JP6249040B2 - Video decoding method - Google Patents

Description

  The present invention relates to a moving picture decoding method for decoding an input stream indicating encoded data of each block obtained by dividing moving picture data.

  The moving image data generally has a very large amount of data. Therefore, a device that handles moving image data compresses the moving image data by encoding it when transmitting the moving image data to another device or when storing the moving image data in the storage device. .

  As a typical moving picture coding standard technology, MPEG (Moving Picture Experts Group phase) -2, MPEG-4, or MPEG-4 AVC / H established by ISO / IEC (International Standardization Organization / International Electrotechnical Commission) . H.264 (H.264 MPEG-4 Advanced Video Coding) is widely used.

  In such an encoding standard technique, only the information of the encoding target picture and the inter encoding method for encoding the encoding target picture using the information of the encoding target picture and the preceding and succeeding pictures are used. Intra-coding method is used.

  In general, the code amount of a picture or block encoded by the inter encoding method is smaller than the code amount of a picture or block encoded by the intra encoding method. In this way, the coding amount of the picture is biased in the sequence depending on the selected coding mode. Similarly, the coding amount selected causes a deviation in the code amount of the block in the picture.

  Therefore, even if the code amount fluctuates with time, a transmission buffer for the data stream is prepared in the transmission source device so that a data stream including a moving image encoded at a constant transmission rate can be transmitted. In addition, a reception buffer for the data stream is prepared in the transmission destination device.

  The delay due to these buffers (hereinafter referred to as buffer delay) is the main factor of the delay (hereinafter referred to as codec delay) from the input of each picture in the encoding device to the display of each decoded picture in the decoding device. . The codec delay includes a decoding delay that is a delay related to decoding and a display delay that is a delay related to display (output).

  By reducing the size of the buffer, buffer delay and codec delay are reduced. However, as the buffer size decreases, the degree of freedom of code amount distribution for each picture also decreases, and as a result, the quality of a moving image to be reproduced deteriorates. The degree of freedom of code amount distribution means the degree of code amount variation.

  MPEG-2 and MPEG-4 AVC / H. H.264 defines the operation of a reception buffer in an ideal decoding device called VBV (Video Buffering Verifier) and CPB (Coded Picture Buffer), respectively. Hereinafter, an ideal decoding device is referred to as an ideal decoding device.

  The moving image encoding device controls the amount of codes so that the reception buffer of the ideal decoding device does not overflow and underflow. The ideal decoding device is defined to perform instantaneous decoding in which the time required for the decoding process is zero. For example, there is a technique for controlling a moving image encoding apparatus related to VBV (see, for example, Patent Document 1).

  The moving picture coding apparatus ensures that the data of the picture is stored in the reception buffer at the time when the picture is decoded by the ideal decoding apparatus so that the reception buffer of the ideal decoding apparatus does not overflow and underflow. The code amount is controlled.

  The underflow of the reception buffer is caused when the video encoding device transmits a stream at a constant transmission rate, and the amount of code of each picture is large, and the video decoding device decodes the picture before the time to be decoded and displayed. This occurs when the transmission of data necessary for the transfer is not completed. That is, the underflow of the reception buffer means that data necessary for decoding a picture does not exist in the reception buffer of the decoding device. In this case, since the video decoding device cannot perform the decoding process, a frame skip occurs.

  Since the moving picture decoding apparatus performs decoding without causing an underflow of the reception buffer, the video is displayed after delaying the stream by a predetermined time from the reception time.

  As described above, in the ideal decoding device, it is defined that the decoding process is instantaneously completed at the processing time 0. Therefore, the input time of the i-th picture (hereinafter also referred to as P (i)) to the moving picture coding apparatus is t (i), and the decoding time of P (i) in the ideal decoding apparatus is dt (i). Then, the time at which the picture can be displayed is dt (i), similar to the decoding time.

  Since the picture display period {t (i + 1) -t (i)} is equal to {dt (i + 1) -dt (i)} in all the pictures, the decoding time dt (i) is equal to the input time t (i). Is a time {dt (i) = t (i) + dly} delayed by a fixed time dly. Therefore, the moving picture encoding apparatus must complete transmission of data necessary for decoding to the reception buffer of the moving picture decoding apparatus by time dt (i).

  FIG. 1 is a diagram illustrating a transition example of the buffer occupation amount of the reception buffer according to the conventional technique. In the example shown in FIG. 1, the horizontal axis represents time, and the vertical axis represents the buffer occupation amount of the reception buffer. A solid line graph 10 represents the buffer occupancy at each time.

  In the reception buffer, the buffer occupancy is restored at a predetermined transmission rate, and data used for decoding the picture is extracted from the buffer at the decoding time of each picture. In the example shown in FIG. 1, the input of P (i) data is started to the reception buffer from time at (i), and the last data of P (i) is input at time ft (i). The ideal decoding device completes the decoding of P (i) at time dt (i), and P (i) can be displayed at the time dt (i).

  While an ideal decoding device performs instantaneous decoding, an actual moving image decoding device requires a predetermined decoding processing time. In general, the decoding processing time for one picture is shorter than the picture display period, but is close to the picture display period.

  Data of P (i) is input to the reception buffer from time at (i) to ft (i), but data required for decoding each block arrives at any time within ft (i) from at (i) It is not guaranteed to do. Therefore, the actual moving picture decoding apparatus starts the decoding process of P (i) from time ft (i). Therefore, if the worst processing time required for decoding one picture is ct, the actual moving picture decoding apparatus can only guarantee that the decoding process is completed only at time ft (i) + ct.

  The video encoding apparatus guarantees that data necessary for decoding P (i) has arrived at the reception buffer by time dt (i), that is, ft (i) ≦ dt ( i). Therefore, when ft (i) is the slowest, ft (i) is equal to dt (i).

  At this time, the time at which the decoding process for the entire P (i) is guaranteed is dt (i) + ct. In order to display all the pictures so that the interval between pictures to be displayed is constant, the moving picture decoding apparatus must delay the display time of each picture by at least time ct from the ideal decoding apparatus.

  MPEG-2 VBV and MPEG-4 AVC / H. In H.264 CPB, the difference between the arrival time of each encoded picture in the video decoding device and the display time of each encoded picture is (ft (i) −at (i) + ct). That is, it is difficult to make the codec delay from each picture input to the encoding device to the corresponding picture output at the decoding device less than the time ct. That is, since the time ct is usually one picture processing time, it is difficult to achieve a codec delay less than one picture processing time.

JP 2003-179938 A

JCTVC-H1003, "High-Efficiency Video Coding (HEVC) text specification draft 6", Joint Collaborative Team on Video Coding of ITU-T SG16 WP3 and ISO / IEC JTC1 / SC29 / WG11, February 2012 MPEG-2 Test Model 5. April 1993.ISO-IEC / JTC1 / SC29 / WG11 / N0400 (http://www.mpeg.org/MPEG/MSSG/tm5/)

  In the prior art, it is difficult to set the codec delay to one picture processing time, but there are the following methods to make the codec delay less than one picture processing time. For example, this method assigns each block in a picture to one of N groups and assigns a decoding start time to each group. A group is, for example, one block line. A block line represents a block row in the horizontal direction of a picture.

  If the amount of generated information in each group can be made uniform, the difference between the decoding start times of successive groups matches the processing time per group, and the time ct becomes the processing time ct / N per group. As a result, the codec delay can be reduced to the processing time per group.

  FIG. 2 is a diagram illustrating an example in which the codec delay is less than one picture time by group division. A graph 17 shown in FIG. 2 represents a time transition of the buffer occupation amount of the conventional method. On the other hand, the graph 15 shown in FIG. 2 represents the time transition of the buffer occupation amount by group division.

  By the group division method, the decoding start time dgt (i, n) of the nth group of P (i) (hereinafter also referred to as G (i, n)) is defined, and the buffer occupation amount is reduced. Each group is decoded for a group decoding time ct / N indicated by reference numeral 16 from the corresponding decoding start time, thereby reducing the delay of the displayable time of each group.

  The group division method reduces the codec delay to the time per group by making the amount of information generated in each group substantially uniform. This codec delay is the worst value when the amount of information generated in each block in the group is extremely biased. In practice, however, the bias in the amount of information generated in each block in the group is reduced by appropriate rate control. can do. In this case, it is theoretically possible to further reduce the codec delay, but it is difficult to realize with the block division method. The reason for this will be described with reference to FIGS.

  FIG. 3 is a diagram illustrating a state of the reception buffer of the video decoding device. In the example shown in FIG. 3, it is expressed by using the accumulated amount of encoded data arrival to the reception buffer and the accumulated value of the encoded data consumed by the decoding process.

