JP5888279B2 - Moving picture coding apparatus, moving picture coding method, moving picture coding program, transmission apparatus, transmission method, and transmission program - Google Patents

Description

  The present invention relates to a moving picture coding technique using motion compensated prediction, and more particularly to a moving picture coding technique for coding filter information that functions as a loop filter.

  MPEG-4 AVC / H.MP, which is a general moving image compression coding. In H.264 (hereinafter referred to as AVC), a deblocking filter is used as a loop filter. The AVC deblocking filter is a technique that divides a target image into fine blocks and corrects pixels near the boundary between adjacent blocks for each block.

  The deblocking filter corrects errors in the vicinity of block boundaries caused by differences in quantization width and prediction methods between blocks in units of pixels, and uses the corrected picture as a reference picture for motion compensation prediction to improve prediction efficiency. It is improving.

  Patent Document 1 discloses a method for controlling an AVC deblocking filter.

WO2009 / 093472

  Since the AVC deblocking filter only corrects pixels near the block boundary, it cannot correct pixels inside the block that are not block boundaries. Therefore, when an error occurs in a pixel inside a block that is not a block boundary, the prediction efficiency may not be improved.

  Under such circumstances, the present inventors recognize the necessity of improving the prediction efficiency even in the case where an error occurs in a pixel inside a block that is not a block boundary in a moving image coding method using a loop filter. It came.

  The present invention has been made in view of such a situation, and an object of the present invention is to improve prediction efficiency by correcting pixels inside a block that is not a block boundary even when an error occurs in the pixel inside the block that is not a block boundary. An object of the present invention is to provide a moving picture coding technique that improves the coding efficiency and improves the coding efficiency.

In order to solve the above-described problem, a moving picture coding apparatus according to an aspect of the present invention is a moving picture coding apparatus that performs coding in units of blocks constituting a picture, wherein the picture is larger than the block. An area setting unit (102) for setting information of divided areas divided into a plurality of tiles , and encoding block image data of a coding target to generate a block code string, and the coding target A block encoding unit (102) for acquiring decoded image data of a block, a loop filter unit (104) for filtering the decoded image data based on a filter parameter of the block to be encoded, and An acquisition flag indicating whether or not the filter parameter of the block to be encoded is acquired from the adjacent encoded block In addition, when the acquisition flag indicates that the acquisition flag is not acquired, a filter parameter encoding unit (105) that encodes information on the filter parameter to generate a filter code string, and a picture parameter set indicating picture characteristics A multiplexing unit (107) that encodes the information of the divided regions, multiplexes the picture parameter set into a code stream, and multiplexes the filter code string and the block code string into the code stream in units of blocks. ). The filter parameter encoding section (105), when the previous SL coded block checks whether contact with the tile boundary, serving as the coding target block is in contact with the tile boundaries, codes the acquired flag Turn into.

Yet another aspect of the present invention is a video encoding method. This method is a moving image coding method that performs coding in units of blocks that constitute a picture, in which the picture is set with information on a divided area obtained by dividing the picture into a plurality of tiles having a size larger than the block. A block encoding step of generating a block code string by encoding the image data of the block to be encoded, and obtaining decoded image data of the block to be encoded; and the decoded A loop filter step for filtering image data based on a filter parameter of the block to be encoded, and whether or not the filter parameter of the block to be encoded is acquired from an adjacent encoded block When encoding the acquisition flag and indicating that the acquisition flag is not acquired, the flag A filter parameter encoding step that encodes information relating to a filter parameter to generate a filter code string, and encodes the information of the divided region into a picture parameter set indicating picture characteristics, and converts the picture parameter set into a code string stream And a multiplexing step of multiplexing the filter code string and the block code string into the code string stream in units of blocks. The filter parameter coding step, a pre-SL coding target block checks whether contact with the tile boundary, serving as the encoding target block may contact the tile boundaries, encodes the acquired flag.

Yet another embodiment of the present invention is a transmission device. The apparatus includes a packet processing unit that packetizes an encoded stream encoded by a moving image encoding method that encodes a block unit constituting a picture to obtain encoded data, and the packetized encoded data A transmission unit for transmission. The encoding method includes: an area setting step for setting information of a divided area obtained by dividing the picture into a plurality of tiles larger than the block; and encoding image data of a block to be encoded into a block A block encoding step for generating a code string and obtaining decoded image data of the block to be encoded, and filtering the decoded image data based on a filter parameter of the block to be encoded A loop filter step to perform and encoding an acquisition flag indicating whether or not to acquire the filter parameter of the block to be encoded from an adjacent encoded block, and indicates that the acquisition flag is not acquired If the filter parameter information is encoded, a filter code string is generated. A filter parameter encoding step, encoding the information of the divided region into a picture parameter set indicating picture characteristics, multiplexing the picture parameter set into a code string stream, and combining the filter code string and the block code string A multiplexing step of multiplexing the code stream in block units. The filter parameter coding step, a pre-SL coding target block checks whether contact with the tile boundary, serving as the encoding target block may contact the tile boundaries, encodes the acquired flag.

Yet another embodiment of the present invention is a transmission method. In this method, a packet processing step for packetizing an encoded stream encoded by a moving image encoding method for encoding in units of blocks constituting a picture to obtain encoded data, and the packetized encoded data A transmission step of transmitting. The encoding method includes: an area setting step for setting information of a divided area obtained by dividing the picture into a plurality of tiles larger than the block; and encoding image data of a block to be encoded into a block A block encoding step for generating a code string and obtaining decoded image data of the block to be encoded, and filtering the decoded image data based on a filter parameter of the block to be encoded A loop filter step to perform and encoding an acquisition flag indicating whether or not to acquire the filter parameter of the block to be encoded from an adjacent encoded block, and indicates that the acquisition flag is not acquired If the filter parameter information is encoded, a filter code string is generated. A filter parameter encoding step, encoding the information of the divided region into a picture parameter set indicating picture characteristics, multiplexing the picture parameter set into a code string stream, and combining the filter code string and the block code string A multiplexing step of multiplexing the code stream in block units. The filter parameter coding step, a pre-SL coding target block checks whether contact with the tile boundary, serving as the encoding target block may contact the tile boundaries, encodes the acquired flag.

  It should be noted that any combination of the above-described constituent elements and a conversion of the expression of the present invention between a method, an apparatus, a system, a recording medium, a computer program, etc. are also effective as an aspect of the present invention.

  According to the present invention, even when an error occurs in a pixel inside a block that is not a block boundary, an improvement in prediction efficiency can be realized by correcting the pixel inside the block that is not a block boundary.

It is a figure explaining the relationship between CTB and CU. It is a figure explaining a tile. It is a figure explaining the structure of the moving image encoder which concerns on 1st Embodiment. It is a flowchart explaining operation | movement of the moving image encoder which concerns on 1st Embodiment. It is a figure explaining a CTB address and a CTB code order. It is a figure explaining the structure of the filter parameter determination part of FIG. It is a flowchart explaining operation | movement of the filter parameter determination part of FIG. It is a flowchart explaining operation | movement of the edge type filter part of FIG. It is a figure explaining the adjacent pixel A and the adjacent pixel B according to edge angle. It is a figure explaining the pixel A and the pixel X of CTB in which the right side and the lower side touch the tile boundary. It is a flowchart explaining operation | movement of the band type filter part of FIG. It is a figure explaining CTB adjacent to encoding object CTB and encoding object CTB. 4 is a flowchart for explaining the operation of a filter parameter encoding unit in FIG. 3. It is a flowchart explaining derivation | leading-out of the effectiveness of adjacent CTB. It is a figure explaining syntax. It is a figure explaining syntax. It is a figure explaining an example of a structure of the encoding stream by 1st Embodiment. It is a figure explaining the structure of the moving image decoding apparatus which concerns on 1st Embodiment. It is a flowchart explaining operation | movement of the moving image decoding apparatus which concerns on 1st Embodiment. It is a flowchart explaining the operation | movement of the filter parameter decoding part of FIG. It is a flowchart explaining operation | movement of the moving image encoder which concerns on 2nd Embodiment. It is a figure explaining an example of the structure of the encoding stream by 2nd Embodiment. It is a flowchart explaining operation | movement of the moving image decoding apparatus which concerns on 2nd Embodiment. It is a figure explaining the syntax of PPS which concerns on 3rd Embodiment. It is a flowchart explaining operation | movement of the moving image encoder which concerns on 3rd Embodiment. It is a flowchart explaining operation | movement of the moving image decoding apparatus which concerns on 3rd Embodiment. It is a flowchart explaining operation | movement of the edge type filter part of 3rd Embodiment. It is a flowchart explaining derivation | leading-out of the effectiveness of adjacent CTB in 3rd Embodiment. It is a figure explaining the syntax of PPS which concerns on 4th Embodiment. It is a flowchart explaining derivation | leading-out of the effectiveness of adjacent CTB in 4th Embodiment. It is a figure for demonstrating the filter type which can be utilized according to the position of CTB in 5th Embodiment. It is a flowchart explaining operation | movement of the edge type filter part in 5th Embodiment. It is a figure explaining the code sequence of loopflitter_type_idx which concerns on 6th Embodiment. It is a flowchart explaining derivation | leading-out of the effectiveness of adjacent CTB in 6th Embodiment. It is a figure explaining the syntax of the loop filter unit which concerns on 7th Embodiment. It is a figure explaining the code sequence of loopflitter_type_idx which concerns on 7th Embodiment It is a flowchart explaining operation | movement of the moving image decoding apparatus 200 which concerns on 4th Embodiment.

[First Embodiment]
First, a technique that is a premise of the embodiment of the present invention will be described.

  Called AVC in 2003 by the joint work of the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC) Joint Technical Committee (ISO / IEC) and the International Telecommunication Union Telecommunication Standardization Sector (ITU-T) A moving picture coding system (14496-10 in ISO / IEC and H.264 standard number in ITU-T) has been established as an international standard. In AVC, a deblocking filter is used as a loop filter.

  First, the deblocking filter will be described. The deblocking filter adaptively changes the strength of the deblocking filter on a block-by-block basis, depending on the quantization parameter and information such as intra prediction (intra-screen prediction) or inter prediction (motion compensation prediction). While controlling the ON / OFF of the deblocking filter on a pixel-by-pixel basis based on the difference in pixel values between pixels near the boundary, errors in the vicinity of the block boundary caused by differences in quantization width and prediction methods between blocks This is a technique for correcting and reducing errors. Deblocking filter improves subjective image quality by reducing block boundary errors that are visually noticeable, and also improves inter prediction efficiency by using an image with reduced block boundary errors as a reference image There is.

  As control information for the deblocking filter, in AVC, a flag for turning on / off the deblocking filter in units of slices and information for adjusting the degree of application of the deblocking filter (ON / OFF in units of pixels) are in the slice header. .

  In recent years, it has become common for hardware and CPUs to adopt a configuration based on parallel processing, and it is an important issue that video coding also supports parallel processing. This problem increases as the size of the image increases. For this reason, in AVC or the like, which is a general moving image encoding method, a moving image picture is divided into a plurality of regions, and slices are used to encode or decode each region in parallel. When processing slices in parallel, deblocking filter processing at each slice boundary becomes an issue. In the AVC deblocking filter, the vertical filter and the horizontal filter are sequentially processed for each block in the raster scan order from the upper left block of the image, and therefore, the deblocking filter at the region boundary cannot be processed in advance. For this reason, AVC introduces a flag for turning on / off the deblocking filter in units of slices, and is used for easily realizing parallel processing including a loop filter.

  However, since the AVC deblocking filter can correct only pixels near the block boundary, there is a possibility that the prediction efficiency may not be improved when an error occurs in a pixel inside the block that is not the block boundary. Therefore, a SAO filter (Sample Adaptive Offset Filter) is introduced as a filter for correcting not only the block boundary but also the pixels inside the block.

