is such that lp_bi_mode_idx is not stored for each PU. By doing so, it is possible to reduce the amount of codes for transmitting an index for specifying the operation mode. Here, the continuation of the merge mode is managed as MergeContinue as internal information, and is determined as “1” when the merge mode is continuously generated twice or more, and “0” otherwise. It should be possible.

  33A to 33D show an example of syntax in the PU when the encoding mode itself is encoded in the operation mode of the fourth embodiment. In FIGS. 33A to 33D, in order to determine whether or not the operation mode of Embodiment 4 is used, Pred_LP_BI is added as a new mode to inter_pred_flag, and when inter_pred_flag is Pred_LP_BI. The syntax is such that encoding information required in the operation mode of the fourth embodiment is arranged. In this syntax, the encoding mode is not the merge mode, but the prediction motion vector information candidates can be used in the operation mode of the fourth embodiment using the merge process or the AMVP process. ing. That is, since the encoding mode is not the merge mode or AMVP mode, in the operation mode of the fourth embodiment, the value of merge_idx arranged in the merge mode as the original encoding mode or the original encoding mode is used. By setting a value different from the values of mvp_idx_l0 and mvp_idx_l1 arranged in the AMVP mode, it is possible to specify the prediction motion vector information effective in the operation mode of Embodiment 4 from the prediction motion vector information candidates It has become a structure.

  Further, from the lp_bi_mode_idx, in the operation mode of the fourth embodiment, the operation mode based on the merge mode or the operation mode based on the AVMP mode is specified and retained as BasePredMode which is internal information, and BasePredMode is Merge, That is, in the merge mode, merge_idx is arranged as shown in FIGS. In addition, when BasePredMode is other than Merge, that is, in the AMVP mode, mvp_idx_lp is arranged as shown in FIGS.

  34A to 34D show another example of syntax in the PU when the encoding mode itself is encoded in the operation mode of the fourth embodiment. In FIGS. 33 (A) to 33 (D), whether or not the operation mode of the fourth embodiment is used is determined based on inter_pred_flag. In FIGS. 34 (A) to 34 (D), syntax higher than PU is used. With reference to lp_bi_mode_flag installed inside, it is configured to determine whether or not the operation mode of the fourth embodiment is used. The other syntax is the same as that shown in FIGS.

  Further, in the syntaxes shown in FIGS. 31A to 31C to FIGS. 34A to 34D, lp_bi_ref_idxl0 and lp_bi_ref_idxl1 may not be arranged, and may be specified from merge_idx and mvp_idx_lp. . Further, in the syntaxes shown in FIGS. 31A to 31C to FIGS. 34A to 34D, lp_bi_ref_idxl0 and lp_bi_ref_idxl1 are not arranged, and based on a predetermined reference analysis process and based on a predetermined rule. The reference destination may be specified.

  Further, in the fourth embodiment, for example, in the syntax as shown in FIGS. 31 (A) to (C) to FIGS. 34 (A) to (D), it is based on the syntax that combines the parts added as the fourth embodiment. You may comprise so that a code sequence may be produced | generated and decoded. Also, the syntaxes shown in FIGS. 31A to 31C to FIGS. 34A to 34D are examples, and in the configurations of the first to third embodiments, the encoding side and the decoding side are used. In this case, any syntax can be used as long as encoding information necessary for the operation mode can be stored.

  The following effects are obtained by the configuration and operation of the moving picture coding apparatus 100 and the moving picture decoding apparatus 200 according to Embodiment 4 that generate and decode a code string including the syntax as described above.