  A graph 20 shown in FIG. 3 represents the accumulated arrival value of the encoded data. The encoded data is transmitted from the video encoding device to the video decoding device at a constant rate R. In the example shown in FIG. 3, the arrival time of the first bit to the reception buffer of the video decoding device, that is, at (0) is set to 0.

  A graph 21 shown in FIG. 3 is an accumulated value of the encoded data consumed by the instantaneous decoding process in units of pictures. After the initial delay dly, the i-th picture P (i) (i = 0,...) Is sequentially decoded in sequence by dt (i). The difference dt (i + 1) −dt (i) between the instantaneous decoding times of two consecutive pictures is constant. The encoded information amount of P (i) is represented by b (i).

  at (i) and ft (i) represent the time at which the first bit and the last bit of the encoded data of P (i) arrive at the video decoding device, respectively. In order not to underflow the reception buffer of the moving picture decoding apparatus, all the encoded data of P (i) must arrive at dt (i). That is, it is necessary that dt (i) ≧ ft (i) and dt (i−1) ≧ at (i) are satisfied.

  The capacity of the reception buffer at each time corresponds to the difference between the graph 20 and the graph 21 at each time. For example, the capacity of the reception buffer after instantaneous decoding of P (0) at time dt (0) is the bit amount indicated by reference numeral 25.

  FIG. 4 is a diagram illustrating a state of the reception buffer focusing on one P (i). FIG. 4 is an enlarged view of a part of FIG. In particular, in the example shown in FIG. 4, when instantaneous decoding is performed in units of pictures, the reception buffer of the video decoding device does not underflow, and at (i) and ft (i) are the latest time, that is, dt ( i) = ft (i) and dt (i−1) = at (i). In the example shown in FIG. 4, it is assumed that the number of groups N is 4, and the number of blocks included in each group and the amount of generated information, dgt (i, n + 1) -dgt (i, n) are uniform.

  A graph 30 illustrated in FIG. 4 represents the accumulated arrival amount of the encoded data to the reception buffer of the video decoding device. A graph 31 is a cumulative value of consumed encoded data when instantaneous decoding is performed in units of pictures.

  A graph 32 is a cumulative value of consumed encoded data when instantaneous decoding of the nth group G (i, n) of P (i) is performed at dgt (i, n).

  The group division method is based on the premise that the amount of information generated in each group is averaged within a picture. In other words, the sum of the amount of generated information of each block in each group of P (i) is b (i) / N. b (i) is the generated information amount of P (i).

  The minimum value of the generated information amount of each block in the group P (i) is 0, and the maximum value is b (i) / N. When each block in P (i) is instantaneously decoded at equal intervals between dt (i−1) and dt (i), the cumulative value graph f (t) of the consumed encoded data is represented by reference numeral 35. It exists in the inside of the square area | region shown by -38.

  When the amount of information generated in each block is uniform, f (t) is a straight line (matching the graph 30) connecting the lower left vertex and the upper right vertex of the rectangular area indicated by reference numerals 35 to 38. When the bit amount of the entire group is generated in the first block, f (t) is a line connecting the left end line and the upper end line of each square. The latter is the worst case in terms of buffer delay.

  In the example shown in FIG. 4, the bits of each block of P (i) arrive during the time from dt (i−1) to dt (i). The arrival time g (x) of the xth bit (x = [1, b (i)]) is expressed by the following equation.

実際の動画像復号装置の挙動を鑑み、Ｐ（ｉ）の各ブロックをｄｔ（ｉ−１）からｄｔ（ｉ）までの間で等間隔にて瞬時復号した場合を考える。ピクチャ内のブロック総数をＭとした場合、Ｐ（ｉ）内のｍ番目のブロックの理想瞬時復号時刻ｐ（ｉ，ｍ）は、以下の式で表される。

Considering the actual behavior of the moving picture decoding apparatus, let us consider a case where each block of P (i) is instantaneously decoded at equal intervals between dt (i−1) and dt (i). When the total number of blocks in the picture is M, the ideal instantaneous decoding time p (i, m) of the mth block in P (i) is expressed by the following equation.

ｆ（ｔ）の形状によっては、ｆ（ｔ）がグラフ３０よりも上側になる。つまり、ｆ（ｐ（ｉ，ｍ））＜ｇ（ｆ（ｐ（ｉ，ｍ）））となり、ブロックの復号に必要となる全てのビットが動画像復号装置の受信バッファに届かず、アンダーフローが発生する。各ブロックのビット量が均一の場合には、ｆ（ｐ（ｉ，ｍ））＝ｇ（ｆ（ｐ（ｉ，ｍ）））となりアンダーフローは発生しないが、バッファ遅延の観点で最悪のケースである。

Depending on the shape of f (t), f (t) is on the upper side of the graph 30. That is, f (p (i, m)) <g (f (p (i, m))), and all the bits necessary for decoding the block do not reach the reception buffer of the moving picture decoding apparatus, resulting in underflow. Will occur. When the bit amount of each block is uniform, f (p (i, m)) = g (f (p (i, m))) and no underflow occurs, but the worst case from the viewpoint of buffer delay It is.

  When the bit amount of the entire group is generated in the first block, the arrival time of all the bits necessary for decoding the first block is delayed by dgt (i, n + 1) −dtg (i, n).

  In the group division method, the moving picture decoding apparatus cannot know the shape of f (t). Therefore, even if the bit arrival delay of the first block of G (i, n) is the worst value dgt (i, n) -dgt (i, n-1), it is guaranteed that underflow will be avoided. To do. Then, it is required to delay the instantaneous decoding time of all the blocks in G (i, n) to dgt (i, n). That is, the decoding start time of the leading block of P (i) is dgt (i, 1), and the prior art has a first problem that the codec delay cannot be further reduced.

  In the prior art, it is assumed that the image can be displayed instantaneously after decoding with the decoding time ct / N. However, Non-Patent Document 1 employs an encoding method called a tile that allows a picture to be divided not only horizontally but also vertically. For this reason, there is a case where the image cannot be displayed instantaneously after decoding with the decoding time ct / N. An example of a case where display is not possible instantaneously will be described with reference to FIG.

  FIG. 5 is a diagram illustrating an example of a case where display is not possible instantaneously. In Non-Patent Document 1, each area obtained by dividing a picture not only horizontally but also vertically is called a tile. In the example shown in FIG. 5, the picture is divided into four tiles.

  Tile 0 (t40), tile 1 (t41), tile 2 (t42), and tile 3 (t43) are arranged in the order of upper left, upper right, lower left, and lower right, and the tiles are processed in this order.

  In addition, within each tile, several groups with multiple blocks are included. In the example shown in FIG. 5, groups 0 to 3 are indicated by s41 to s44. At this time, the decoding is performed in the order of the groups, and the scanning order or the decoding order as indicated by sc41 to sc42 is obtained.

  In contrast to this decoding order, the display order may be a raster scan depending on the display, and at this time, the order is as indicated by reference sign sc43. At this time, there is a case where the group cannot be displayed instantaneously even after the decoding process is completed.

  For example, consider immediately after decoding of group 0 (s41). At this time, the CTB in the upper left half of the picture included in tile 0 (t40), for example, block b41 and block b42, can be displayed because it belongs to group 0 (s41). However, since the CTB included in the upper right half of the picture included in tile 1 (t41), for example, block b45 and block b46 belong to group 2 (s43), they are not decoded, and therefore cannot be displayed. Absent.

  When the display is a raster scan, it is structured to display in order from the left end to the right end of the screen. Therefore, when attempting to display the upper row of the picture, the blocks belonging to group 2 (s43) must also be displayed. Therefore, it is necessary to wait until group 2 (s43) is decoded and can be displayed.

  The time required to complete the decoding of group 2 (s43) is the time when all the blocks through which sc41 and sc42 pass in the scan order are decoded.

  In the group division method, decoding can be performed earlier, but nothing is considered regarding the displayable time. For this reason, the prior art has a second problem in that it is necessary to wait for a time corresponding to one picture in order to guarantee that a picture is displayed.

  Non-Patent Document 1 defines an operation when the amount of bits necessary for decoding a picture is larger than the amount of bits that can be accumulated in a buffer, such as when the picture is complex.

  FIG. 6 is a diagram for explaining the operation when the amount of bits necessary for decoding a picture is larger. The moving image encoding apparatus adjusts the code amount so that the accumulation of the rate R indicated by the rate 51 determined as in the graph 50 shown in FIG. 6 does not exceed the accumulation 52 of the extracted bit amount of the picture. Do.