  Here, an outline of the SAO filter will be described. The SAO filter is a filter that corrects a pixel by adding a predetermined offset value to all pixels inside the block as well as pixels near the block boundary. Since all pixels in the block are processed, it is difficult to perform detailed adaptive processing using information such as differences between blocks as in the case of a deblocking filter. It is necessary to improve the performance by adjusting the offset value for each area. Although the accuracy of error correction is improved by adjusting the offset value in units of fine areas, it is necessary to suppress this because the amount of code for encoding the offset values in units of fine areas increases. Details of the SAO filter will be described later.

  Hereinafter, a moving picture coding apparatus and a moving picture decoding apparatus capable of effectively operating the SAO filter are realized.

(Encoding tree block)
In the present embodiment, an input image signal (picture) is divided into coding tree block units and subjected to coding and decoding processing. The coding tree block (CTB) is hierarchically divided into four to form a coding unit (CU). Each CU in the CTB is predictively encoded using intra prediction or inter prediction.

  In this embodiment, the CTB is 64 (pixels) × 64 (pixels), and the maximum number of divisions is three. Therefore, there are 8 × 8, 16 × 16, 32 × 32, and 64 × 64 as CU sizes. FIG. 1 is a diagram for explaining the relationship between CTB and CU.

  In this embodiment, CTB is 64 (pixels) × 64 (pixels) and the maximum number of divisions of CTB is 3, but the combination is not limited to this. Further, the CU may be further divided into one to four prediction blocks, and each prediction block may be predictively encoded using intra prediction or inter prediction.

(Picture)
A picture is a single image serving as a unit for encoding and decoding. The pictures include an I picture using only intra prediction, a P picture using intra prediction and inter prediction, and a B picture using intra prediction and inter prediction of uni prediction and bi prediction.

(tile)
In the present embodiment, a tile that can define a picture as an area divided in the vertical direction and the horizontal direction is introduced.

  Next, tiles will be described with reference to the drawings. FIG. 2 is a diagram illustrating tiles. FIG. 2A shows an example in which a picture is divided into two in the vertical direction, and the picture is composed of tile 0 and tile 1. FIG. 2B shows an example in which the picture is divided into two parts in the vertical and horizontal directions, and the picture is composed of tile 0 to tile 3. The tiles in the picture are processed in raster scan order. Also, CTBs in the tile are processed in raster scan order. Here, for ease of explanation, the tiles have the same size in the picture, but the tiles need not have the same size. As shown in FIGS. 2A and 2B, a tile is defined as a rectangular region.

(slice)
In the present embodiment, a slice includes CTBs in a tile in raster scan order. Here, it is assumed that there is one slice in a picture, but a slice can be composed of one or more CTBs in one tile. That is, the tile can be divided into one or more slices. A slice can also include multiple tiles. That is, a plurality of tiles can be combined into one slice.

  In this embodiment, by introducing tiles, it is possible to divide a region in the vertical direction that cannot be realized by slicing. By dividing the region in the vertical direction, the number of horizontal pixels of each tile can be reduced, and the amount of line memory can be reduced when each tile is independently processed (parallel processing). On the other hand, since the vertical boundary becomes longer by dividing the tile, the processing of the vertical boundary is recognized as coding efficiency, visual discomfort, and the like. The larger the screen, the greater the effect. In addition, since CTBs in tiles are processed in raster scan order and slices include CTBs in tiles in raster scan order, combining the tiles and slices has the same problem with the vertical boundaries of slices. May occur. Note that, when an area is configured by only slices, the vertical boundary is generated only at the start CTB of the slice, so the problem with respect to the vertical boundary is small.

  Hereinafter, with reference to the drawings, details of a moving image encoding apparatus, a moving image encoding method, a moving image encoding program, a moving image decoding apparatus, a moving image decoding method, and a moving image decoding program according to a preferred embodiment of the present invention will be described. explain. In the description of the drawings, the same elements are denoted by the same reference numerals and redundant description is omitted.

(Configuration of moving picture coding apparatus 100)
FIG. 3 is a diagram for explaining the configuration of the moving picture coding apparatus 100 according to the first embodiment. The moving image encoding apparatus 100 is an apparatus that encodes an input moving image in units of one picture.

  The moving image encoding apparatus 100 is realized by hardware such as an information processing apparatus including a CPU (Central Processing Unit), a frame memory, and a hard disk. The moving image encoding apparatus 100 realizes functional components described below by operating the above components.

  A moving image encoding apparatus 100 according to the present embodiment includes an image data acquisition unit 101, a CTB encoding unit 102, a filter parameter determination unit 103, a filter parameter encoding unit 105, a loop filter 104, a frame memory 106, and a code string multiplexing unit. 107 and an encoding control unit 110. The encoding control unit 110 includes a tile setting unit 111, a loop filter setting unit 112, and an area information setting unit 113.

(Function and operation of moving picture coding apparatus 100)
FIG. 4 is a flowchart for explaining the operation of the moving picture coding apparatus 100 according to the first embodiment. Hereinafter, the function and operation of each unit will be described with reference to FIGS. 3 and 4. Here, the size of the input picture is 768 × 384, and the bit depth is 8 bits. In this case, there are 12 CTBs in the horizontal direction and 6 in the vertical direction, and there are 72 CTBs in the picture. Hereinafter, the number of CTBs in the horizontal direction of the picture is NumCtbInPicWidth. Since the bit depth is 8 bits, the pixel value is 0 to 255. The input picture is encoded with M = 3 and N = 15, which is a general GOP (Group of Picture), and the I picture is encoded with only I slice, the P picture is encoded with only P slice, and the B picture is encoded with only B slice. Shall be Here, the picture size is set to 768 × 384 and set to a multiple of CTB, but the present invention is not limited to this. Further, although GOP is set to M = 3 and N = 15, it may be set to N = 1 or the like which is only an I picture, and is not limited thereto.

  First, the tile setting unit 111 sets a tile (S100). Thereafter, unless otherwise specified, it is assumed that a picture is composed of four tiles as shown in FIG.

  Next, the loop filter setting unit 112 sets loop filter information (S101). Here, the loop filter information will be described. The loop filter information is a loop filter interleave flag (loopfilter_interleave_flag). The loop filter interleave flag is a flag indicating whether the encoding unit of the filter parameter is multiplexed in units of CTB or multiplexed in units of pictures. If the loop filter interleave flag is 1, it indicates that the filter parameter is multiplexed in CTB units, and if the loop filter interleave flag is 0, it indicates that the filter parameters are multiplexed in picture units. Hereinafter, it is assumed that the flag is 1-bit information having a value of 0 or 1.

  In the present embodiment, loopfilter_interleave_flag is set to 1. That is, the filter parameters are multiplexed on a CTB basis.

  First, the region information setting unit 113 sets a CTB address and a CTB code order for CTB (S102). Here, the CTB address and the CTB code order will be described. FIG. 5 is a diagram for explaining the CTB address and the CTB code order.

  FIG. 5A shows the CTB address set for each CTB in the picture. As shown in FIG. 5A, the CTB address is set by incrementing by 1 in the raster scan order for each CTB in the picture. FIG. 5B shows the CTB code order set for each CTB in the picture when the picture is composed of four tiles as shown in FIG. The first tile has a CTB code order from 0 to 17, the second tile has a CTB code order from 18 to 35, the third tile has a CTB code order from 36 to 53, and the fourth tile has The CTB code order is 54 to 71. As shown in FIG. 5B, the CTB code order is set by incrementing by one in the raster scan order for each CTB in the tile. As shown in FIG. 5, when the tile is divided in the vertical direction, a block in which the CTB address (CTB address order) and the order in which the CTB is encoded is not the same occurs. As shown in FIG. 5, the CTB address is determined in the raster scan order corresponding to the picture, and the CTB code order is determined in the raster scan order corresponding to the tile.

  As shown in FIG. 5, the CTB address does not depend on the tile and is set according to the size of the picture, and the CTB code order is set according to the size of the picture and the size of the tile. That is, the CTB address indicates the position in the picture, and the positional relationship of the CTB in the picture can be obtained by the CTB address. Each or both of the CTB address and CTB code order may be used to identify which tile a CTB belongs to.

  Also, the encoding control unit 110 sets and manages information necessary for encoding a picture, such as SPS, PPS, and slice header, as necessary. The SPS, PPS, and slice header will be described later. Then, an image data acquisition unit 101, a CTB encoding unit 102, a filter parameter determination unit 103, a filter parameter encoding unit 105, a loop filter 104, a frame memory 106, and a code string multiplexing unit 107 are controlled to generate an encoded stream. To do.

  Note that information necessary for encoding pictures such as tiles, loop filter information, SPS, PPS, and slice headers is shared within the moving image encoding apparatus 100, and the description of the flow of these data is particularly important. Omitted unless otherwise noted.

  Next, the following processing is repeated for all CTBs in the picture in accordance with the CTB code order (S103 to S111).

  Next, the image data acquisition unit 101, based on the position in the picture of the encoding target CTB, which is the CTB to be encoded, from the image data input from the terminal 10, the original CTB corresponding to the encoding target CTB. Image data is acquired (S104). Then, the image data acquisition unit 101 supplies the CTB original image data to the CTB encoding unit 102. The image data acquisition unit 101 supplies the CTB original image data to the filter parameter determination unit 103. Here, the tile to which the encoding target CTB belongs is set as the encoding target tile, and the slice to which the encoding target CTB belongs is set as the encoding target slice.

  Next, the entropy code part is set (S105). Here, the setting of the entropy code unit will be described. If the encoding target CTB is the first CTB of a slice or tile, the CTB encoding unit 102 initializes an entropy encoding unit inside the CTB encoding unit 102, and the filter parameter encoding unit 105 is a filter parameter encoding unit 105. The entropy code part inside is initialized. It is assumed that the entropy coding unit inside the CTB coding unit 102 and the entropy coding unit inside the filter parameter coding unit 105 are CABAC of arithmetic coding used in AVC.

  Next, the CTB encoding unit 102 performs intra-screen prediction (intra prediction) and motion compensation prediction (inter prediction) on the CTB original image data supplied from the image data acquisition unit 101 based on the slice type, and generates an error. The signal is calculated, the error signal is encoded using orthogonal transformation or quantization processing, a CTB code string is generated, and local decoding is performed to generate decoded image data (S106). Note that the encoding process performed by the CTB encoding unit 102 may be any method that may cause encoding distortion due to the hierarchical encoding of the CTB and quantization, and detailed description thereof is omitted here. To do. Also, since local decoding is a technique performed in general video encoding, detailed description thereof is omitted here, but the decoded image data is obtained by decoding an encoded stream output from the video encoding device 100. It is image data obtained at this time. The CTB encoding unit 102 supplies the generated CTB code sequence to the code sequence multiplexing unit 107. Also, the CTB encoding unit 102 supplies the decoded image data to the filter parameter determination unit 103 and the loop filter 104.

  Next, the filter parameter determination unit 103 determines a filter parameter based on the CTB original image data supplied from the image data acquisition unit 101 and the decoded image data supplied from the CTB encoding unit 102 (S107). The filter parameter determination unit 103 supplies the filter parameters to the filter parameter encoding unit 105 and the loop filter 104. Details of the filter parameter determination unit 103 will be described later.

  The loop filter 104 executes a loop filter on the decoded image data supplied from the CTB encoding unit 102 based on the filter parameter supplied from the filter parameter determination unit 103 (S108). New decoded image data obtained by executing the loop filter is supplied to the frame memory 106.

  Next, the filter parameter encoding unit 105 encodes the filter parameter supplied from the filter parameter determination unit 103 into a loop filter unit (S109), and supplies the loop filter unit to the code string multiplexing unit 107. The filter parameter encoding unit 105 and the loop filter unit will be described later.