  First, for example, whether or not the operation mode of the fourth embodiment is performed when the encoding mode is the merge mode by using a code string including syntax as shown in FIGS. 31 (A) to (C). Can be switched. In addition, when the encoding mode is the merge mode (hereinafter, the encoding merge mode for distinction) and the operation mode of the fourth embodiment is performed, the merge_idx specified in the encoding merge mode is used to determine that of the fourth embodiment. Among the motion vector predictor candidates used in the operation mode, the motion vector predictor information and the reference information that are best in the coding merge mode are identified from the candidates and used in the operation mode of the fourth embodiment. Will be able to. Therefore, since it is not necessary to newly include information for specifying predicted motion vector information and reference information in the code string in the operation mode of the fourth embodiment, for example, as shown in FIGS. By using a syntax that does not place lp_bi_ref_idxl0 and lp_bi_ref_idxl1, the amount of code required for the information for specifying the predicted motion vector information and the reference information can be reduced, and the code of the code string Efficiency can be improved.

  Also, for example, by using a code string including syntax as shown in FIGS. 31A to 31C, whether or not to perform the operation mode of the fourth embodiment is controlled in units of the encoding merge mode. Will be able to. Thereby, it is possible to reduce the frequency of including the coding information required in the operation mode of Embodiment 4 in the code string, and to improve the coding efficiency of the code string.

  Also, it is assumed that the predicted motion vector information and reference information that are best in the coding merge mode are not the best predicted motion vector information and reference information in the operation mode of Embodiment 4, but the best among other candidates. Even if there is something, lp_bi_merge_idx is arranged after lp_bi_mode_idx in the syntax as shown in FIGS. 31A to 31C or the syntax as shown in FIGS. As a result, the predicted motion vector information that is the best in the operation mode of the fourth embodiment can be newly selected from the candidates. Thereby, the information amount of residual motion vector information, a prediction signal, or a synthetic | combination prediction signal can be reduced, and the encoding efficiency of a code sequence can be improved. Here, lp_bi_merge_idx, lp_bi_ref_idxl0, and lp_bi_ref_idxl1 are the best in the operation mode of the fourth embodiment by using any one of these, a combination thereof, or a syntax that can be arranged. Predicted motion vector information and reference information can be obtained.

  Further, for example, mvd_lp for specifying one piece of residual motion vector information can be included by using a code string including syntax as shown in FIGS. 31 (A) to (C). Then, the motion vector information is generated by deriving the corrected motion vector information from the predicted motion vector information that is best in the coding merge mode, and adding the residual motion vector information, and the motion vector information is used to generate the motion vector predictor. A prediction signal on the list side different from the information can be generated. As a result, the motion vector information for specifying the prediction signal can be corrected, the L0 side and L1 side prediction signals obtained in the original encoding merge mode, and the synthesis obtained by synthesizing them. It becomes possible to obtain a prediction signal and a synthesized prediction signal with better quality than the prediction signal.

  Also, it is assumed that the predicted motion vector information and reference information that are best in the coding merge mode are not the best predicted motion vector information and reference information in the operation mode of Embodiment 4, but the best among other candidates. In the case where there is something, by placing lp_bi_merge_idx after lp_bi_mode_idx in the syntax as shown in FIGS. 31 (A) to (C) or the syntax as shown in FIGS. 31 (A) to (C), Prediction motion vector information that is best selected in the operation mode of Embodiment 4 can be newly selected from candidates, residual motion vector information is identified from mvd_lp, and a prediction motion vector selected in the encoding merge mode From the information candidates, the best predicted motion vector information and reference information are identified, and the corrected predicted motion vector information is derived based on these information. Generates motion vector information by calculation, it is possible to generate a prediction signal of the different lists side to the predicted motion vector information by the motion vector information. As a result, the motion vector information for specifying the prediction signal can be corrected, the L0 side and L1 side prediction signals obtained in the original encoding merge mode, and the synthesis obtained by synthesizing them. It becomes possible to obtain a prediction signal and a synthesized prediction signal with better quality than the prediction signal.

  Also, by using a code string including syntax as shown in FIGS. 32A to 32C, only one operation mode of the fourth embodiment is included while the encoding merge mode continues. By doing so, it becomes possible not to include the operation mode every time a predetermined processing unit, for example, PU unit. As a result, the amount of code required for the information for specifying the operation mode can be reduced, and the coding efficiency of the code string can be improved.