  However, the amount of bits required for encoding is not sufficient for the amount of bits accumulated in the buffer, for example, the picture is complex, and underflow may occur. For example, it is a case like the graph 53 shown in FIG.

  When underflow occurs, the decoding device does not start decoding at the decoding time dt (0) of the original picture, as shown in the graph 54 in FIG. 6, and the time when the bits necessary for decoding are received by the buffer. Perform decryption on dt ′.

  In general, the display timing of the delayed picture is the timing dt (1) at which the next picture should be displayed. The picture that should originally be displayed at the time of dt (1) is decoded but skipped.

  However, in Non-Patent Document 1, there is a third problem that the rule when underflow occurs in group units is not clearly defined.

  Therefore, the technique disclosed below is a moving picture encoding apparatus, moving image, which can realize further lower delay when realizing a codec delay of less than one picture time in order to solve at least the second problem. An object is to provide an image encoding method, a moving image decoding apparatus, and a moving image decoding method.

  A moving image encoding method according to an aspect of the disclosure is a moving image decoding method for decoding an input stream indicating encoded data of each block obtained by dividing each picture included in the moving image data, wherein the encoding is performed from the input stream. Which group of blocks in which each block constituting the picture is divided into predetermined units from the input stream, the decoded block of each block is acquired and decoded in a frame memory Group information indicating whether it belongs to a group, and extracting group output delay information including a group output delay associated with a division shape when a picture is divided into predetermined units from the input stream, and the group information and the group output delay The output time of the first group of the picture is calculated based on the information and becomes the output time , Sequentially output the decoded data of each block of the group belonging to the decoded output target picture, during the decoding, detecting the occurrence of underflow in the group unit, at the output time of the first group of the picture, When underflow is detected in the group constituting the picture, the computer executes a process of outputting another decoded block stored in the frame memory instead of the picture.

  According to the disclosed technique, when a codec delay of less than one picture time is realized, a further reduction in delay can be realized.

The figure which shows the example of a transition of the buffer occupation amount of the reception buffer by a prior art. The figure which shows the example which made the codec delay less than 1 picture time by group division. The figure which shows the mode of the receiving buffer of a moving image decoding apparatus. The figure which shows the mode of the reception buffer which paid its attention to one P (i). The figure which shows an example when it cannot display instantaneously. The figure for demonstrating operation | movement when the bit amount required for decoding of a picture is larger. 1 is a block diagram illustrating an example of a schematic configuration of a moving image encoding device according to Embodiment 1. FIG. The figure showing the accumulation value of the coding data at the time of paying attention to P (i). The figure for demonstrating display delay. The figure which shows the relationship between the accumulation value of the bit amount of the encoding data which arrived at the receiving buffer, and the accumulation value of the generated information amount in each block in P (i). The figure for demonstrating calculation of group output time information. 5 is a flowchart illustrating an example of a moving image encoding process according to the first embodiment. 6 is a flowchart illustrating an example of output processing according to the first embodiment. FIG. 9 is a block diagram illustrating an example of a schematic configuration of a moving image decoding apparatus according to a second embodiment. 10 is a flowchart illustrating an example of a moving image decoding process according to the second embodiment. 10 is a flowchart illustrating an example of output processing according to the second embodiment. FIG. 10 is a block diagram illustrating an example of a schematic configuration of a moving image encoding device according to a third embodiment. The figure for demonstrating generation | occurrence | production of underflow. The figure for demonstrating the process at the time of underflow generation | occurrence | production. 9 is a flowchart illustrating an example of processing of a moving image encoding device according to Embodiment 3. FIG. 10 is a block diagram illustrating an example of a schematic configuration of a moving image decoding apparatus according to a fourth embodiment. 10 is a flowchart illustrating an example of processing of a moving image decoding apparatus according to a fourth embodiment. FIG. 10 is a block diagram illustrating an example of a schematic configuration of a moving image processing apparatus according to a fifth embodiment.

  Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. The moving image encoding apparatus described in the embodiment encodes each picture included in moving image data in units of groups, and outputs a bit stream as encoded data.

  The picture may be either a frame or a field. The frame is one still image in the moving image data, while the field is a still image obtained by extracting only odd-numbered data or even-numbered data from the frame.

  The moving image to be encoded may be a color moving image or a monochrome moving image.

[Example 1]
<Configuration>
FIG. 7 is a block diagram illustrating an example of a schematic configuration of the video encoding device 100 according to the first embodiment. The moving image encoding apparatus 100 includes an encoding processing unit 110, a code amount control unit 120, a group determination unit 130, a decoding time determination unit 140, and an output time determination unit 150.

  The encoding processing unit 110 includes an orthogonal transform unit 111, a quantization unit 112, and an entropy encoding unit 113.

  The code amount control unit 120 includes a quantization value calculation unit 121, a buffer occupation amount calculation unit 122, and a bit counter 123.

  The code amount control unit 120 sets the code amount so that the data necessary for the output of all the blocks included in the group reaches the decoding buffer of the decoding device by the output time indicated by the output delay time and the decoding is completed. Control.

  The group determination unit 130 includes a group configuration determination unit 131 and a group information addition unit 132.

  The decoding time determination unit 140 includes a group decoding time calculation unit 141, a group decoding delay determination unit 142, and a group decoding delay information addition unit 153.

  The output time determination unit 150 includes a group output time calculation unit 151, a group output delay determination unit 152, and a group output delay information addition unit 152.

  Each of these units included in the video encoding device 100 is mounted on the video encoding device 100 as a separate circuit. Alternatively, these units included in the video encoding device 100 may be mounted on the video encoding device 100 as a single integrated circuit in which circuits for realizing the functions of the units are integrated. Alternatively, these units included in the video encoding device 100 may be functional modules realized by a computer program executed on a processor included in the video encoding device 100.

  The encoding target picture included in the moving image data is divided into blocks by a control unit (not shown), and is input to the orthogonal transform unit 111 for each block. Each block has, for example, 16 × 16 pixels.

  The orthogonal transform unit 111 calculates an intra prediction value or an inter prediction value from a locally decoded picture stored in a frame memory (not shown), calculates a difference from the input block, and calculates a block prediction error. Further, orthogonal transformation is performed on the block prediction error.

  The quantization unit 112 performs a quantization operation on the orthogonally transformed block prediction error. A quantization parameter (control information) in the quantization operation is given from the quantization value calculation unit 121. The resulting quantized orthogonal transform coefficients and intra prediction or inter prediction parameters (intra prediction direction, motion vector information, etc.) are output to the entropy encoding unit 113 as block compressed data. The quantized orthogonal transform coefficient is subjected to inverse quantization and inverse orthogonal transform by a local decoding unit (not shown), and then a local decoding block is generated by adding an intra prediction value or an inter prediction value, and then a frame memory (not shown) To accumulate.

  The entropy encoding unit 113 performs entropy encoding on the block compressed data output from the quantization unit 112.

  The quantization value calculation unit 121 calculates the quantization value of each block from the reception buffer state of the ideal decoding device and the generated information amount upper limit of the block to be encoded next, which is output from the buffer occupation amount calculation unit 122. .

  The buffer occupancy amount calculation unit 122 outputs the accumulated bit amount value of the encoded data output from the bit counter 123, the group information output from the group configuration determination unit 131, and the group output from the group decoding delay determination unit 142. From the decoding time and the decoding delay of the group, the state of the reception buffer of the ideal decoding device and the upper limit of the generated information amount of the block to be encoded next are calculated.

  The bit counter 123 counts the number of output bits of the entropy encoding unit 113 and outputs a cumulative value of encoded data.

  The group configuration determining unit 131 determines a group to which each block belongs for a plurality of blocks. For example, the group configuration determination unit 131 determines a group to which a block being encoded belongs, according to a predetermined method using block count information received from a control unit (not shown) and encoding method designation information received from a control unit (not shown). To do.

  The block count information is information indicating the number of each block included in the picture. For example, the number for the upper left block of the picture is set to 1, and a number is assigned to each block according to the raster scan order. The largest number is assigned to the block at the lower right corner of the picture. The block count information may include a number assigned to each block according to another order.

  The group configuration determination unit 131 preferably determines a plurality of groups so that the number of blocks included in each group is as equal as possible in order to equalize the decoding processing time for each group.

  For example, if the group configuration determining unit 131 divides each block into groups in units of block lines, the number of blocks included in each group can be made equal for an arbitrary picture size.

  For example, if the picture size is 1920 pixels × 1088 pixels equivalent to high definition television broadcasting (HDTV), and the block size is 16 pixels × 16 pixels, the number of block lines is 68. Therefore, in this case, each block included in the encoding target picture is classified into one of 68 groups.