  Next, the code string multiplexing unit 107 combines the CTB code string supplied from the CTB encoding unit 102 and the loop filter unit supplied from the filter parameter encoding unit 105 together with the SPS, PPS, slice header, and the like as necessary. The encoded stream is multiplexed (S110), and the encoded stream is output from the terminal 11. Details of the code string multiplexing unit 107 will be described later.

  When the process is completed for all CTBs in the picture, the process ends.

(Filter parameter determination unit 103)
Next, a detailed configuration of the filter parameter determination unit 103 will be described. FIG. 6 is a diagram for explaining the configuration of the filter parameter determination unit 103. As shown in FIG. 6, the filter parameter determination unit 103 includes a filter parameter setting unit 120, a loop filter execution unit 121, an error measurement unit 122, and a filter parameter determination unit 123. The terminal 12 is connected to the CTB encoding unit 102, the terminal 13 is connected to the image data acquisition unit 101, and the terminal 14 is connected to the filter parameter encoding unit 105 and the loop filter 104. The loop filter execution unit 121 includes a filter type determination unit 130, an edge type filter 131, and a band type filter 132.

  Subsequently, the operation of the filter parameter determination unit 103 will be described. FIG. 7 is a flowchart for explaining the operation of the filter parameter determination unit 103. Hereinafter, the operation of the filter parameter determination unit 103 will be described with reference to FIGS. 6 and 7.

  First, the filter parameter setting unit 120 generates a filter parameter candidate list based on the decoded image data supplied from the terminal 12 and the original image data supplied from the terminal 13 (S120), and the generated filter parameter candidate list Is supplied to the loop filter execution unit 121. Details of the filter parameter setting unit 120 will be described later.

  Next, the following processing is repeated for all filter parameters registered in the filter parameter candidate list (S121 to S126).

  First, the loop filter execution unit 121 applies a loop filter to the decoded image data supplied from the terminal 12 based on the filter parameter to generate new decoded image data (S122), and converts the decoded image data into an error. It supplies to the measurement part 122. Details of the loop filter execution unit 121 will be described later.

  The error measurement unit 122 calculates an error evaluation value between the decoded image data supplied from the loop filter execution unit 121 and the original image data supplied from the terminal 13 (S123). Here, the error evaluation value is the sum of absolute differences (SAD) of the decoded image data and the original image data in the block.

  Next, the error measuring unit 122 checks whether the error evaluation value is the minimum (S124). If the error evaluation value is minimum (Y in S124), the filter parameter is supplied to the filter parameter determination unit 123, and the filter parameter determination unit 123 holds the filter parameter (S125).

  If the error evaluation value is not minimum (N in S124), the next filter parameter is processed (S126). When the processing of all the filter parameters registered in the filter parameter candidate list is completed, the filter parameter having the smallest error evaluation value held by the filter parameter determination unit 123 is determined as the filter parameter (S127), and the processing Exit.

  Here, SAD is used as the error evaluation value, but it is only necessary to be able to calculate the filter parameter, and the present invention is not limited to this, and SSD or the like may be used. Further, the evaluation may be performed in consideration of not only an error between the decoded image data and the original image data but also a code amount necessary for transmitting the filter parameter.

(Filter parameter setting unit 120)
Next, details of the filter parameter setting unit 120 will be described.

  First, filter parameters will be described. The filter parameters include a filter type, bandwidth, band position, offset 0, offset 1, offset 2, offset 3, and offset 4.

  There are six filter types: filter type 0, filter type 1, filter type 2, filter type 3, filter type 4, and filter type 5.

  Filter type 0 indicates that no loop filter is processed, filter type 1 indicates that the loop filter is a band type filter, and filter types 2 to 5 indicate that the loop filter is an edge type filter.

  If the error evaluation value between the decoded image data and the original image data is equal to or smaller than the threshold value, the filter parameter setting unit 120 generates a filter parameter candidate list including only the filter type 0 as the filter type, and the error evaluation value is equal to or smaller than the threshold value. Otherwise, a filter parameter candidate list including all combinations of filter parameters that can be taken by the filter parameters is generated.

(Loop filter execution unit 121)
Next, details of the loop filter execution unit 121 will be described. The loop filter execution unit 121 includes a filter type determination unit 130, an edge type filter 131, and a band type filter 132.

Subsequently, the operation of the loop filter execution unit 121 will be described.
First, the filter type determination unit 130 determines the filter type. If the filter type is 0, the decoded image data supplied from the terminal 12 is output to the error measuring unit 122 as decoded image data after applying the filter. If the filter type is 1, the band type filter 132 applies a band type filter to the decoded image data supplied from the terminal 12 based on the filter parameter, and outputs the decoded image data after the filter application to the error measurement unit 122. To do. If the filter type is 2 to 5, the edge type filter 131 applies the edge type filter to the decoded image data supplied from the terminal 12 based on the filter parameter, and the decoded image data after the filter application is used as the error measurement unit 122. Output to.

(Edge type filter 131)
Next, details of the edge type filter 131 will be described. The edge type filter 131 includes an edge angle setting unit 141, an adjacent pixel setting unit 142, an edge type determination unit 143, an edge offset value setting unit 144, and an edge offset value addition unit 145.

  FIG. 8 is a flowchart for explaining the operation of the edge type filter 131. Hereinafter, the operation of the edge type filter 131 will be described with reference to FIG.

  First, the edge angle setting unit 141 determines the edge angle (S130). The edge angle n is a filter type n. n is 2,3,4,5. Next, the following processing is repeated for the 64 pixels X in the coding tree block in the raster scan order (S131 to S139). X is 0, 1, 2,...

  First, the adjacent pixel setting unit 142 sets the adjacent pixel A and the adjacent pixel B according to the edge angle (S132). The adjacent pixel A and the adjacent pixel B corresponding to the edge angle will be described later. Next, the adjacent pixel setting unit 142 checks whether the adjacent pixel A or the adjacent pixel B is outside the picture (S133). If the adjacent pixel A or the adjacent pixel B is outside the picture (Y in S133), the edge offset value setting unit 144 sets the offset value to 0 (S137), and the edge offset value adding unit 145 sets the offset value to the pixel X. Are added to calculate the processing pixel X ′ (S138). If the adjacent pixel A or the adjacent pixel B is not out of the picture (N in S133), it is checked whether loopfilter_interleave_flag is 1 (S140). If loopfilter_interleave_flag is 1 (Y in S140), the adjacent pixel setting unit 142 checks whether the adjacent pixel A or the adjacent pixel B is outside the tile (S134). The adjacent pixel A or the adjacent pixel B being out of the tile means that the adjacent pixel A or the adjacent pixel B is included in a CTB belonging to a tile different from the pixel X. If loopfilter_interleave_flag is not 1 (N in S140), the edge type determination unit 143 determines the edge type (S135).

  If the adjacent pixel A or the adjacent pixel B is outside the tile (Y in S134), the edge offset value setting unit 144 sets the offset value to 0 (S137), and the edge offset value adding unit 145 sets the offset value to the pixel X. Are added to calculate the processing pixel X ′ (S138). If the adjacent pixel A or the adjacent pixel B is not outside the tile (N in S134), the edge type determination unit 143 determines the edge type (S135). The edge type determination will be described later. Next, the edge offset value setting unit 144 sets an offset value (S136). Next, the edge offset value adding unit 145 calculates the processing pixel X ′ by adding the offset value to the pixel X (S138). The calculation of the processing pixel X ′ will be described later.

(Adjacent pixel A and adjacent pixel B according to edge angle)
FIG. 9 is a diagram for explaining the adjacent pixel A and the adjacent pixel B according to the edge angle. In the case of the edge angle 1, the adjacent pixel A is the left pixel of the pixel X, and the adjacent pixel B is the right pixel of the pixel X (FIG. 9A). In the case of the edge angle 2, the adjacent pixel A is a pixel above the pixel X, and the adjacent pixel B is a pixel below the pixel X (FIG. 9B). In the case of the edge angle 3, the adjacent pixel A is the upper right pixel of the pixel X, and the adjacent pixel B is the lower left pixel of the pixel X (FIG. 9C). In the case of the edge angle 4, the adjacent pixel A is the upper left pixel of the pixel X, and the adjacent pixel B is the lower right pixel of the pixel X (FIG. 9D). As described above, by causing the edge type filter to act on the horizontal direction, the vertical direction, the right diagonal direction, and the left diagonal direction, it is possible to correct errors generated in the respective directions. Here, edge angle 1, edge angle 2, edge angle 3, and edge angle 4 are assigned to filter type 2, filter type 3, filter type 4, and filter type 5, respectively.

  As described above, in the edge type filter 131, when the filter parameter is multiplexed in CTB units (loopfilter_interleave_flag is 1), if the adjacent pixel A or the adjacent pixel B is outside the tile, the offset value is set to 0. By preventing the loop filter from acting, the operation of the loop filter can be completed in units of CTB. Therefore, as shown in FIG. 10, the pixel A and the pixel X before applying the filter, which are necessary for applying the loop filter to the pixel X in the CTB whose right side or lower side is in contact with the tile boundary, as shown in FIG. There is no need to be held across boundaries.

(Edge type judgment)
The edge type is determined by a code sum SS that is a value obtained by adding the sum of the difference between the pixel X and the adjacent pixel A and the difference between the pixel X and the adjacent pixel B. When the pixel value of the pixel m is P (m), the code sum SS is calculated by Expression (1).
SS = Sign (P (X) -P (A)) + Sign (P (X) -P (B)); Formula (1)
Here, Sign (i) is a function that returns -1 if the input value i is less than 0, and returns 1 if the input value i is 0 or more.

  The edge types when SS is 0, -2, -1, 1, 2, are 0, 1, 2, 3, 4, respectively. Offset 0, offset 1, offset 2, offset 3, and offset 4 are set as offset values of edge types 0, 1, 2, 3, and 4, respectively. Here, when SS is 0, since the pixel X, the adjacent pixel A, and the adjacent pixel B are arranged in a straight line, the offset 0 is set to 0 because there is no error. When SS is not 0, each of offset 0 to offset 4 is set to an appropriate value in order to correct distortion.

(Calculation of processing pixel X ′ of edge type filter)
The processing pixel X ′ after applying the edge type filter of the pixel X is calculated by Expression (2).
P (X ′) = P (X) + OFFSET [SS]; Formula (2)
Here, OFFSET [m] is an offset m (m = 0, 1,..., 4).

  As described above, in the edge type filter, it is possible to adaptively correct the error in units of pixels by the sign of the difference from the adjacent pixel in the horizontal direction, vertical direction, right diagonal direction, and left diagonal direction. It becomes.

(Band type filter 132)
Next, details of the band type filter 132 will be described. The band type filter 132 includes a hand width setting unit 151, a band position setting unit 152, a band offset value setting unit 153, and a band offset value adding unit 154.

  FIG. 11 is a flowchart for explaining the operation of the band type filter 132. Hereinafter, the operation of the band type filter will be described with reference to FIG.

  First, the hand width setting unit 151 sets a bandwidth (S150).

(Bandwidth)
Here, the bandwidth will be described. The bandwidth is a bandwidth when the range of pixel values that can be taken by the input image is divided into a plurality of bands. In this embodiment, the bandwidth is 32. In this embodiment, since the input image is 8 bits and has a pixel value range from 0 to 255, 0 to 31 (band 0), 32 to 63 (band 1), 64 to 95 (band 2), It is divided into eight bands 96 to 127 (band 3), 128 to 159 (band 4), 160 to 191 (band 5), 192 to 223 (band 6), and 224 to 255 (band 7).