  33A to 33D, the encoding mode itself is a code string including syntax that is encoded in the operation mode of the fourth embodiment, so that the encoding merge mode is performed. However, prediction motion vector information and reference information candidates obtained by the merge mode, and prediction motion vector information obtained by the AMVP mode although the encoding mode is not the AMVP mode (hereinafter, the encoded AMVP mode for distinction). Also, from among the candidates for reference information, the predicted motion vector information that is the best in the operation mode of the fourth embodiment can be newly selected from the candidates. Thereby, the information amount of residual motion vector information, a prediction signal, or a synthetic | combination prediction signal can be reduced, and the encoding efficiency of a code sequence can be improved. Here, lp_bi_merge_idx, lp_bi_ref_idxl0, and lp_bi_ref_idxl1 are the best in the operation mode of the fourth embodiment by using any one of these, a combination thereof, or a syntax that can be arranged. Predicted motion vector information and reference information can be obtained.

  Further, for example, by using a code string including syntax as shown in FIGS. 33A to 33D, whether or not to perform the operation mode of the fourth embodiment is determined, the unit of the operation mode of the fourth embodiment. It becomes possible to control with. As a result, it is possible to increase the frequency with which the encoding information required in the operation mode of the fourth embodiment is included in the code string, and fine control can be performed. Can be improved.

  In addition, referring to lp_bi_mode_flag installed in the syntax higher than the PU as shown in FIGS. 34A to 34D, it is determined whether or not the operation mode of the fourth embodiment is used. By using a code string including a simple syntax, whether or not to perform the operation mode of the fourth embodiment can be controlled in a larger processing unit. Thereby, it is possible to reduce the frequency of including the coding information required in the operation mode of Embodiment 4 in the code string, and to improve the coding efficiency of the code string.

  The encoded stream, which is the code sequence of the moving image output from the moving image encoding apparatus 100 according to the fourth embodiment described above, has specific data so that it can be decoded according to the encoding method used there. The moving picture decoding apparatus 200 according to the fourth embodiment corresponding to the moving picture encoding apparatus 100 according to the fourth embodiment can decode an encoded stream having this specific data format.

(Embodiment 5)
When a wired or wireless network is used to exchange an encoded stream between the moving image encoding apparatus 100 and the moving image decoding apparatus 200, the encoded stream is converted into a data format suitable for the transmission form of the communication path. May be transmitted. In this case, for example, as shown in FIG. 35, the encoded stream output from the moving image encoding apparatus 100 according to the fifth embodiment is converted into encoded data having a data format suitable for the transmission form of the communication channel and transmitted to the network. And a moving image receiving apparatus 3550 that receives encoded data from the network, restores the encoded data to the encoded stream, and supplies the encoded stream to the moving image decoding apparatus 200 according to the fifth embodiment.

  The moving image transmitting apparatus 3500 includes a memory that buffers the encoded stream output from the moving image encoding apparatus 100 according to the fifth embodiment, a packet processing unit 3511 that packetizes the encoded stream, and packetized encoded data. And a transmission unit 3512 that transmits the data via a network. The moving image reception device 3550 includes a reception unit 3561 that receives packetized encoded data via a network, a memory that buffers the received encoded data, a packet process on the encoded data, and an encoded stream. And a packet processing unit 3562 provided to the moving picture decoding apparatus 200 according to the fifth embodiment.

(Embodiment 6)
FIG. 36 is a block diagram illustrating an example in which the video encoding device 100 and the video decoding device 200 according to Embodiments 1 to 5 are configured by an information processing device 3600 including a computer. The information processing apparatus 3600 according to the sixth embodiment includes a central processing control device 3603 typified by a CPU, a recording medium and a communication device 3606, and a temporary storage device 3605 typified by a memory and an external storage device typified by an HDD. The processing in the first to fifth embodiments may be executed by software processing using the encoding program and decoding program stored in 3604. In addition, the program can be provided by being recorded on a computer-readable recording medium, provided from a server through a wired or wireless network, or provided as data broadcasting of terrestrial or satellite digital broadcasting. is there.