  The number of blocks included in the group may be set to a value between 1 and the number of blocks in the entire screen.

  The group configuration determination unit 131 notifies the buffer occupation amount calculation unit 122 of the identification information of the group to which the encoding target block belongs. The group configuration determination unit 131 notifies the group decoding time calculation unit 141 and the group output time calculation unit 151 of information on blocks included in each group. The group configuration determination unit 131 may notify the group decoding time calculation unit 141 and the group output time calculation unit 151 of the index of the block located at the head of each group.

  The group information adding unit 132 adds group information indicating the number of groups in the picture and block information in each group to the encoded data.

  The group decoding time calculation unit 141 calculates the decoding time of each group from the group information output from the group configuration determination unit 131 and notifies the group decoding delay determination unit 142 of it.

  The group decoding delay determination unit 142 determines the decoding delay of each group, and notifies the buffer occupation amount calculation unit 122 and the group decoding delay information addition unit 143 together with the decoding time of each group. The determined decoding delay is notified as delay information.

  Group decoding delay information adding section 143 receives the decoding time and decoding delay of the group and adds them to the encoded data as group decoding delay information.

  The group output time calculation unit 151 calculates the output time (also referred to as display time) of each group from the encoding method designation information received from the control unit (not shown) and the group information output from the group configuration determination unit 131. The output time information is notified to the group output delay determination unit 152.

  The group output delay determining unit 152 determines the output delay of each group from the output time of each group, and notifies the group output delay information adding unit 153 of the output delay information.

  The group output delay information adding unit 153 receives the output time and output delay of each group, and adds the output delay time from the decoding time for each group to the encoded data as group output delay information.

<< Decoding delay >>
Consider a case where each block in the i-th picture P (i) is instantaneously decoded at equal intervals from dt (i−1) to dt (i). In this case, the accumulated value graph f (t) of the consumed encoded data can reduce the block transmission delay by appropriate rate control such as setting the lower limit / upper limit of the information amount per block. In addition, by notifying this information to the video decoding device, it is possible to further advance the earliest decoding start time of the block. This will be described with reference to FIG.

  FIG. 8 is a diagram illustrating a cumulative value of encoded data when attention is paid to P (i). The graph 60 shows the accumulated arrival amount of the encoded data whose rate is R. A graph 61 is a cumulative value of consumed encoded data when instantaneous decoding is performed in units of pictures.

  Reference numerals 62 to 66 are cumulative values of encoded data consumed for decoding in the groups (G0 to G4) represented by reference numerals 67 to 71.

  Looking at the relationship between the existence range of each group and the graph 60, in G (1) to G (4), the rate is always above the accumulated value of the encoded data, so G (1) to G (4 Underflow does not occur even when instantaneous decoding of each block of) is performed at equal intervals between dt (i−1) and dgt (i, 1).

  In G (0), since the accumulated value of the encoded data of G (0) exceeds the rate, underflow occurs. In order to avoid this underflow, it is sufficient that the accumulated value of the encoded data does not exceed the rate, and the minimum value is the interval Δt.

  Δt is smaller than dgt (i, n) −dgt (i, n−1) regardless of which group is generated. The moving picture decoding apparatus uses the maximum value of Δt of each group in P (i) and sets the decoding start time of the first block of P (i) to dt (i−1) + Δt (i). , Instantaneous decoding can be performed at equal intervals without underflowing all blocks.

  In the entire sequence, from the maximum value Δt of Δt (i) of all pictures, the start block decoding start time dinit of the first picture is made as shown in the following formula without underflowing all blocks of all pictures: Instantaneous decoding can be performed at equal intervals.

またＰ（ｉ）内でｎ番目のグループの復号開始が可能となる時刻の最早値ｒ（ｉ，ｎ）は、以下の式で表される。

Further, the earliest value r (i, n) at which decoding of the nth group can be started within P (i) is expressed by the following equation.

動画像符号化装置にて、Δｔがｄｇｔ（ｉ，ｎ）−ｄｇｔ（ｉ，ｎ−１）より小さい値になるように各ピクチャ、各グループの発生情報量を制御し、Δｔの値を明示的に動画像復号装置に伝送する。動画像復号装置にて、グループＧ（ｉ，ｎ）の瞬時復号時刻をｒ（ｉ， ｎ）とすることで、確実に各ブロックの復号開始時刻を保証することができる。

In the moving picture coding apparatus, the amount of generated information for each picture and each group is controlled so that Δt becomes a value smaller than dgt (i, n) −dgt (i, n−1), and the value of Δt is clearly indicated. To the video decoding device. By setting the instantaneous decoding time of the group G (i, n) to r (i, n) in the video decoding device, the decoding start time of each block can be reliably guaranteed.

  The group on the video decoding device side does not necessarily match the group notified from the video encoding device. When the group on the video decoding device side matches the group notified from the video encoding device, r (i, n) = dgt (i, n).

≪Display delay≫
By explicitly setting the display delay of the target group as additional extension information, it is possible to notify the decoding device of the earliest display timing and to minimize the display delay. For example, a display delay designation method in the case of tile division and group division as shown in FIG. 5 will be described with reference to FIGS.

  In the case shown in FIG. 5, the display with the largest display delay is the uppermost display of group 0 (s41). In order to start the display of the uppermost row of group 0 (s41), it is necessary that at least the decoding of the pixel values of the uppermost row of pictures of group 2 (s43) has been completed. Therefore, the display delay is explicitly notified as additional extension information.

FIG. 9 is a diagram for explaining display delay. The time when the uppermost display of the group 0 (s41) becomes possible is ogt (0) shown in FIG. ogt (0) is set later than the group 2 extraction time dgt (2). The display time at this time is expressed by the following expression, assuming that the picture is decoded at a certain speed.
ogt (0) = dgt (0) + (dgt (2) −dgt (1)) + 1 / L (dgt (3) −dgt (2)) (5)
L represents the total number of lines in the vertical direction in the group 2 represented by s43, and l represents the number of lines at the upper right end of the picture in the group 2 represented by s43. l / L (dgt (3) −dgt (2)) is the time when decoding of the upper right end of the picture of group 2 represented by s43 is completed, assuming that it takes one group time to decode the group. Represent.

  That is, the displayable time is the time from the instantaneous decoding time of group 0 represented by s41 to the instantaneous decoding time of group 2 represented by s43 with respect to the decoding time dgt (0) of group 0 represented by s41. Add. Further, the displayable time is obtained by adding the time actually taken until the decoding of the upper right corner of the group 2 picture is completed.

  On the video encoding device side, it is possible to notify the decoding device side of an appropriate time considering the actual decoding time by explicitly transmitting the time represented by the above formula as additional extension information. Can guarantee a display with little delay.

  In the above example, the time at which the decoding of the upper right end of the picture of the group 2 in the display time is completed is the time dgt (3) −dgt at which the decoding of all the groups 2 represented by s43 is actually completed. Also in (2), since the time when decoding of one picture is completed can be notified earlier than the displayable time, display with a small delay can be guaranteed.

<< Decoding time calculation >>
A method for calculating group decoding time information in the first embodiment will be described. In the following explanation, M is the total number of blocks included in the current picture.

  First, the group decoding time calculation unit 141 decodes a decoding time dt (i) of P (i) delayed by a predetermined delay time dly from the input time t (i) of the i-th picture P (i) in the encoding order. Based on {= t (i) + dly}, a decoding time dgt (i, n) representing a time at which the nth group G (i, n) of the picture P (i) is decoded is calculated. Alternatively, the group decoding time calculation unit 141 uses {dgt (i, n) −dgt (i, n−1)} equivalent to dgt (i, n) instead of dgt (i, n) as the decoding time. May be calculated. Further, the group decoding time calculation unit 141 may round the decoding time so as to be a multiple of an appropriate unit, for example, 1/90000 second.

  The group decoding time calculation unit 141 sets the decoding time of each group to equal the time required for decoding processing of each block included in each group by the number of groups N, etc. Decide to split. In this case, the decoding time dgt (i, n) of G (i, n) (n = 1, 2,..., N) is calculated according to the following equation.

なお、ｄｔ（ｉ）はＰ（ｉ）の復号時刻である。ｄ（ｉ＋１）−ｄ（ｉ）はｉによらず一定であり、以降ｓと表現する。

Note that dt (i) is the decoding time of P (i). d (i + 1) -d (i) is constant regardless of i, and is hereinafter expressed as s.

  Furthermore, the group decoding time calculation unit 141 may determine the decoding time dgt (i, n) (n ≧ 2) of the group encoded / decoded after the second as the following equation.