  By setting the bandwidth in this way and increasing the bandwidth, a wide range of pixel values can be corrected at once, and by reducing the bandwidth, the pixel values in a narrow range can be corrected at once. can do. For example, when the range of pixel values that can be taken by a pixel in a picture is wide, the bandwidth is increased. When the range of pixel values that can be taken by a pixel in a picture is narrow, the bandwidth is reduced. In this way, the decoded image can be appropriately corrected according to the characteristics of the picture. In this embodiment, the bandwidth is described as a fixed value. However, the range of pixel values that can be taken by the input image may be divided into one or more bands, and may be 16 or 128. Also, the band type filter can be effectively operated by multiplexing and transmitting the bandwidth to PPS or the like and adaptively changing it according to the range of pixel values that can be taken by the pixels in the picture.

  Following step S150, the following processing is repeated for the three band positions (S151 to S158). Here, the three band positions will be described. The first band position is a band position included in the filter parameter. The second band position is a band position obtained by shifting the first band position by one. For example, if the first band position is band 2, the second band position is band 3. Similarly, the third band position is a band position obtained by shifting the second band position by one.

  The band position setting unit 152 sets the b-th band position (S152). Here, b is 0 and 1. The band position is a value indicating one of the bands divided by the band width, and is an integer from 0 to 7 here. For example, band position 0 indicates band 0 and band position 2 indicates band 3.

  Subsequent to step S152, the band offset value setting unit 153 sets an offset value for each band (S153). The offset value of the band indicated by the band position is set to offset 0, and the offset value of the band other than the band indicated by the band position is set to 0.

  Next, the following processing is performed in the raster scan order for the 64 pixels X in the encoding tree block of pixel values (S154 to S157). X is 0, 1,.

  First, the band offset value addition unit 154 determines xb that is a band to which the pixel X belongs (S155). Next, the band offset value addition unit 154 calculates the processing pixel X ′ by adding the offset value Ob [xb] of the band to which the pixel X belongs to the pixel X (S156).

(Calculation of processing pixel X ′ of band type filter)
The processed pixel X ′ of the pixel X after application of the band type filter is calculated by Expression (3).
P (X ′) = P (X) + Ob [xb]; Formula (3)

  As described above, since the band type filter can correct only pixels having a pixel value within a specific range, it has an effect of reducing noise that is likely to occur in a flat region. In particular, for a color difference signal having a resolution reduced from that of a luminance signal, such as 4: 2: 0 format, the effect of the band type filter is great because errors are more likely to occur in DC than in the luminance signal and there are few edges. Further, since adjacent pixels are not referred to, it is possible to realize a filtering processing amount and a memory amount smaller than edge type offset.

(Loop filter 104)
Next, the loop filter 104 will be described. Since the loop filter 104 has the same function and operation as the loop filter execution unit 121, description thereof is omitted here.

(Filter Parameter Encoding Unit 105)
Next, details of the filter parameter encoding unit 105 will be described. FIG. 12 is a diagram for explaining an encoding target CTB and a CTB adjacent to the encoding target CTB. FIG. 13 is a flowchart for explaining the operation of the filter parameter encoding unit 105. Hereinafter, the operation of the filter parameter encoding unit 105 will be described with reference to FIGS. 12 and 13.

  First, available_left_ctb_flag and available_above_ctb_flag, which are the validity of adjacent CTBs, are derived (S160). Next, it is checked whether available_left_ctb_flag is 1 (S161). If available_left_ctb_flag is 1 (Y in S161), it is checked whether the filter parameters of the CTB located on the left of the encoding target CTB and the CTB to be encoded are the same (S162). Here, the CTB located to the left of the encoding target CTB is A of CTB in FIG. If the filter parameters of the CTB to be encoded and the CTB located to the left of the encoding target CTB are the same (Y in S162), loopfilter_merge_left_flag is set to 1 and loopfilter_merge_left_flag is encoded (S163), and the process ends. If the filter parameters of the CTB located to the left of the encoding target CTB and the CTB to be encoded are not the same (N in S162), the loopfilter_merge_left_flag is set to 0 and the loopfilter_merge_left_flag is encoded (S164).

  When available_left_ctb_flag is not 1 (N of S161), following step S164, it is checked whether available_above_ctb_flag is 1 (S165). If available_above_ctb_flag is 1 (Y in S165), it is checked whether the filter parameters of the CTB to be encoded and the CTB located on the CTB to be encoded are the same (S166). Here, the CTB located above the encoding target CTB is B of the CTB in FIG. If the filter parameters of the CTB to be encoded and the CTB located above the encoding target CTB are the same (Y in S166), loopfilter_merge_above_flag is set to 1 and loopfilter_merge_above_flag is encoded (S167), and the process ends. If the filter parameters of the CTB to be encoded and the CTB located above the encoding target CTB are not the same (N in S166), the loopfilter_merge_above_flag is set to 0 and the loopfilter_merge_above_flag is encoded (S168). Following step S168, loopfilter_cod_param (), which is a loop filter encoding parameter, is encoded. Here, the loop filter encoding parameter is a code string obtained by encoding all elements of the filter parameter.

  Note that the filter parameter encoding unit 105 holds the filter parameters of all the CTBs in the picture in order to refer to the filter parameters of the CTBs located on the left and above the CTB to be encoded.

  Here, when encoding the filter parameters in CTB units, if all the filter parameters are encoded as loop filter encoding parameters, an enormous amount of code is required for encoding the filter parameters, and the image data is encoded. Therefore, it is impossible to allocate a sufficient amount of code to the CTB code string, which may result in a decrease in encoding efficiency. Further, when different filter parameters are used for CTBs belonging to the same object, a difference in the characteristics of the loop filter within the same object may be visually recognized as block distortion. Therefore, generally, there is a high possibility that the filter parameters are set so that the filter parameters are the same within the same object. Therefore, as described above, when the filter parameter of the encoding target CTB is the same as the filter parameter of the encoded CTB adjacent to the encoding target CTB, the encoding target CTB is used as the filter parameter of the encoding target CTB. By encoding a 1-bit flag indicating that the encoded CTB filter parameter adjacent to the code is used, the code amount required for the loop filter encoding parameter is reduced to 1 bit without visual degradation. It is possible to reduce the coding efficiency. Although details of the loop filter encoding parameter will be described later, the code amount required for encoding the loop filter encoding parameter is a code amount sufficiently larger than 1 bit. On the other hand, when the filter parameter of the encoding target CTB is not the same as the filter parameter of the encoded CTB adjacent to the encoding target CTB, the encoded parameter adjacent to the encoding target CTB is used as the filter parameter of the encoding target CTB. By encoding a 1-bit flag indicating that no CTB filter parameter is to be used and by encoding a loop filter encoding parameter, the filter parameter can be changed for each CTB.

  Here, derivation of available_left_ctb_flag and available_above_ctb_flag, which are the validity of adjacent CTBs, will be described. FIG. 14 is a flowchart for explaining the derivation of the effectiveness of adjacent CTBs. First, available_left_ctb_flag and available_above_ctb_flag are set to 1 (S180). Next, it is checked whether the left side of the encoding target CTB is in contact with the picture boundary (S181). If the left side of the encoding target CTB is in contact with the picture boundary (Y in S181), the available_left_ctb_flag is set to 0 (S184). If the left side of the encoding target CTB is not in contact with the picture boundary (N in S181), it is checked whether the left side of the encoding target CTB is in contact with the tile boundary (S182). Here, the fact that the left side of the encoding target CTB is in contact with the tile boundary means that the encoding target CTB and the CTB located to the left of the encoding target CTB belong to different tiles. On the other hand, the fact that the left side of the encoding target CTB is not in contact with the tile boundary means that the encoding target CTB and the CTB located to the left of the encoding target CTB belong to the same tile. If the left side of the encoding target CTB is in contact with the tile boundary (Y in S182), it is checked whether loopfilter_interleave_flag is 1 (S183). If loopfilter_interleave_flag is 1 (Y in S183), available_left_ctb_flag is set to 0 (S184). If loopfilter_interleave_flag is not 1 (N in S183), step S184 is skipped. When the left side of the encoding target CTB is not in contact with the tile boundary (N in S182), Step S183 and Step S184 are skipped.

  Next, it is checked whether the upper side of the encoding target CTB is in contact with the picture boundary (S185). If the upper side of the encoding target CTB is in contact with the picture boundary (Y in S185), the available_above_ctb_flag is set to 0 (S188). If the upper side of the encoding target CTB is not in contact with the picture boundary (N in S185), it is checked whether the upper side of the encoding target CTB is in contact with the tile boundary (S186). Here, the fact that the upper side of the encoding target CTB is in contact with the tile boundary means that the CTB located on the encoding target CTB and the encoding target CTB belong to different tiles. On the other hand, the fact that the upper side of the encoding target CTB is not in contact with the tile boundary means that the encoding target CTB and the CTB located on the encoding target CTB belong to the same tile. If the upper side of the CTB to be encoded is in contact with the tile boundary (Y in S186), it is checked whether loopfilter_interleave_flag is 1 (S187). If loopfilter_interleave_flag is 1 (Y in S187), available_above_ctb_flag is set to 0 (S188). If loopfilter_interleave_flag is not 1 (N in S187), step S188 is skipped. When the upper side of the encoding target CTB is not in contact with the tile boundary (N in S186), Step S187 and Step S188 are skipped.

  Here, when the filter parameters are multiplexed in CTB units (loopfilter_interleave_flag is 1) and the encoding process is performed independently (in parallel) in tile units, the left side of the encoding target CTB is in contact with the tile boundary. If it is permitted to obtain the filter parameters from the CTB located on the left for the encoding target CTB, the encoding target CTB, the filter parameter, and the CTB located on the left are determined until the filter parameter of the CTB located on the left is determined. Since it is not possible to check whether the filter parameters are the same, latency may occur. Therefore, for the CTB to be encoded whose CTB left side is in contact with the tile boundary, use of the filter parameter of the CTB located on the left is prohibited. Furthermore, by setting the available_left_ctb_flag to 0 so that the loopfilter_merge_left_flag is not encoded, the code amount of the loopfilter_merge_left_flag is reduced, and a decrease in encoding efficiency is suppressed. Also, by not encoding the loopfilter_merge_left_flag, it is possible to prohibit the use of the filter parameter of the CTB positioned explicitly on the left. Similarly, when the upper side of the CTB to be encoded is in contact with the tile boundary, the available_above_ctb_flag is set to 0 so that loopfilter_merge_above_flag is not encoded, thereby suppressing a decrease in encoding efficiency, and an explicitly located CTB The use of filter parameters can be prohibited.

(Syntax)
Next, the syntax according to the present embodiment will be described. In encoding, each element included in the syntax is encoded in an encoded stream. On the other hand, in decoding, each element included in the syntax is decoded from the encoded stream, and the same value as the element encoded in the encoding can be obtained. Thereafter, unless otherwise specified, each element constituting the syntax becomes a code string with fixed-length bits. 15 and 16 are diagrams for explaining the syntax. Hereinafter, the syntax will be described with reference to FIGS. 15 and 16. Here, for ease of explanation, each element constituting the syntax is a fixed-length bit and becomes a code string, but may be a variable-length bit.

  First, SPS (Sequence Parameter Set) will be described. SPS is a parameter set that defines a parameter group for determining characteristics of a sequence (encoded stream). The picture size, bit depth, CTB size, number of CTB divisions, and the like are defined.