  In addition, the moving image encoding device, the moving image encoding method, the moving image decoding device, the moving image decoding method, and the moving image decoding program according to the first to sixth embodiments of the present invention are applied to moving images. Any device, method, program, system, and the like for encoding and decoding an image may be used, and the present invention is not particularly limited. For example, broadcasting devices represented by TV, mobile phones, smartphones, portable music players, game devices, video conference devices, monitoring devices, DVD-R / RW, BD-R / RW, HDD, SD, holographic memory, etc. The present invention may be applied to an image recording / reproducing apparatus using a recording medium that can be additionally written and rewritten, an imaging recording / reproducing apparatus such as a digital camera or a camcorder, an image recording / editing apparatus such as authoring, a moving image distribution apparatus, and the like.

  The present invention has been described based on some embodiments. It is understood by those skilled in the art that these embodiments are exemplifications, and that various modifications can be made to combinations of the respective constituent elements and processing processes, and such modifications are also within the scope of the present invention. By the way.

  DESCRIPTION OF SYMBOLS 100 moving image encoding apparatus, 101 input part, 102 picture storage part, 103 motion information storage part, 104 motion estimation part, 105 reference analysis part, 106 1st correction prediction signal generation part, 107 2nd correction prediction signal generation part, 108 mode determination unit, 109 subtraction unit, 110 residual coding unit, 111 residual decoding unit, 112 addition unit, 113 code string generation unit, 114 code string storage unit, 115 code amount control unit, 116 output unit, 117 pieces Direction prediction signal generation unit, 118 bi-prediction signal generation unit, 119 intra prediction unit, 120 encoding control unit, 200 moving image decoding device, 201 input unit, 202 picture storage unit, 203 decoded sequence generation unit, 204 residual decoding unit , 205 addition unit, 206 mode identification unit, 207 motion information storage unit, 20 Motion prediction unit, 209 reference analysis unit, 210 first modified prediction signal generation unit, 211 second modified prediction signal generation unit, 212 output unit, 213 unidirectional prediction signal generation unit, 214 bi-prediction signal generation unit, 215 intra prediction unit 216 Decoding control unit.

Claims (8)