さらに、グループ復号時刻決定部１４１は、２番目以降に符号化・復号されるグループの復号時刻ｄｇｔ（ｉ，ｎ）（ｎ≧２）を、次式のように決定してもよい。

Further, the group decoding time determination unit 141 may determine the decoding time dgt (i, n) (n ≧ 2) of the group encoded / decoded after the second as the following equation.

グループ復号遅延決定部１４２は、ピクチャ全体のブロック遅延最大値Δｔを符号化開始前に決定する。Δｔは次式で表される範囲の値にする。

The group decoding delay determination unit 142 determines the block delay maximum value Δt of the entire picture before the start of encoding. Δt is set to a value in the range represented by the following equation.

バッファ占有量算出部１２２は、理想復号装置の受信バッファのバッファ占有量、及び次に符号化するブロックの発生情報量上限を、以下のように計算する。

The buffer occupancy calculation unit 122 calculates the buffer occupancy of the reception buffer of the ideal decoding device and the generated information amount upper limit of the block to be encoded next as follows.

  FIG. 10 shows the accumulated value of the bit amount of the encoded data arriving at the reception buffer of the ideal decoding device and the accumulated value of the generated information amount in each block in P (i) in the encoding process of P (i). It is a figure which shows the relationship.

  A graph 72 represents the cumulative value R (t) of the bit amount of the encoded data that has arrived at the reception buffer of the ideal decoding device. The graph 75 is obtained by shifting the graph 72 to the left by Δt, and is R ′ (t). There is a relationship of R ′ (t) = R (t + Δt).

  B (i) shown in FIG. 10 represents the accumulated value of the generated encoded data from P (0) to P (i). b (i) represents the generated information amount of the entire P (i) and is the same as B (i) -B (i-1).

  The graph 73 shows a straight line V where the value at time dt (i−1) is B (i−1), the value at time dt (i) is B (i), and the slope is b (i) / s. (T). s is one picture time, which is the same as dt (i) -dt (i-1).

  The graph 73 shows that each block is decoded at equal intervals from time dt (i−1) to dt (i), and the generated information amount is equal to b (i) / M. This corresponds to the consumption curve f (t) of the encoded data.

  A graph 74 is an actual consumption data f (t) of the block-unit encoded data, and a point 77 is a cumulative value of the encoded data consumption of the block unit when decoding up to the m-th block.

  In order for the ideal decoding apparatus not to cause an underflow of the reception buffer when the group n is decoded at the group decoding start early time r (i, n) calculated from the group decoding time information, the following conditions must be satisfied: There is. The quantized value calculation unit 121 calculates the quantized value so that the following condition is always satisfied.

領域７６は、ｆ（ｔ）が、時刻ｄｔｇ（ｉ，ｕ−１）からｄｔｇ（ｉ，ｕ）までの間で取ることが許される範囲を示す。

A region 76 indicates a range in which f (t) is allowed to be taken between time dtg (i, u-1) and dtg (i, u).

≪Quantization value calculation≫
The quantization value calculation method of the block m in the quantization value calculation part 121 is demonstrated below. In the first embodiment, the number of blocks included in each group is equal to M / N.

  In starting the processing of the first block of the nth group G (i, n) to which the block m belongs, the target information amount T (i, n) of G (i, n) is calculated according to the following equation. Here, n = Ceil (m * N / M).

Ｔ（ｉ）はＰ（ｉ）全体の目標情報量であり、Ｔ'（ｉ，ｎ）はＧ（ｉ，ｎ）の実発生情報量である。Ｔ（ｉ）は、公知の方式を用いて、Ｐ（０）からＰ（ｉ−１）までの実発生情報量の総和によって決定される。

T (i) is the target information amount of the entire P (i), and T ′ (i, n) is the actual generated information amount of G (i, n). T (i) is determined by the sum total of the actual amount of information generated from P (0) to P (i-1) using a known method.

  For example, according to the quantization value calculation method in the standardization organization reference software Test Model 5 (see Non-Patent Document 2) in MPEG-2, the quantization value calculation unit 121 determines that the actual generated information amount of G (i, n) is T (i , N), the quantized value is calculated so as to approach.

  Next, the quantized value calculation unit 121 calculates the expected value b ′ (i, n) of the generated information amount accumulated value in P (i) when the encoding process for the entire G (i, n) is completed, and n The difference d1 from the accumulated amount of generated information B (i, n−1) in P (i) before entropy coding of the th group is compared with a predetermined threshold value DTH1.

  b ′ (i, n) is calculated by the following equation.

閾値ＤＴＨ１は、次式で表される。

The threshold value DTH1 is expressed by the following equation.

ｂ０は量子化値をその取り得る値の範囲のうちの最大値とした場合に、各ブロックで発生する最大の符号量である。（（Ｍ／Ｎ）−ｍ）は、Ｇ（ｉ，ｍ）で符号化処理が終わっていないブロック数に相当する。ｏｆｆｓｅｔはマージン項である。

b0 is the maximum code amount generated in each block when the quantized value is the maximum value in the range of possible values. ((M / N) -m) corresponds to the number of blocks in G (i, m) that have not been encoded. offset is a margin term.

  When d1 <DTH1, the quantized value calculation unit 121 sets the quantized value as the maximum value.

  For b0, the code amount of the block when the frequency coefficients are all 0 may be used. When d1 <DTH1, the quantization value calculation unit 121 determines the quantization value so that all the frequency coefficients of the encoding target block are quantized to zero. With this control, if the average value of the code amount of the remaining blocks that have not been subjected to the encoding process in the group does not exceed b0, T (i, n) ≧ T ′ (i, n), that is, f (dtg (i , N)) ≦ V (dtg (i, n)). And it is guaranteed that the reception buffer of the ideal decoding device will not underflow.

  As described above, if the quantization value calculation unit 121 actually transmits the output stream from the video encoding device 100 to the video decoding device according to the predetermined rate R, the reception buffer of the video decoding device does not underflow. In addition, the code amount of the moving image data can be controlled.

  The quantization value calculation unit 121 notifies the quantization unit 112 of the obtained quantization value.

  As described above, by performing buffer control for each group and adding a group coding delay, it is possible to appropriately reduce the buffer delay without causing overflow or underflow.

≪Output time calculation≫
Next, a method for calculating group output time information in the first embodiment will be described. FIG. 11 is a diagram for explaining calculation of group output time information.

  In the following explanation, M is the total number of blocks included in the current picture. Also, let the width and height of the picture, the width and height of the tile, and the width and height of the CTB be (widthp, heightp), (widthth, heightt), and (widthc, heightc), respectively. Here, it is assumed that all the tiles (t80 to t83) have the same size, and the tiles are processed in the raster scan order sc83. That is, in the example shown in FIG. 11, the processing is performed in the order of tile 0 (t80), tile 1 (t81), tile 2 (t82), and tile 3 (t83).

  Furthermore, in the example shown in FIG. 11, it is assumed that a group has 17 CTBs, and the number of CTBs is the same in all groups. At this time, the group 0 (s81) is from the CTB index 0 in the picture to the fourth row in the third column.

  In this way, the uppermost CTB column of the upper right tile 1 (t81) is included in the group 2 (s83). Therefore, when the display is displayed in raster scan order, at least group 0 (s81) can be displayed only after group 2 (s83) is decoded.

  When the group 2 (s83) is displayed after being decoded, first, assuming that the decoding is instantaneous decoding and the extraction timing of the group k is d (k), the output time ogt (0) of the group 0 (s81) ) Is expressed as:

また、復号に１ピクチャ時間ｓかかり、ピクチャ内のグループの数をＮとする場合、グループの復号に必要な時間は、ｓ／Ｎとなる。つまり、瞬時復号時の復号時刻ｄｇｔを利用すると、グループ２の復号が完了する時刻ｄｇｔ'（２）とグループ０（ｓ８１）の表示時刻ｏｇｔ（０）は次式のように表される。

In addition, when decoding takes one picture time s and the number of groups in a picture is N, the time required for decoding a group is s / N. That is, when the decoding time dgt at the time of instantaneous decoding is used, the time dgt ′ (2) when the decoding of the group 2 is completed and the display time ogt (0) of the group 0 (s81) are expressed as follows.

ここで、動画像符号化装置１００は、前復号ピクチャの復号時刻より、グループの出力時刻を減算した出力遅延時間を復号装置に通知することで、復号装置側において、表示時刻を保証することが出来る。

Here, the moving picture coding apparatus 100 can guarantee the display time on the decoding apparatus side by notifying the decoding apparatus of the output delay time obtained by subtracting the group output time from the decoding time of the previous decoded picture. I can do it.