  Next, PPS (Picture Parameter Set) will be described. FIG. 15A is a diagram for explaining an example of the syntax of PPS. PPS is a parameter set that defines a group of parameters for determining picture characteristics. In PPS, loopfilter_interleave_flag, pic_loopfilter_param (), tiles_coding_flag, num_tile_columns_minus1, num_tile_rows_minus1, uniform_spacing, and cin

  Tile_coding_flag, num_tile_columns_minus1, num_tile_rows_minus1, uniform_spacing_flag, column_width, and column_height are PPS syntax elements. tiles_coding_flag is 1 if the picture is composed of a plurality of tiles, and 0 is that the picture is not composed of tiles. num_tile_columns_minus1 and num_tile_rows_minus1 indicate numbers for dividing the tile in the vertical direction and the horizontal direction, respectively. uniform_spacing_flag is information for determining whether or not to divide a picture by a predetermined method. Here, the predetermined method indicates that the picture is divided into tiles having the same size. column_width [i] and column_height [i] indicate the width and height of the i-th tile, respectively. In the case of FIG. 2B, num_tile_columns_minus1 and num_tile_rows_minus1 are encoded as 1, and uniform_spacing_flag is encoded as 1. If the tile configuration is not changed in sequence units, the syntax elements of the tile information can be installed in the SPS.

(Slice header)
Next, the slice header will be described. FIG. 15B is a diagram for explaining an example of the syntax of the slice header. The slice header is header information that defines a parameter group for determining the characteristics of the slice. In the slice header, slice_address, slice_type, num_of_tiles, and tile_position are set according to the syntax, encoded in moving picture encoding, and decoded in moving picture decoding.

  The slice_address (slice address) indicates the CTB address of the first CTB included in the slice. For slice_type (slice type), I slice using only intra prediction, P slice using intra prediction and inter prediction (motion compensated prediction), intra prediction, and inter prediction of uni prediction and bi prediction are used. A value indicating a B slice. num_of_tiles indicates the number of tiles included in the slice. tile_position indicates the number of bytes from the first byte of the slice header of the first byte of the tile.

(Syntax for loop filter)
FIG. 16A is a diagram for explaining an example of the syntax of the picture loop filter parameter (pic_loopfilter_param ()) in the PPS. In the picture loop filter parameter, loopfilter_unit () is set according to the syntax, encoded in moving picture encoding, and decoded in moving picture decoding. num_of_ctbs is the total number of CTBs in the picture, and loopfilter_unit () of all the CTBs in the picture is stored in the PPS.

  FIG. 16B is a diagram for explaining an example of the syntax of the loop filter encoding parameter (loopfilter_cod_param ()). In the loop filter encoding parameter, loopfilter_type_idx, loopfilter_band_position, and loopfilter_offset are set according to the syntax, encoded in moving image encoding, and decoded in moving image decoding.

  loopfilter_type_idx is an index indicating the filter type of the filter parameter, and is an integer from 0 to 5. loopfilter_type_idx is a 3-bit code string. loopfilter_band_position is a band position of the filter parameter, and is a 3-bit code string indicating an integer from 0 to 7. loopfilter_offset is an offset of the filter parameter. When loopfilter_type_idx is 1, it is an integer from −32 to 31 and is a 6-bit code string, and when loopfilter_type_idx is from 2 to 5, it is an integer from 0 to 31. It becomes a 5-bit code string.

  Here, the code amount required for the edge type filter is 3 bits of loopfilter_type_idx and 5 bits of loopfilter_offset, which is 23 bits. The code amount required for the band type filter is 3 bits, 3 bits of loopfilter_type_idx, 3 bits of loopfilter_band_position, and 6 bits of loopfilter_offset, which is 24 bits.

  Here, in this embodiment, the offset numbers of the edge type filter and the band type filter are set to 4 and 3, respectively. This is because the amount of code required for each of the edge type filter and the band type filter is balanced because the filter parameter of the band type filter has a larger code amount per filter than the filter parameter of the edge type filter. By balancing the amount of code required for each of the edge type filter and the band type filter, the encoding time and the decoding time of the loop filter encoding parameters of the edge type filter and the band type filter can be balanced. Circuit design and software design can be facilitated. Here, the offset numbers of the edge type filter and the band type filter are set to 4 and 3, respectively. For example, the offset numbers of the edge type filter and the band type filter are set to 2 and 1, respectively. May be.

  FIG. 16C illustrates an example of the syntax of the loop filter unit (loopfilter_unit ()). In the loop filter unit, loopfilter_merge_left_flag, loopfilter_merge_above_flag, and loopfilter_cod_param () are installed according to the syntax, encoded in moving image encoding, and decoded in moving image decoding.

(Function and operation of code string multiplexing unit 107)
Hereinafter, the configuration of the encoded stream multiplexed by the code string multiplexing unit 107 will be described. FIG. 17 is a diagram for explaining an example of the configuration of the encoded stream according to the first embodiment. FIG. 17 shows the structure of an encoded stream when loopfilter_interleave_flag is 1. The SPS is encoded by the code string multiplexing unit 107 and multiplexed at the head of the encoded stream. The PPS is encoded by the code string multiplexing unit 107 and multiplexed at the head of the picture. The slice header is encoded by the code string multiplexing unit 107 and multiplexed at the head of the slice. Following the slice header, the following processing is repeated in the CTB code order for the CTBs belonging to the slice. First, the CTB loop filter unit is multiplexed, and then the CTB code string is multiplexed.

  Here, a general synchronization code is added to the SPS, PPS, and slice header, and the SPS, PPS, and slice header can be separated from the encoded stream. In addition, although a header is not given to the tile, it can be separated from the encoded stream by the tile_position of the slice header.

  As described above, by setting loopfilter_interleave_flag to 1 and multiplexing the CTB filter parameters in the CTB code order in units of CTBs, it is possible to realize a moving picture coding apparatus that suppresses transmission delay of filter parameters.

(Configuration of moving picture decoding apparatus 200)
FIG. 18 is a diagram illustrating the configuration of the video decoding device 200 according to the first embodiment. The moving picture decoding apparatus 200 is an apparatus that receives the encoded stream generated by the moving picture encoding apparatus 100 and reproduces a decoded picture in units of pictures.

  The moving picture decoding apparatus 200 is realized by hardware such as an information processing apparatus including a CPU (Central Processing Unit), a frame memory, and a hard disk. The moving picture decoding apparatus 200 realizes functional components described below by operating the above components.

  The moving picture decoding apparatus 200 according to the present embodiment includes a code string separation unit 201, a CTB decoding unit 202, a loop filter 203, a filter parameter decoding unit 204, a frame memory 205, and a decoding control unit 210.

(Function and operation of moving picture decoding apparatus 200)
The decoding control unit 210 manages information necessary for decoding pictures such as region information, loop filter information, SPS, PPS, and slice header, and includes a code string separation unit 201, a CTB decoding unit 202, a loop filter 203, The filter parameter decoding unit 204 and the frame memory 205 are controlled to decode the encoded stream. Note that information necessary for decoding pictures such as region information, loop filter information, SPS, PPS, and slice header is assumed to be shared in the video decoding device 200, and description of the flow of these data is omitted. .

  FIG. 19 is a flowchart for explaining the operation of the moving picture decoding apparatus 200 according to the first embodiment. FIG. 18 shows a decoding operation in units of pictures. Hereinafter, the function and operation of each unit will be described with reference to FIGS. 18 and 19.

  In the present embodiment, it is assumed that an encoded stream in which loopfilter_interleave_flag is set to 1 is input.

  First, the code string separation unit 201 acquires a picture stream from an encoded stream input from the terminal 20 (S500). Here, the picture stream is an encoded stream in units of pictures, and includes SPS when SPS is present before the first CTB code string of the picture stream.

  Next, the code string separation unit 201 decodes the SPS if the tile stream includes the SPS, and acquires the size of the picture. Further, the PPS is decoded according to the syntax of FIG. 15, and loop filter information and tile information are acquired (S501).

  The decoding control unit 210 derives region information from the picture size and tile information. Here, it is assumed that the same information as the region information set by the moving image encoding apparatus 100 is derived as the region information. Note that the relationship between the CTB address (CTB address order) set by the picture size and tile information and the order in which the CTB is encoded is also the moving picture coding apparatus 100 set by the moving picture coding apparatus 100. It is the same as the set relationship.

  Next, the following processing is repeatedly performed for all CTBs included in the picture stream (S502 to S508).

  First, if there is a slice header, the slice header is decoded according to the syntax of FIG. 15 to obtain a slice address, the slice address is set as a CTB address, and the CTB code order corresponding to the CTB address is obtained from the region information. If there is no slice header, the CTB code order is incremented by 1, and a CTB address corresponding to the CTB code order is obtained from the region information (S503). Here, the CTB address is derived from the CTB code order based on the relationship between the CTB address and the CTB code order shown in FIG. For example, the CTB address of a CTB whose CTB code order is 8 is 14. In the next CTB, the CTB code order is incremented by 1 to 9 and the CTB address is 15. As described above, the CTB address is obtained from the CTB code order based on the relationship between the CTB address and the CTB code order.

  Next, the CTB decoding unit 202 sets the entropy decoding unit inside the CTB decoding unit 202 (S504). Here, the entropy decoding unit inside the CTB decoding unit 202 is a CABAC capable of decoding the code string generated by the video encoding device 100, and is initialized at the start of a slice or tile.

Next, the code string separation unit 201 supplies the loop filter unit of the CTB to be decoded to the filter parameter decoding unit 204, and the filter parameter decoding unit 204 decodes the loop filter unit to the filter parameter (S505), This is supplied to the loop filter 203. Details of the filter parameter decoding unit 204 will be described later.
Next, the code string separation unit 201 supplies the CTB code string of the CTB to be decoded to the CTB decoding unit 202, and the CTB decoding unit 202 decodes the CTB code string of the CTB to be decoded and obtains the decoded image data of the CTB to be decoded. Acquired (S506), and supplies the decoded image data to the loop filter 203.

  Next, the loop filter 203 executes a loop filter on the decoded image data supplied from the CTB decoding unit 202 based on the filter parameter supplied from the code string separation unit 201 (S507), and a new decoded image is obtained. Data is generated and the decoded image data is supplied to the frame memory 205. Since the loop filter 203 has the same function as the loop filter 104 of the video encoding device 100, the description thereof is omitted.

(Function and operation of filter parameter decoding unit 204)
Next, functions and operations of the filter parameter decoding unit 204 will be described.
FIG. 20 is a flowchart for explaining the operation of the filter parameter decoding unit 204. Hereinafter, the function and operation of the filter parameter decoding unit 204 will be described with reference to FIG.

  First, the filter parameter decoding unit 204 derives available_left_ctb_flag and available_above_ctb_flag (S510). Here, the method for deriving available_left_ctb_flag and available_above_ctb_flag is the same as that in the process described with reference to FIG. 14 in which “encoding target CTB” is replaced with “decoding target CTB”.

  Next, it is checked whether available_left_ctb_flag is 1 (S511). If available_left_ctb_flag is 1 (Y in S511), loopfilter_merge_left_flag is decoded (S512). Next, it is inspected whether loopfilter_merge_left_flag is 1 (S513). If loopfilter_merge_left_flag is 1 (Y in S513), the CTB filter parameter located to the left of the decoding target CTB is used as the decoding target CTB filter parameter (S514), and the process is terminated.

  If available_left_ctb_flag is not 1 (N in S511) or loopfilter_merge_left_flag is not 1 (N in S513), it is checked whether available_above_ctb_flag is 1 (S515). If available_above_ctb_flag is 1 (Y in S515), loopfilter_merge_above_flag is decoded (S516). Next, it is checked whether loopfilter_merge_above_flag is 1 (S517). If loopfilter_merge_above_flag is 1 (Y in S517), the CTB filter parameter located above the decoding target CTB is used as the filtering parameter of the decoding target CTB (S518), and the process is terminated. If available_above_ctb_flag is not 1 (N in S515) or loopfilter_merge_above_flag is not 1 (N in S517), loopfilter_cod_param () is decoded (S519), and the process is terminated.

  As described above, the loop filter_interleave_flag is set to 1, the CTB filter parameters are separated and decoded in CTB units in the CTB code order, and the loop filter is executed in CTB units, thereby decoding all CTBs in the picture. Thus, it is possible to realize a low-delay moving image decoding apparatus that can output decoded image data immediately after the completion of. Note that the filter parameter decoding unit 204 holds the filter parameters of all the CTBs in the picture in order to refer to the filter parameters of the CTBs located on the left and the top of the decoding target CTB.