Obtain a code string, perform decoding string generation processing based on a predetermined syntax structure, and at least residual motion vector information, information after residual coding, and code belonging to a region to be decoded from the code string A decoded sequence generation unit for generating the optimization parameters;
A predetermined motion prediction process is performed based on the motion information belonging to the decoded area to generate first reference information, and the first reference picture specified based on the first reference information A motion prediction unit that generates predicted motion vector information;
Second prediction motion vector information in a second reference picture specified based on second reference information is derived from the first prediction motion vector information, and the derived second prediction motion vector information is stored in the remaining prediction motion vector information. A modified prediction signal generation unit that generates motion vector information in addition to the difference motion vector information, and generates a prediction signal from the second reference picture based on the generated motion vector information;
A residual decoding unit that performs a predetermined residual decoding process on the residual encoded information and generates decoded residual information;
An addition unit that generates a decoded signal by adding the prediction signal generated by the modified prediction signal generation unit to the decoded residual information;
An image decoding apparatus comprising: The modified prediction signal generator is
Generating a first prediction signal from the first reference picture based on the first prediction motion vector information;
Generating a second prediction signal from the second reference picture based on the second prediction motion vector information;
Combining the first prediction signal and the second prediction signal to generate a combined prediction signal;
The adding unit is
Adding a combined prediction signal generated by the modified prediction signal generation unit to decoding residual information to generate a decoded signal;
The image decoding apparatus according to claim 1. This image decoding apparatus
The encoding parameter is a decoding parameter, and an encoding mode, a motion prediction mode, and an operation mode are specified from the decoding parameter,
The modified prediction signal generator is
The second predictive motion vector information corresponding to the operation mode is derived by the derivation process specified based on the operation mode, and the derived second motion vector information corresponding to the operation mode is derived. Prediction residual motion vector information is added to residual motion vector information to generate motion vector information corresponding to the motion mode, and based on the generated motion vector information, the second reference picture corresponds to the motion mode. Generating a second prediction signal;
The image decoding apparatus according to claim 1, wherein the image decoding apparatus is an image decoding apparatus. This image decoding apparatus
The encoding parameter is a decoding parameter, and an encoding mode, a motion prediction mode, and an operation mode are specified from the decoding parameter,
The motion prediction unit
Operates based on a motion prediction mode for distinguishing a plurality of motions, identifies motion vector prediction candidates corresponding to the motion prediction mode by motion prediction processing identified by the motion prediction mode,
The modified prediction signal generator is
From the prediction motion vector information corresponding to the motion prediction mode, the prediction motion vector information corresponding to the operation mode is specified as the first prediction motion vector information.
The image decoding apparatus according to claim 1, wherein the image decoding apparatus is an image decoding apparatus. This image decoding apparatus
The encoding parameter is a decoding parameter, and an encoding mode, a motion prediction mode, and an operation mode are specified from the decoding parameter,
When the encoding mode is AMVP (Advanced Motion Vector prediction) mode,
The motion prediction unit
According to the AMVP mode of the encoding mode, the best predicted motion vector information is determined from the predicted motion vector information candidates by AMVP processing,
The modified prediction signal generator is
In accordance with the AMVP mode of the encoding mode, the determined predicted motion vector information is acquired and used as the first predicted motion vector information.
The image decoding apparatus according to claim 1, wherein the image decoding apparatus is an image decoding apparatus. This image decoding apparatus
The encoding parameter is a decoding parameter, and an encoding mode, a motion prediction mode, and an operation mode are specified from the decoding parameter,
When the encoding mode is merge mode,
The motion prediction unit
According to the merge mode of the encoding mode, the best prediction motion vector information is determined from prediction motion vector information candidates by merge processing,
The modified prediction signal generator is
According to the merge mode of the encoding mode, to obtain the determined motion vector predictor information, the first motion vector predictor information,
The image decoding apparatus according to claim 1, wherein the image decoding apparatus is an image decoding apparatus. Obtain a code string, perform decoding string generation processing based on a predetermined syntax structure, and at least residual motion vector information, information after residual coding, and code belonging to a region to be decoded from the code string Generating activation parameters; and
A predetermined motion prediction process is performed based on the motion information belonging to the decoded area to generate first reference information, and the first reference picture specified based on the first reference information Generating predicted motion vector information;
Second prediction motion vector information in a second reference picture specified based on second reference information is derived from the first prediction motion vector information, and the derived second prediction motion vector information is stored in the remaining prediction motion vector information. Generating motion vector information in addition to the difference motion vector information, and generating a prediction signal from the second reference picture based on the generated motion vector information;
Performing a predetermined residual decoding process on the residual encoded information to generate decoded residual information;
Adding the predicted signal to the decoded residual information to generate a decoded signal;
An image decoding method comprising: Obtain a code string, perform decoding string generation processing based on a predetermined syntax structure, and at least residual motion vector information, information after residual coding, and code belonging to a region to be decoded from the code string A function to generate a parameter,
A predetermined motion prediction process is performed based on the motion information belonging to the decoded area to generate first reference information, and the first reference picture specified based on the first reference information A function of generating predicted motion vector information;
Second prediction motion vector information in a second reference picture specified based on second reference information is derived from the first prediction motion vector information, and the derived second prediction motion vector information is stored in the remaining prediction motion vector information. A function of generating motion vector information in addition to the difference motion vector information, and generating a prediction signal from the second reference picture based on the generated motion vector information;
A function of performing a predetermined residual decoding process on the residual encoded information and generating decoded residual information;
A function of generating the decoded signal by adding the prediction signal to the decoded residual information;
An image decoding program that causes a computer to execute the above.