  In addition, even in a post filter such as a deblocking filter in HEVC disclosed in Non-Patent Document 1, there is a case in which it is necessary to wait for a group to be decoded later in order to display the group. To do. Even in such a case, it is possible to achieve a display delay of less than one picture time by appropriately setting a display delay in consideration of a decoding time of a group to be decoded later.

<< Output stream >>
In order to share the group to which each block belongs, the group decoding delay, and the group output delay with the moving picture decoding apparatus, the moving picture encoding apparatus 100 includes at least group information that represents a block that belongs to each group, and group decoding delay information. , And group output delay information is added to the output data stream and notified to the video decoding device. The output data stream is also simply called an output stream.

  Therefore, the group decoding delay information adding unit 143 adds, for example, group decoding delay to the header information of the output data stream at each picture or at a predetermined picture interval.

  Further, the group output delay information adding unit 153 adds, for example, a group output delay to the header information of the output data stream at each picture or at a predetermined picture interval.

  Further, the group information adding unit 132 adds, for example, group information to the header information of the output data stream at each picture or at a predetermined picture interval.

  The header information is, for example, a sequence header defined in MPEG-2, or H.264. H.264 can be used as a sequence parameter set or supplemental enhancement information. The decoding time for each group is a picture header (Picture Header) defined in MPEG-2 or H.264. It may be added to header information that is always attached to each picture, such as a slice header defined in H.264.

  When the groups are determined so that the number of blocks included in each group is equal, the moving picture coding apparatus 100 notifies the moving picture decoding apparatus that all the blocks are equally divided into N groups. For this purpose, the group number N is notified from the group configuration determining unit 131 to the group information adding unit 132 as group information.

  The group information adding unit 132 encodes the group information. MPEG-2 and H.264 In H.264, encoding is performed in units of 16 × 16 pixel blocks called macroblocks, and the number of blocks does not exceed the range that can be expressed normally by 20 bits. Since the maximum value of the number N of groups is at most equal to the maximum value of the number of blocks, encoding of N may be performed with a fixed bit length.

  If the number of blocks included in each group is not necessarily equal, the group configuration determination unit 131 notifies the group information addition unit 132 of the group number N and the index information of the first block of each group as group information. Is done.

  First, the group information adding unit 132 encodes the number N of groups, and sequentially encodes index information of the first block of each group. As the encoding method for the index information of the head block, for example, a fixed bit length encoding method is used. Further, the group information adding unit 132 may use another encoding method such as a variable length encoding method such as a Huffman code in order to encode the number N of groups and the index information of the first block of each group.

<Operation>
Next, the operation of the moving image encoding apparatus 100 according to the first embodiment will be described. FIG. 12 is a flowchart illustrating an example of a moving image encoding process according to the first embodiment.

  In step S100, the group decoding delay Δt is first determined at the start of the sequence encoding operation. Δt is determined to be smaller than the time of the group in which the number of blocks included in the sequence is minimum.

  In step S101, the group decoding delay information adding unit 143 adds group information and group decoding time delay information to the data stream.

  In step S102, the group configuration determination unit 131 first determines a group in a picture when starting encoding of each picture. The number of groups in each picture in the sequence and the number of blocks included in each group can be determined for each picture. Alternatively, the same number of groups may be used for all pictures in the sequence, and the number of blocks included in each group may be equalized.

  In step S103, the group decoding delay determining unit 142 calculates a group decoding delay for each group (step S103).

  In step S104, the buffer occupation amount calculation unit 122 estimates the buffer state of the reception buffer of the ideal decoding device and the generated information amount upper limit of the group to be encoded next when starting decoding of each group.

  In step S105, the quantized value calculation unit 121 stores all data in the group in the reception buffer by the earliest start time of group decoding based on the buffer state of the reception buffer and the upper limit of the generated information amount of the group to be encoded next. Compute the quantized value of the block to arrive.

  In step S106, the encoding processing unit 110 encodes the block using the calculated quantization value.

  Next, output processing of the moving image encoding apparatus 100 according to the first embodiment will be described. FIG. 13 is a flowchart illustrating an example of output processing according to the first embodiment.

  In step S200, the output time determination unit 150 extracts group information from the data stream.

  In step S201, the group output delay determination unit 152 determines group output delay information. The group output delay time can be determined as described above.

  In step S202, the group output delay information adding unit 153 adds group output delay information to the data stream.

  As described above, according to the first embodiment, when realizing a codec delay of less than one picture time, further delay reduction can be realized by advancing group decoding or output.

[Example 2]
Next, a moving picture decoding apparatus according to the second embodiment will be described. In the second embodiment, the stream encoded by the moving image encoding apparatus 100 in the first embodiment is appropriately decoded.

<Configuration>
FIG. 14 is a block diagram illustrating an example of a schematic configuration of the video decoding device 200 according to the second embodiment. The moving picture decoding apparatus 200 includes a reception buffer 205, a block decoding unit 210, a frame memory 211, a group output unit 212, a decoding time calculation unit 220, an output time calculation unit 230, and a group information extraction unit 240. Have.

  The group information extraction unit 240 extracts group information indicating a group obtained by dividing each block at a predetermined interval from an input stream (also referred to as an input stream).

  The decoding time calculation unit 220 includes a group decoding delay information extraction unit 221 and a group decoding time calculation unit 222.

  The output time calculation unit 230 includes a group output delay information extraction unit 231 and a group output time calculation unit 232.

  Each of these units included in the video decoding device 200 is mounted on the video decoding device 200 as a separate circuit. Alternatively, these units included in the video decoding device 200 may be mounted on the video decoding device 200 as a single integrated circuit in which circuits for realizing the functions of the respective units are integrated. Alternatively, these units included in the video decoding device 200 may be functional modules realized by a computer program executed on a processor included in the video decoding device 200.

  The reception buffer 205 receives and buffers the stream transmitted from the moving image encoding apparatus 100.

  The block decoding unit 210 acquires data from the reception buffer 205 at the group decoding start time output from the group decoding time calculation unit 222, performs decoding processing sequentially from the first block, and sequentially outputs the decoded blocks. The decoding start time is also simply referred to as decoding time.

  The frame memory 211 stores the decoded block output from the block decoding unit 210. The frame memory 211 also functions as a decoding buffer that is buffered before the output target group is output, for example. The decoding buffer may have a configuration different from that of the frame memory 211.

  The group output unit 212 outputs the corresponding group at the group output time output from the group output time calculation unit 232.

  The group decoding delay information extraction unit 221 extracts group decoding delay information from the input stream that is encoded data.

  The group decoding time calculation unit 222 calculates the decoding start time of each group from the group information output from the group information extraction unit 240 and the group decoding delay information output from the group decoding delay information extraction unit 221.

  The group decoding time calculation unit 222 calculates, for example, the decoding start time dtb (i) of the first block of the i-th picture P (i) by the following equation.

グループ出力遅延情報抽出部２３１は、符号化データである入力ストリームから、グループ出力遅延情報を抽出する。

The group output delay information extraction unit 231 extracts group output delay information from the input stream that is encoded data.

  The group output time calculation unit 232 calculates the output time of each group from the group information output from the group information extraction unit 240 and the group output delay information output from the group output delay information extraction unit 231.

  The moving picture decoding apparatus 200 calculates the decoding start time of each decoding group based on the notified number N of groups and the decoding delay information of the groups. Further, the output time of each decoding group is calculated based on the notified number N of groups and the group output delay information.

<Operation>
Next, the operation of the video decoding device 200 in the second embodiment will be described. FIG. 15 is a flowchart illustrating an example of a moving image decoding process according to the second embodiment. In step S300 illustrated in FIG. 15, the group information extraction unit 240 first extracts group information from the data stream when starting decoding of each picture.

  In step S301, the group decoding delay information extraction unit 221 extracts group decoding delay information from the data stream.

  In step S302, the group decoding time calculation unit 222 calculates the decoding start time of the first group.

  The number of decoding groups in each picture in the sequence and the number of blocks included in each decoding group can be determined for each picture. Alternatively, the same number of decoding groups may be used for all pictures in the sequence, and the number of blocks included in each decoding group may be equalized. Furthermore, the decoding group may be the same as the group described in the block decoding time information.

  In step S303, the block decoding unit 210 waits until the group decoding time in the group decoding loop.

  In step S304, the block decoding unit 210 acquires data from the reception buffer 205 and decodes each block.

  In step S305, the group decoding time calculation unit 222 calculates the decoding start time of the next group.