  Further, when the left side of the decoding target CTB is in contact with the tile boundary, if the filter parameter of the decoded CTB located to the left of the decoding target CTB is used as the filtering parameter of the decoding target CTB, the left side of the decoding target CTB is in contact with The filter parameter of the CTB to be decoded cannot be determined until the processing of the current CTB is completed. Therefore, when the left side of the decoding target CTB is in contact with the tile boundary, a loopfilter_merge_left_flag that is a flag indicating whether or not to use the filter parameter of the decoded CTB located to the left of the decoding target CTB as the filter parameter of the decoding target CTB. Is not decoded, the filter parameter of the decoded CTB located to the left of the CTB to be decoded is made unusable. In this way, even when the left side of the decoding target CTB is in contact with the tile boundary, it is possible to obtain a filter parameter immediately after decoding the loop filter unit of the decoding target CTB. Then, without waiting for the end of the CTB processing with which the left side of the decoding target CTB is in contact, the processing of the loop filter 203 is operated in units of CTB, and the output delay of the decoded image data after applying the loop filter is minimized. can do.

  Similarly, when the upper side of the decoding target CTB is in contact with the tile boundary, the flag indicates whether or not to use the decoded CTB filter parameter located above the decoding target CTB as the filter parameter of the decoding target CTB. By not decoding the loopfilter_merge_above_flag, the filter parameters of the decoded CTB located above the CTB to be decoded are made unusable. In this way, even when the upper side of the decoding target CTB is in contact with the tile boundary, it is possible to obtain the filter parameters immediately after decoding the loop filter unit of the decoding target CTB. Then, without waiting for the end of the CTB process with which the upper side of the decoding target CTB is in contact, the processing of the loop filter 203 is operated in units of CTBs, and the output delay of the decoded image data after applying the loop filter is minimized. can do.

[Second Embodiment]
Next, a second embodiment will be described. The second embodiment is different from the first embodiment in that loopfilter_interleave_flag is set to 0. That is, the filter parameter is encoded on a picture-by-picture basis.
Hereinafter, differences from the first embodiment will be described.

  First, the operation of the moving picture coding apparatus 100 according to the present embodiment will be described. FIG. 21 is a flowchart for explaining the operation of the moving picture coding apparatus 100 according to the second embodiment. As shown in FIG. 21, steps S107 to S110 are executed after all CTBs in the picture have been encoded. According to the order of the CTB addresses, the processes of step S107, step S108, and step S109 are repeatedly performed for all CTBs in the picture (S112 to S113).

  Next, for all CTBs of the picture, the code sequence multiplexing unit 107 converts the CTB code sequence supplied from the CTB encoding unit 102 and the loop filter unit supplied from the filter parameter encoding unit 105 to SPS, as necessary. The encoded stream is multiplexed together with the PPS, slice header, etc. (S110). Here, the configuration of the encoded stream multiplexed by the code string multiplexing unit 107 will be described. FIG. 22 is a diagram for explaining an example of the configuration of an encoded stream according to the second embodiment. FIG. 22 shows the structure of an encoded stream when loopfilter_interleave_flag is 0. The SPS is multiplexed at the head of the encoded stream. PPS is multiplexed at the head of a picture. At this time, picture loop filter parameters in which loop filter units are grouped in units of pictures are multiplexed in the PPS. A slice header is multiplexed at the head of the slice. Following the slice header, the CTB code sequence is multiplexed in the CTB code order for the CTBs belonging to the slice.

  Further, the setting operation of the entropy coding unit in step S105 is different. In the present embodiment, when the encoding target CTB is the first CTB of a slice or tile, the filter parameter encoding unit 105 does not initialize the entropy encoding unit inside the filter parameter encoding unit 105.

  Here, for filter parameter encoding, loopfilter_interleave_flag is set to 1 in the first embodiment. Therefore, when the left side of the encoding target CTB is in contact with the tile boundary, available_left_ctb_flag is set to 0. In the second embodiment, loopfilter_interleave_flag is set to 0, so that available_left_ctb_flag is set to 1 even when the left side of the encoding target CTB is in contact with the tile boundary. Therefore, even when the left side of the encoding target CTB is in contact with the tile boundary, it is possible to encode the loopfilter_merge_left_flag, and the encoding efficiency can be improved. Similarly, even when the upper side of the encoding target CTB is in contact with the tile boundary, the loopfilter_merge_left_flag can be encoded, and the encoding efficiency can be improved.

  Also, by setting loopfilter_interleave_flag to 0 and multiplexing the filter parameters in units of pictures, an encoding device that suppresses the code amount of the filter parameters can be realized.

  Subsequently, the operation of the moving picture decoding apparatus 200 according to the present embodiment will be described. FIG. 23 is a flowchart for explaining the operation of the moving picture decoding apparatus 200 according to the second embodiment. As shown in FIG. 23, step S505 and step S507 are executed after decoding of all CTBs in the picture. According to the order of the CTB addresses, the processes in steps S505 and S507 are repeated for all CTBs in the picture (S510 to S511).

  Here, step S505 in the present embodiment will be described. The code string separation unit 201 supplies the loop filter unit of the CTB to be decoded in the picture loop filter parameter to the filter parameter decoding unit 204, and the filter parameter decoding unit 204 decodes the loop filter unit to the filter parameter (S505), The filter parameter is supplied to the loop filter 203.

  Further, the operation of the loop filter 203 is different. In the present embodiment, since the loop filter is executed after decoding of all CTBs in the picture, the loop filter 203 holds the decoded image data of all CTBs in the picture.

  Also, the setting operation of the entropy decoding unit in step S504 is different. In the present embodiment, even if the decoding target CTB is the first CTB of a slice or tile, the filter parameter decoding unit 204 does not initialize the entropy decoding unit in the filter parameter decoding unit 204 according to the first embodiment. Different from form.

  As described above, the video decoding apparatus 200 according to the present embodiment can decode the encoded stream generated by the video encoding apparatus 100 according to the present embodiment and output decoded image data. .

[Third Embodiment]
Next, a third embodiment will be described. This embodiment is different from the first embodiment in the pixel to which the loop filter is applied. In the first embodiment, the loop filter is operated so as to complete with CTB, but in this embodiment, it is possible to execute the loop filter across CTB.
Hereinafter, differences from the first embodiment will be described.

  First, the syntax of PPS is different. FIG. 24 is a diagram for explaining the syntax of PPS according to the third embodiment. The difference from the first embodiment is that a loop filter_cross_tiles_flag is added to the PPS. loopfilter_cross_tiles_flag is a 1-bit flag indicating whether or not to permit processing of the loop filter across tile boundaries. If the loop filter_cross_tiles_flag is 1, the loop filter is allowed to process across the tile boundary, and if the loop filter_cross_tiles_flag is 0, the loop filter is not allowed to process across the tile boundary. In the present embodiment, it is assumed that loopfilter_cross_tiles_flag is 1.

  FIG. 25 is a flowchart for explaining the operation of the moving picture coding apparatus 100 according to the third embodiment. The operation of the moving picture encoding apparatus 100 according to the present embodiment, which is different from the first embodiment, will be described using FIG. Subsequent to step S102, the filter parameters of all CTBs of the picture are determined (S112). Here, the method for determining the filter parameters of all CTBs of a picture can be realized by, for example, a general two-pass encoding or a method of generating by referring to the data of the previous picture.

  Further, step S115 is performed instead of step S107. In step S115, the filter parameter of the CTB to be encoded is acquired for the filter parameter determined in step S112.

  Further, following step S111, a loop filter is executed for pixels straddling the tile boundary (S116). Pixels that straddle a tile boundary are pixels that are located on the left of the tile boundary and pixels that are located on the right of the tile boundary if the tile boundary is in the vertical direction, and above the tile boundary if the tile boundary is in the horizontal direction. And a pixel located below the tile boundary. Step S116 is a process necessary when performing independent processing (parallel processing) from step S103 to step S111 in units of tiles.

  As described above, the filter parameters of all CTBs in a picture are determined before the encoding process in CTB units, and the loop filter is executed after the encoding process in CTB units for all CTBs in the picture is completed. Thus, since the output of the encoded stream can be output by an encoding process in units of CTB, a low-delay moving image encoding apparatus can be realized.

  FIG. 26 is a flowchart for explaining the operation of the moving picture decoding apparatus 200 according to the third embodiment. The operation of the moving picture decoding apparatus 200 according to the present embodiment, which is different from the first embodiment, will be described using FIG. Step S507 is not performed after step S506. Subsequent to step S508, the processes of steps S512 and S507 are repeated for all CTBs in the picture in the order of CTB addresses (S510 to S511). Hereinafter, step S512 and step S507 will be described.

  If the loopfilter_merge_left_flag of the processing target CTB is 1, the filter parameter is acquired from the CTB located to the left of the processing target CTB. If the loop filter_merge_left_flag of the processing target CTB is 0 and the loop filter_merge_above_flag of the processing target CTB is 1, the filter parameter is acquired from the CTB located above the processing target CTB. If the loop filter_merge_left_flag of the processing target CTB is 0 and the loop filter_merge_above_flag of the processing target CTB is 0, the file parameter of the processing target CTB is acquired (S512).

  The loop filter 104 performs a loop filter on the decoded image data of the processing target CTB based on the filter parameter of the processing target CTB, generates new decoded image data, and stores the decoded image data in the frame memory 205. Supply (S507). Here, it is assumed that the loop filter 104 holds the decoded image data of all CTBs in the picture, but the decoded image data supplied from the CTB decoding unit 202 is supplied to the frame memory 205 to be stored, and step S507 is performed. Then, the decoded image data of the processing target CTB may be acquired from the frame memory 205.

  Next, the operation of the edge type filter 131 will be described. FIG. 27 is a flowchart for explaining the operation of the edge type filter 131 according to the third embodiment. Step S140 is different from that of the first embodiment. In step S140, if loopfilter_cross_tiles_flag is 1, the inspection result in step S134 on whether pixel A or pixel B is out of the tile is skipped, and the operations after step S135, which is a step of executing an edge type filter, are performed.

  In this way, even when loopfilter_interleave_flag is set to 1 and the transmission delay of the filter parameter is reduced, if loopfilter_cross_tiles_flag is 1, the edge type filter refers to pixel A and pixel B outside the tile, so It becomes possible to apply a loop filter to two pixels that are in contact with each other, and the distortion of the tile boundary can be reduced. In addition, since the image quality of the reference picture used in motion compensation prediction is improved, the coding efficiency can be improved.

  Next, derivation of available_left_ctb_flag and available_above_ctb_flag, which are the effectiveness of adjacent CTBs in the operations of the filter parameter encoding unit 105 and the filter parameter decoding unit 204, will be described. FIG. 28 is a flowchart for explaining the derivation of the effectiveness of adjacent CTBs in the third embodiment. The difference from the first embodiment is that step S190 and step S191 are added. Steps S190 and S191 check whether loopfilter_cross_tiles_flag is 1. Hereinafter, step S190 and step S191 will be described.

If loopfilter_cross_tiles_flag is 1 (Y in S190), step 182, step 183, and step 184 are skipped, and available_left_ctb_flag is set to 1 regardless of whether the left side of the encoding target CTB is in contact with the tile boundary or not. The If loopfilter_cross_tiles_flag is not 1 (N in S190), step S182 is executed.
If loopfilter_cross_tiles_flag is 1 (Y in S190), step 186, step 187, and step 188 are skipped, and available_above_ctb_flag is set to 1 regardless of whether the upper side of the encoding target CTB is in contact with the tile boundary or not. The If loopfilter_cross_tiles_flag is not 1 (N in S191), step S186 is executed.