  In step S306, the block decoding unit 210 outputs the decoded block to the frame memory.

  Next, output processing of the moving image decoding apparatus 200 according to the second embodiment will be described. FIG. 16 is a flowchart illustrating an example of output processing according to the second embodiment.

  In step S400, the group information extraction unit 240 first extracts group information from the data stream when starting decoding of each picture.

  In step S401, the group output delay information extraction unit 231 next extracts group output delay information from the data stream.

  In step S402, the group output time calculation unit 232 calculates the first group of P (i) from the group information output from the group information extraction unit 240 and the group output delay information output from the group output delay information extraction unit 231. Output start time is calculated. In step S403, the decoding block belonging to the group is output according to the output start time of the first group of P (i).

  As described above, according to the second embodiment, it is possible to appropriately decode the stream encoded by the moving image encoding apparatus 100 according to the first embodiment.

[Example 3]
Next, a moving picture coding apparatus according to the third embodiment will be described. In the third embodiment, what kind of processing is performed when underflow occurs in units of groups is defined.

<Configuration>
FIG. 17 is a block diagram illustrating an example of a schematic configuration of a video encoding device 300 according to the third embodiment. The moving image encoding apparatus 301 includes an encoding processing unit 310, a code amount control unit 320, a group determination unit 330, a decoding time determination unit 340, and an output time determination unit 350.

  The encoding processing unit 310, the group determining unit 330, the decoding time determining unit 340, and the output time determining unit 350 are the encoding processing unit 110, the group determining unit 130, the decoding time determining unit 140, and the output time determining unit 150 shown in FIG. The same processing is performed.

  The code amount control unit 320 includes a quantization value calculation unit 321, a buffer occupation amount calculation unit 322, a bit counter 323, and a filler addition unit 324.

  When the data necessary for decoding all the blocks included in the group is transmitted to the decoding device at a predetermined transmission rate, the code amount control unit 320 performs decoding until the time indicated by the determined display time. The code amount is controlled so as to reach the reception buffer.

  The quantization value calculation unit 321 and the bit counter 323 perform the same processing as the quantization value calculation unit 121 and the bit counter 123 illustrated in FIG.

  In addition to the operation of the buffer occupancy calculation unit 122 shown in FIG. 7, the buffer occupancy calculation unit 322 has a group generated information amount that exceeds the target value, and all data in the group is received by the reception buffer of the ideal decoding device by the decoding start time Check if a buffer underflow condition occurs.

  When a buffer underflow state is detected, the filler adding unit 324 is instructed to insert dummy data at the end of the processed picture, and a buffer underflow state is notified to an overall control unit (not shown). When notified of the buffer underflow state, the overall control unit (not shown) controls to skip the encoding process for the next picture to be encoded.

  The filler adding unit 324 inserts dummy data at the end of the processed picture. The amount of dummy data to be inserted is instructed from the buffer occupation amount calculation unit 322.

  The filler adding unit 324 adds filler data to the output stream when the data necessary for decoding all the blocks included in the group does not reach the reception buffer of the decoding device by the display time. Also, the filler adding unit 324 controls the data necessary for decoding the last block of the picture including the group not to reach the reception buffer of the decoding device by the display time by adding filler data.

<Processing at underflow>
First, the case where underflow occurs in a group in a picture will be considered with reference to FIG. FIG. 18 is a diagram for explaining the occurrence of underflow. As shown in the graph 90 of FIG. 18, basically, even when the decoding time in units of groups is defined, the encoding device side uses the decoding time scheduled according to the information sent to the decoding device with additional information such as an SEI message. The code amount is adjusted so that it can be decoded.

  However, as in the graph 91 shown in FIG. 18, when a group underflow occurs in dgt (0), the decoding is not performed until the bits necessary for the decoding are received in the buffer, as described above. is there.

  It should be noted that the display of one picture needs to be guaranteed and the display must be delayed by one picture even if a group underflow occurs. This is because when a group underflow occurs, it waits until a bit necessary for decoding of one group is received in the buffer. The next decoding timing is dgt ′ shown in the graph 91 shown in FIG.

  In this case, since the subsequent decoding time is also delayed by that amount, even when the picture to which the group belongs is decoded and the displayed time dt (0) is reached, the decoding of all the groups is not completed. Delay one picture display.

  Here, a case is considered in which underflow has not occurred as a picture even though underflow has occurred as a group. Since underflow has occurred in the group unit, group decoding is delayed, one picture display is delayed, and an attempt is made to skip the next picture.

  However, since there is no underflow in units of pictures, a contradictory situation occurs in which a picture is displayed at a normal time. In this case, since the decoding of the group is delayed, the decoding is not completed at the normal picture timing, and an appropriate picture cannot be output.

  Furthermore, even at the display timing of the next picture, the decoding necessary for the next picture is not finished, and an appropriate picture cannot be output. As described above, a picture appropriate for the display timing of the picture is not decoded.

  Therefore, as shown in FIG. 19, when an underflow occurs in a group, it is assumed that an underflow has occurred as a picture, and one picture display is delayed, and the next picture to be displayed is skipped. Then stipulate. Thereby, since the next picture can be skipped, the display time can be corrected correctly.

  FIG. 19 is a diagram for explaining processing when underflow occurs. For example, in the example shown in FIG. 19, when an underflow occurs in dgt (1), the amount of pictures to be decoded in dt (1) is larger than the encoded stream arrival rate 96 as indicated by reference numeral 95. It is assumed that underflow has occurred in dt (1) despite being small. As a result, the display of one picture is delayed, the picture displayed at dt (1) is displayed at dt (2), and the picture to be displayed at dt (2) is skipped.

  On the encoding device side, if underflow occurs in a group, filler data is packed into encoded data of a picture. As a result, it is possible to skip the next picture by intentionally causing underflow, so that the display time can be corrected correctly.

≪Underflow detection, stream editing≫
An underflow detection method and a bitstream editing method in the moving image encoding apparatus according to the third embodiment will be described below.

  First, it is assumed that the code amount control unit 320 performs the same operation as in the first embodiment. Underflow is detected by the buffer occupation amount calculation unit 322. At this time, if at least one group does not satisfy the condition (2), it is detected that an underflow has occurred in the group included in the picture.

  At this time, the buffer occupation amount calculation unit 322 notifies the filler addition unit 324 of underflow occurrence information. The filler adding unit 324 receives the underflow occurrence information and performs processing such that the picture display is skipped when it is confirmed that the underflow has occurred.

  For example, by adding Filler data to the output stream, an underflow in units of pictures is intentionally generated, and display of pictures is skipped. Since the method for adding Filler data can be easily inferred, description thereof is omitted here.

  By performing the above processing, it is possible to prevent the picture display order from being shifted by skipping the picture display.

<Operation>
Next, the operation of the moving picture coding apparatus 300 according to the third embodiment will be described. FIG. 20 is a flowchart illustrating an example of processing of the moving image encoding device according to the third embodiment.

  In step S500, the buffer occupancy calculation unit 322 confirms whether or not underflow occurs in units of groups based on the buffer occupancy of the reception buffer of the decoding device.

  In step S501, when the buffer occupation amount calculation unit 322 determines that underflow occurs in units of groups, the buffer occupancy amount calculation unit 322 instructs the filler addition unit 324 to edit the output stream so that underflow occurs in units of pictures. The filler adding unit 324 edits the output stream so that the picture display is skipped. A picture that has underflowed is also called a big picture.

As described above, according to the third embodiment, it is possible to appropriately process even when an underflow occurs in units of groups.
[Example 4]
Next, the moving picture decoding apparatus according to the fourth embodiment will be described. In the fourth embodiment, the encoded data encoded by the moving picture encoding apparatus according to the third embodiment can be appropriately decoded.

<Configuration>
FIG. 21 is a block diagram illustrating an example of a schematic configuration of the video decoding device 400 according to the fourth embodiment. The video decoding device 400 includes a reception buffer 405, a group decoding delay information extraction unit 421, a group output delay information extraction unit 431, a group decoding time calculation unit 422, a group output time calculation unit and 432, and group information extraction. Unit 440, block decoding unit 410, frame memory 411, group output unit 412, and display control unit 413.

  Each of these units included in the video decoding device 400 is implemented in the video decoding device 400 as a separate circuit. Alternatively, these units included in the video decoding device 400 may be mounted on the video decoding device 400 as one integrated circuit in which circuits that realize the functions of the units are integrated. Alternatively, these units included in the video decoding device 400 may be functional modules realized by a computer program executed on a processor included in the video decoding device 400.