  As described above, when loopfilter_cross_tiles_flag is set to 1, in encoding of filter parameters, loopfilter_merge_left_flag and flag of loopfilter_merge_above_flag, which is a flag indicating that the filter parameter is acquired across the tile boundary, and encoding of boundaries of tile are permitted. When the filter parameters are the same across the two, the coding efficiency can be improved by setting the loop filter_merge_left_flag and the loop filter_merge_above_flag to 1 and setting the code amount required for the filter parameter to 1 bit.

  Also, when loopfilter_cross_tiles_flag is set to 1, in decoding of a file unit, loopfilter_merge_left_flag, loopfilter_merge_above_flag, and loop of loopfilter_code_param () after decoding of CT before decoding and looping before CT of the loop before decoding For a CTB having a loopfilter_merge_left_flag of 1, a filter parameter is obtained from the CTB located to the left of the CTB. For a CTB having a loopfilter_merge_above_flag of 1, the filter parameter is obtained from the CTB located above the CTB. , Filter parameters Without stopping the decoding processing of data, the loop filter can be performed accurately after decoding all the CTB picture is finished.

  Here, by decoding the loopfilter_cod_param () having the most complicated structure in the syntax element of the file unit in advance, the delay for acquiring the filter parameter can be minimized.

  In order to suppress the execution delay of the loop filter, in step S505, the loop filter is executed for the CTB that does not use the filter parameter outside the tile boundary, and after all the CTB decoding processes are completed. A loop filter may be executed for a CTB that uses filter parameters outside the tile boundary.

[Fourth Embodiment]
Next, a fourth embodiment will be described. The fourth embodiment is different from the third embodiment in filtering at the tile boundary and control of filter parameter encoding.
Hereinafter, differences from the third embodiment will be described.

  First, the syntax of PPS is different. FIG. 29 is a diagram for explaining the syntax of the PPS according to the fourth embodiment. The difference from the third embodiment is that a loop filter_interleave_cross_tiles_flag is added to the PPS. loopfilter_interleave_cross_tiles_flag is a 1-bit flag indicating whether or not to allow acquisition of filter parameters across tile boundaries. If loopfilter_interleave_cross_tiles_flag is 1, it is permitted to acquire filter parameters across tile boundaries, and if loopfilter_interleave_cross_tiles_flag is 0, it is not permitted to acquire filter parameters across tile boundaries. In the present embodiment, it is assumed that loopfilter_interleave_cross_tiles_flag is 0.

  Next, derivation of available_left_ctb_flag and available_above_ctb_flag, which are the effectiveness of adjacent CTBs in the operations of the filter parameter encoding unit 105 and the filter parameter decoding unit 204, will be described. FIG. 30 is a flowchart for explaining the derivation of the effectiveness of adjacent CTBs in the fourth embodiment. Steps S190 and S191 are different from those of the third embodiment. Steps S190 and S191 check whether loopfilter_interleave_cross_tiles_flag is 1.

Next, the operation of the video decoding device 200 will be described. FIG. 37 is a flowchart for explaining the operation of the moving picture decoding apparatus 200 according to the fourth embodiment. The operation of the moving picture decoding apparatus 200 according to the present embodiment, which is different from the third embodiment, will be described with reference to FIG.
Subsequent to step S506, a loop filter is executed for pixels other than the pixels in contact with the tile boundary (S520). Further, following step S508, a loop filter is executed for pixels that are in contact with the tile boundary (S521).

  As described above, the loop filter_interleave_flag is set to 1, the loop filter_cross_tiles_flag is set to 1, and the loop filter_interleave_cross_tiles_flag is set to 0, so that the process related to the loop filter after the decoding process of all CTBs is completed is set to be a border filter. The loop filter execution delay can be minimized while reducing tile boundary distortion. In addition, the decoding pixels that need to be stored in order to execute the loop filter after the decoding process of all the CTBs in the picture is completed can be only the adjacent pixels A and B for the pixels that touch the tile boundary. The storage capacity of the decoded image data can be reduced. Further, by setting loopfilter_interleave_flag to 1, loopfilter_cross_tiles_flag to 1, and loopfilter_interleave_cross_tiles_flag to 1, the same effect as the third embodiment can be obtained.

[Fifth Embodiment]
Next, a fifth embodiment will be described. The fifth embodiment is different from the first embodiment in usable filter types. In the first embodiment, the filter types that can be used in the CTB located at the tile boundary and the other CTBs are the same. However, in the present embodiment, the CTB located at the tile boundary and other CTBs can be used. Different filter types.
Hereinafter, differences from the first embodiment will be described.

  FIG. 31 is a diagram for explaining filter types that can be used according to the CTB position according to the fifth embodiment. The filter types that can be used according to the CTB position will be described with reference to FIG. With respect to CTB in which available_left_ctb_flag and available_above_ctb_flag are both 0 and the left and upper sides of the CTB do not touch the picture boundary and tile boundary, use of all filter types from filter type 0 to filter type 5 is used as in the first embodiment. And assigns loopflitter_type_idx from 0 in ascending order of filter type. For the CTB whose left side is in contact with the picture boundary or tile boundary, only the filter type 1 and the filter type 2 are permitted to be used, and 0 and 1 are assigned to the filter type 1 and the filter type 2 as loopflitter_type_idx, respectively. For the CTB whose upper side is in contact with the picture boundary or tile boundary, use of only the filter type 1 and the filter type 3 is permitted, and 0 and 1 are assigned to the filter type 1 and the filter type 3 as loopflitter_type_idx, respectively.

  Next, the operation of the edge type filter 131 will be described. FIG. 32 is a flowchart for explaining the operation of the edge type filter 131 in the fifth embodiment. The difference is that step S133 for checking whether the pixel A or the pixel B is a picture boundary, step S134 for checking whether the pixel A or the pixel B is a tile boundary, and step S137 for setting the offset value to 0 are omitted.

  As described above, when loopfilter_interleave_flag is 1, for CTB in which the left side of CTB is in contact with the picture boundary or tile boundary, filter type 2 and filter type 4 are filter types that refer to pixels outside the picture boundary or tile boundary. By disabling the filter type 5, there is no need to refer to pixels outside the picture boundary or tile boundary, so that the step S 133, step S 134, and step S 137 of the edge type filter 131 are not required, and the edge The processing amount of the type filter 131 can be reduced. Similarly, for the CTB whose upper side is in contact with the picture boundary or tile boundary, the filter type 1, filter type 4, and filter type 5, which are filter types referring to pixels outside the picture boundary or tile boundary, are invalidated. Thus, since it is not necessary to refer to pixels outside the picture boundary or tile boundary, steps S133, S134, and S137 of the edge type filter 131 become unnecessary, and the processing amount of the edge type filter 131 is reduced. be able to.

  Also, by assigning loopflitter_type_idx only to valid filter types, for the CTB in which the left or upper side of the CTB is in contact with the picture boundary or tile boundary, the coding amount of the loopflitter_type_idx is reduced to a 1-bit flag, thereby improving the coding efficiency. Can be improved. Here, the filter type 0, that is, the filter type that does not process the loop filter, can be dealt with by setting the offset as 0 in the filter type 1 or the filter type 2.

  Further, in the moving picture coding apparatus according to the third embodiment or the fourth embodiment, by limiting the filter type at the tile boundary as in the present embodiment, the low delay which is always completed by CTB. It is possible to realize a moving image encoding device and a moving image decoding device.

[Sixth Embodiment]
Next, a sixth embodiment will be described. The sixth embodiment is different from the fourth embodiment in the derivation of the code string of loopflitter_type_idx and available_left_ctb_flag and available_above_ctb_flag.
Hereinafter, differences from the fourth embodiment will be described.

  First, the code strings of loopflitter_type_idx are different. FIG. 33 is a diagram for explaining a code string of loopflitter_type_idx according to the sixth embodiment. “0” of loopflitter_type_idx indicates the filter type 0, and the code string (bin) is “0”. 1 of loopflitter_type_idx indicates filter type 1 and the code string (bin) is “10”. Loopfilter_type_idx of 2 indicates filter type 2, and the code string (bin) is “1100”. “3” of loopflitter_type_idx indicates the filter type 3, and the code string (bin) is “1101”. 4 of loopflitter_type_idx indicates the filter type 4 and the code string (bin) is “1110”. The loopfitter_type_idx of 5 indicates the filter type 5 and the code string (bin) is “1111”.

  Here, since the filter parameter of the band type filter has a larger code amount per filter than the filter parameter of the edge type filter, by assigning a code amount shorter than the edge type filter loopfilter_type_idx to the loopfilter_type_idx of the band type filter, the edge type The code amount required for each of the filter and the band type filter can be balanced.

  By balancing the amount of code required for each of the edge type filter and the band type filter, the encoding time and the decoding time of the loop filter encoding parameters of the edge type filter and the band type filter can be balanced. Circuit design and software design can be facilitated. Here, the loop type filter loop_type_idx of the band type filter is set to 2 bits and the edge type filter loop filter_type_idx is set to 4 bits, but is not limited to this if the loop type filter loop filter_type_idx is shorter than the edge type filter loop filter_type_idx.

  Next, derivation of available_left_ctb_flag and available_above_ctb_flag is different. Next, derivation of available_left_ctb_flag and available_above_ctb_flag, which are the effectiveness of adjacent CTBs in the operations of the filter parameter encoding unit 105 and the filter parameter decoding unit 204, will be described. FIG. 34 is a flowchart for explaining the derivation of the effectiveness of adjacent CTBs in the sixth embodiment. The difference from the fourth embodiment is that step S196 and step S197 are added. Hereinafter, step S196 and step S197 will be described.

  It is checked whether the filter type of the filter parameter of the CTB located to the left of the encoding target CTB is filter type 0 (filter OFF) (S196). If the filter type of the filter parameter of the CTB located to the left of the encoding target CTB is the filter type 0 (Y in S196), available_left_ctb_flag is set to 0 (S184). If the filter type of the filter parameter of the CTB located to the left of the encoding target CTB is not the filter type 0 (N in S196), S182 is executed.

  It is checked whether or not the filter type of the filter parameter of the CTB located above the encoding target CTB is the filter type 0 (filter OFF) (S197). If the filter type of the filter parameter of the CTB located above the encoding target CTB is the filter type 0 (Y in S197), the available_above_ctb_flag is set to 0 (S188). If the filter type of the filter parameter of the CTB located above the encoding target CTB is not the filter type 0 (N in S197), S186 is executed.

  Here, when the filter type is filter type 0, loopfilter_type_idx is 1 bit, and therefore it is more likely that a smaller amount of code is required than when encoding using loopfilter_merge_left_flag or loopfilter_merge_above_flag. This means that if the filter type of the filter parameter of the CTB located on the left is filter type 0, one bit of loopfilter_merge_left_flag is sufficient, but the filter type of the filter parameter of the CTB located on the left is not filter type 0. If the filter type of the filter parameter of the CTB located is filter type 0, loopfilter_merge_left_flag and loopfilter_merge_above_flag are 2 bits, and the filter type of the CTB filter parameter located on the left is not filter type 0, but the CTB filter located above If the filter type of the parameter is not filter type 0, loopfilter_merge_left_flag Loopfilter_merge_above_flag, it is because the 3-bit Loopfliter_type_idx.

  Therefore, when the filter type is filter type 0, it is possible to improve the coding efficiency by coding only the loop filter coding parameter without coding the loopfilter_merge_left_flag and the loopfilter_merge_above_flag.

[Seventh Embodiment]
Next, a seventh embodiment will be described. The seventh embodiment differs from the sixth embodiment in the syntax of the loop filter unit and the code string of loopflitter_type_idx.
Hereinafter, differences from the sixth embodiment will be described.

  First, the syntax of the loop filter unit is different. FIG. 35 is a diagram for explaining the syntax of the loop filter unit according to the seventh embodiment. An enable_loopfilter_flag which is a loop filter permission flag is added. enable_loopfilter_flag is a flag for controlling ON / OFF of the loop filter of the CTB. A loop filter is executed for a CTB whose enable_loopfilter_flag is 1, and a loop filter is not executed for a CTB whose enable_loopfilter_flag is 0.

  Next, the code string of loopflitter_type_idx is different. FIG. 36 is a diagram for describing a code string of loopflitter_type_idx according to the seventh embodiment. “0” of loopflitter_type_idx indicates filter type 1, and the code string (bin) is “0”. 1 of loopflitter_type_idx indicates filter type 2, and the code string (bin) is “100”. 2 of loopflitter_type_idx indicates the filter type 3 and the code string (bin) is “101”. 3 of loopflitter_type_idx indicates the filter type 4 and the code string (bin) is “110”. 4 of loopflitter_type_idx indicates the filter type 5 and the code string (bin) is “111”.

  Next, derivation of available_left_ctb_flag and available_above_ctb_flag, which are the validity of adjacent CTBs in the operations of the filter parameter encoding unit 105 and the filter parameter decoding unit 204, is different. Step S196 and step S197 are different from the sixth embodiment, and step S196 and step S197 will be described.

  It is checked whether or not the loop filter permission flag of the filter parameter of CTB located to the left of the encoding target CTB is 0 (S196). If the loop filter permission flag of the filter parameter of the CTB located to the left of the encoding target CTB is 0 (Y in S196), the available_left_ctb_flag is set to 0 (S184). If the loop filter permission flag of the filter parameter of the CTB located to the left of the encoding target CTB is not 0 (N in S196), S182 is executed.

  It is checked whether the loop filter permission flag of the filter parameter of the CTB located above the encoding target CTB is 0 (S197). If the loop filter permission flag of the filter parameter of the CTB located above the encoding target CTB is 0 (Y in S197), the available_above_ctb_flag is set to 0 (S188). If the loop filter permission flag of the filter parameter of CTB located above the encoding target CTB is not 0 (N in S197), S186 is executed.

  In the video encoding device and video decoding device according to the embodiments described above, tile boundaries have been described as region boundaries. However, CTBs within a tile are processed in raster scan order, and slices are raster scanned with CTBs within the tile. When tiles and slices are configured so as to be included in order, a combination of tiles and slices can be applied similarly to slice boundaries by replacing the tile boundaries with slice boundaries.

  In the moving picture encoding apparatus and moving picture decoding apparatus according to the embodiments described above, the SAO filter is used as the loop filter. However, for example, a deblocking filter and a SAO filter can be combined.

  The moving image encoded stream output from the moving image encoding apparatus of the embodiment described above has a specific data format so that it can be decoded according to the encoding method used in the embodiment. Therefore, the moving picture decoding apparatus corresponding to the moving picture encoding apparatus can decode the encoded stream of this specific data format.

  When a wired or wireless network is used to exchange an encoded stream between a moving image encoding device and a moving image decoding device, the encoded stream is converted into a data format suitable for the transmission form of the communication path. It may be transmitted. In that case, a video transmission apparatus that converts the encoded stream output from the video encoding apparatus into encoded data in a data format suitable for the transmission form of the communication channel and transmits the encoded data to the network, and receives the encoded data from the network Then, a moving image receiving apparatus that restores the encoded stream and supplies the encoded stream to the moving image decoding apparatus is provided.

  The moving image transmitting apparatus is a memory that buffers the encoded stream output from the moving image encoding apparatus, a packet processing unit that packetizes the encoded stream, and transmission that transmits the packetized encoded data via the network. Part. The moving image receiving apparatus generates a coded stream by packetizing the received data, a receiving unit that receives the packetized coded data via a network, a memory that buffers the received coded data, and packet processing. And a packet processing unit provided to the video decoding device.

  The above processing relating to encoding and decoding can be realized as a transmission, storage, and reception device using hardware, and is stored in a ROM (Read Only Memory), a flash memory, or the like. It can also be realized by firmware or software such as a computer. The firmware program and software program can be recorded on a computer-readable recording medium, provided from a server through a wired or wireless network, or provided as a data broadcast of terrestrial or satellite digital broadcasting Is also possible.

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

  DESCRIPTION OF SYMBOLS 100 moving image encoder, 101 image data acquisition part, 102 CTB encoding part, 103 filter parameter determination part, 104 loop filter, 105 filter parameter encoding part, 106 frame memory, 107 code sequence multiplexing part, 110 encoding control Section, 111 tile setting section, 112 loop filter setting section, 113 area information setting section, 120 filter parameter setting section, 121 loop filter execution section, 122 error measurement section, 123 filter parameter determination section, 130 filter type determination section, 131 edge Type filter, 132 band type filter, 141 Edge angle setting unit, 142 Adjacent pixel setting unit, 143 Edge type determination unit, 144 Edge offset value setting unit, 145 Edge Offset value addition unit, 151 hand width setting unit, 152 band position setting unit, 153 band offset value setting unit, 154 band offset value addition unit, 200 video decoding device, 201 code string separation unit, 202 CTB decoding unit, 203 loop Filter, 204 filter parameter decoding unit, 205 frame memory, 210 decoding control unit.

Claims (6)

A moving image encoding apparatus that encodes in units of blocks constituting a picture,
An area setting unit for setting information of a divided area obtained by dividing the picture into a plurality of tiles having a size larger than the block;
A block encoding unit that encodes image data of a block to be encoded to generate a block code string, and obtains decoded image data of the block to be encoded;
A loop filter unit for filtering the decoded image data based on a filter parameter of the block to be encoded;
When encoding the acquisition flag indicating whether or not the filter parameter of the block to be encoded is acquired from the adjacent encoded block, and indicating that the acquisition flag is not acquired, A filter parameter encoding unit that encodes information to generate a filter code string;
The division region information is encoded into a picture parameter set indicating picture characteristics, the picture parameter set is multiplexed into a code stream, and the filter code string and the block code string are combined into the code string stream in units of blocks. A multiplexing unit for multiplexing,
The filter parameter encoding unit prior Symbol encoded subject to block checks whether contact with the tile boundary, when serving as the encoding target block is in contact with the tile boundary, that encoding the acquisition flag A video encoding apparatus characterized by the above. A video encoding method for encoding in units of blocks constituting a picture,
An area setting step for setting information of a divided area obtained by dividing the picture into a plurality of tiles having a size larger than the block;
A block encoding step of generating a block code string by encoding image data of a block to be encoded, and obtaining decoded image data of the block to be encoded;
A loop filter step for filtering the decoded image data based on a filter parameter of the block to be encoded;
When encoding the acquisition flag indicating whether or not the filter parameter of the block to be encoded is acquired from the adjacent encoded block, and indicating that the acquisition flag is not acquired, A filter parameter encoding step for encoding information to generate a filter code string;
The division region information is encoded into a picture parameter set indicating picture characteristics, the picture parameter set is multiplexed into a code stream, and the filter code string and the block code string are combined into the code string stream in units of blocks. A multiplexing step for multiplexing,
The filter parameter coding step, prior Symbol encoded subject to block checks whether contact with the tile boundary, when serving as the encoding target block is in contact with the tile boundary, that encoding the acquisition flag A video encoding method characterized by the above. A moving image encoding program for encoding in units of blocks constituting a picture,
An area setting step for setting information of a divided area obtained by dividing the picture into a plurality of tiles having a size larger than the block;
A block encoding step of generating a block code string by encoding image data of a block to be encoded, and obtaining decoded image data of the block to be encoded;
A loop filter step for filtering the decoded image data based on a filter parameter of the block to be encoded;
When encoding the acquisition flag indicating whether or not the filter parameter of the block to be encoded is acquired from the adjacent encoded block, and indicating that the acquisition flag is not acquired, A filter parameter encoding step for encoding information to generate a filter code string;
The division region information is encoded into a picture parameter set indicating picture characteristics, the picture parameter set is multiplexed into a code stream, and the filter code string and the block code string are combined into the code string stream in units of blocks. A computer to execute a multiplexing step for multiplexing;
The filter parameter coding step, prior Symbol encoded subject to block checks whether contact with the tile boundary, when serving as the encoding target block is in contact with the tile boundary, that encoding the acquisition flag A video encoding program characterized by the above. A packet processing unit that packetizes an encoded stream encoded by a moving image encoding method that encodes in units of blocks constituting a picture to obtain encoded data;
A transmission unit for transmitting the packetized encoded data,
The encoding method is:
An area setting step for setting information of a divided area obtained by dividing the picture into a plurality of tiles having a size larger than the block;
A block encoding step of generating a block code string by encoding image data of a block to be encoded, and obtaining decoded image data of the block to be encoded;
A loop filter step for filtering the decoded image data based on a filter parameter of the block to be encoded;
When encoding the acquisition flag indicating whether or not the filter parameter of the block to be encoded is acquired from the adjacent encoded block, and indicating that the acquisition flag is not acquired, A filter parameter encoding step for encoding information to generate a filter code string;
The division region information is encoded into a picture parameter set indicating picture characteristics, the picture parameter set is multiplexed into a code stream, and the filter code string and the block code string are combined into the code string stream in units of blocks. A multiplexing step for multiplexing,
The filter parameter coding step, prior Symbol encoded subject to block checks whether contact with the tile boundary, when serving as the encoding target block is in contact with the tile boundary, that encoding the acquisition flag A transmitter characterized by the above. A packet processing step of packetizing an encoded stream encoded by a moving image encoding method for encoding in units of blocks constituting a picture to obtain encoded data;
Transmitting the packetized encoded data, and
The encoding method is:
An area setting step for setting information of a divided area obtained by dividing the picture into a plurality of tiles having a size larger than the block;
A block encoding step of generating a block code string by encoding image data of a block to be encoded, and obtaining decoded image data of the block to be encoded;
A loop filter step for filtering the decoded image data based on a filter parameter of the block to be encoded;
When encoding the acquisition flag indicating whether or not the filter parameter of the block to be encoded is acquired from the adjacent encoded block, and indicating that the acquisition flag is not acquired, A filter parameter encoding step for encoding information to generate a filter code string;
The division region information is encoded into a picture parameter set indicating picture characteristics, the picture parameter set is multiplexed into a code stream, and the filter code string and the block code string are combined into the code string stream in units of blocks. A multiplexing step for multiplexing,
The filter parameter coding step, prior Symbol encoded subject to block checks whether contact with the tile boundary, when serving as the encoding target block is in contact with the tile boundary, that encoding the acquisition flag A transmission method characterized by the above. A packet processing step of packetizing an encoded stream encoded by a moving image encoding method for encoding in units of blocks constituting a picture to obtain encoded data;
Transmitting the packetized encoded data to the computer, and
The encoding method is:
An area setting step for setting information of a divided area obtained by dividing the picture into a plurality of tiles having a size larger than the block;
A block encoding step of generating a block code string by encoding image data of a block to be encoded, and obtaining decoded image data of the block to be encoded;
A loop filter step for filtering the decoded image data based on a filter parameter of the block to be encoded;
When encoding the acquisition flag indicating whether or not the filter parameter of the block to be encoded is acquired from the adjacent encoded block, and indicating that the acquisition flag is not acquired, A filter parameter encoding step for encoding information to generate a filter code string;
The division region information is encoded into a picture parameter set indicating picture characteristics, the picture parameter set is multiplexed into a code stream, and the filter code string and the block code string are combined into the code string stream in units of blocks. A multiplexing step for multiplexing,
The filter parameter coding step, prior Symbol encoded subject to block checks whether contact with the tile boundary, when serving as the encoding target block is in contact with the tile boundary, that encoding the acquisition flag A transmission program characterized by.

Images