≪Underflow detection, stream editing≫
An underflow detection method and a bitstream editing method in the video decoding device 400 according to the fourth embodiment will be described.

  First, it is assumed that the block decoding unit 410 performs the same operation as in the second embodiment. Underflow is detected by the block decoding unit 410. The block decoding unit 410 receives bit amount information from an entropy decoding unit (not shown).

  At this time, if at least one group does not satisfy the condition (2), it is detected that an underflow has occurred in the group included in the picture. For example, the graph 91 shown in FIG. 18 indicates that underflow has occurred at dgt (1).

  At this time, the block decoding unit 410 notifies the display control unit 413 of underflow occurrence information. The display control unit 413 receives the underflow occurrence information, confirms that an underflow has occurred, and performs processing to skip the picture display.

  That is, when an underflow occurs in the group dgt (l) of the picture whose decoding time is dt (k), even if a bit amount that can be decoded as a picture is stored in the buffer in dt (k), dt (K) is displayed in dt (k + 1). Also, the picture that was supposed to be displayed at dt (k + 1) is skipped.

  For example, in the example shown in FIG. 19, the picture that should have been displayed at dt (1) is displayed at dt (2), and the picture that should have been displayed at dt (2) is skipped. In this example, it is assumed that decoding is instantaneous, and that output (display) can be performed simultaneously with decoding.

  Accordingly, it is possible to prevent the picture display order from being shifted by skipping the picture display.

<Operation>
Next, the operation of the video decoding device 400 according to the fourth embodiment will be described. FIG. 22 is a flowchart illustrating an example of processing of the video decoding device 400 according to the fourth embodiment.

  In step S600, the block decoding unit 410 confirms whether or not an underflow occurs in units of groups based on the buffer occupation amount of the reception buffer 405.

  If the block decoding unit 410 determines in step S601 that underflow has occurred in units of groups, the block decoding unit 410 notifies the display control unit 413 of underflow occurrence information. When the underflow occurrence information is notified, the display control unit 413 corrects the display picture timing to be correct.

  As described above, according to the fourth embodiment, it is possible to appropriately decode the encoded data encoded by the moving image encoding apparatus 300 according to the third embodiment.

[Example 5]
FIG. 23 is a block diagram illustrating an example of a schematic configuration of the moving image processing apparatus 500. The moving image processing apparatus 500 is an example of the moving image encoding apparatus or the moving image decoding apparatus described in each embodiment. As illustrated in FIG. 23, the moving image processing apparatus 500 includes a control unit 501, a main storage unit 502, an auxiliary storage unit 503, a drive device 504, a network I / F unit 506, an input unit 507, and a display unit 508. These components are connected to each other via a bus so as to be able to transmit and receive data.

  The control unit 501 is a CPU that controls each device, calculates data, and processes in a computer. The control unit 501 is an arithmetic device that executes a program stored in the main storage unit 502 or the auxiliary storage unit 503. The control unit 501 receives data from the input unit 507 or the storage device, calculates and processes the data, and then displays the display unit 508. Or output to a storage device.

  The main storage unit 502 is a ROM (Read Only Memory), a RAM (Random Access Memory), or the like, and a storage device that stores or temporarily stores programs and data such as an OS and application software that are basic software executed by the control unit 501. It is.

  The auxiliary storage unit 503 is an HDD (Hard Disk Drive) or the like, and is a storage device that stores data related to application software or the like.

  The drive device 504 reads the program from the recording medium 505, for example, a flexible disk, and installs it in the storage device.

  The recording medium 505 stores a predetermined program. The program stored in the recording medium 505 is installed in the moving image processing apparatus 500 via the drive apparatus 504. The installed predetermined program can be executed by the moving image processing apparatus 500.

  The network I / F unit 506 has a communication function connected via a network such as a LAN (Local Area Network) or a WAN (Wide Area Network) constructed by a data transmission path such as a wired and / or wireless line. This is an interface between the device and the moving image processing apparatus 500.

  The input unit 507 includes a keyboard having cursor keys, numeric input, various function keys, and the like, a mouse, a slide pad, and the like for selecting keys on the display screen of the display unit 508. The input unit 507 is a user interface for a user to give an operation instruction to the control unit 501 or input data.

  The display unit 508 has an LCD (Liquid Crystal Display) or the like, and performs display according to display data input from the control unit 501. Note that the display unit 508 may be provided outside, and in that case, the moving image processing apparatus 500 includes a display control unit.

  As described above, the moving image encoding process or the moving image decoding apparatus described in each of the above-described embodiments may be realized as a program for causing a computer to execute. By installing this program from a server or the like and causing it to be executed by a computer, the above-described moving image encoding process or moving image decoding apparatus can be realized.

  It is also possible to record the moving picture encoding program or the moving picture decoding program on the recording medium 505 and cause the computer or portable terminal to read the recording medium 505 on which the program is recorded to realize the above-described processing. is there.

  The recording medium 505 is a recording medium that records information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, or a magneto-optical disk, and information is electrically stored such as a ROM or flash memory. Various types of recording media such as a semiconductor memory for recording can be used. Note that the recording medium 505 does not include a carrier wave.

  The program executed by the moving image processing apparatus 500 has a module configuration including each unit described in each embodiment. As actual hardware, when the control unit 501 reads out and executes a program from the auxiliary storage unit 503, one or more of the above-described units are loaded onto the main storage unit 502, and one or more of the units are main. It is generated on the storage unit 502.

  Further, the moving image encoding processing described in each of the above embodiments may be implemented in one or a plurality of integrated circuits.

  The moving picture encoding apparatus according to each of the embodiments described above is used for various purposes. For example, the moving image encoding device or moving image decoding device is incorporated in a video camera, a video transmission device, a video reception device, a videophone system, a computer, or a mobile phone.

  All examples and specific terms listed herein are intended for instructional purposes to help the reader understand the concepts contributed by the inventor to the present invention and the promotion of the technology. It should be construed that it is not limited to the construction of any example herein, such specific examples and conditions, with respect to showing the superiority and inferiority of the present invention. Although embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions and modifications can be made thereto without departing from the spirit and scope of the present invention.

  Although each embodiment has been described in detail above, it is not limited to a specific embodiment, and various modifications and changes can be made within the scope described in the claims. It is also possible to combine all or a plurality of the components of the above-described embodiments.

In addition, the following additional notes are disclosed regarding each of the above embodiments.
(Appendix 1)
A moving image encoding apparatus that divides and encodes each picture included in moving image data into a plurality of blocks,
A group configuration determining unit that determines a group to which each block belongs, for a plurality of blocks;
A group information adding unit for adding group information to which each block belongs to an output stream;
A group output time calculator for decoding and calculating an output time for each group;
A group output delay determining unit for determining an output delay for each group;
A group output delay information adding unit for adding an output delay time from the decoding time for each group to an output stream;
A code amount control unit that controls the code amount so that the data necessary for the output of all the blocks included in the group reaches the decoding buffer of the decoding device by the output time indicated by the output delay time and the decoding is completed When,
An encoding processing unit that performs encoding based on control information of the code amount control means;
Video encoding device (Appendix 2)
A moving image coding method for dividing and coding each picture included in moving image data into a plurality of blocks,
For multiple blocks, determine the group to which each block belongs,
Adding group information to which each block belongs to an output stream;
Calculate the output time for each group,
Determine the output delay for each group,
An output delay time from the decoding time for each group is added to the output stream,
Data necessary for output of all the blocks included in the group reaches the decoding buffer of the decoding device by the output time indicated by the output delay time, and controls the code amount so that decoding is completed,
An image encoding method in which a computer executes a process of performing encoding based on the controlled code amount.
(Appendix 3)
A moving picture decoding apparatus for decoding input data indicating encoded data of each block obtained by dividing each picture included in moving picture data,
A group information extraction unit that extracts group information indicating a group obtained by dividing each block at a predetermined interval from the input stream;
An output delay information extraction unit for extracting group output delay information including the output delay of the group from the input stream;
An output time calculation unit for calculating a group output time based on the group information and the group output delay information;
When the output time is reached, an output unit that sequentially outputs the decoded output target groups;
A video decoding device comprising:
(Appendix 4)
A moving picture decoding method for decoding input data indicating encoded data of each block obtained by dividing each picture included in moving picture data,
A video decoding device for decoding data,
Extracting group information indicating a group obtained by dividing each block at a predetermined interval from the input stream,
Extracting group output delay information including group output delay from the input stream;
Calculating a group output time based on the group information and the group output delay information;
A moving picture decoding method in which a computer executes a process of sequentially outputting decoded output target groups at the output time.

100, 300 Video encoding device 200, 400 Video decoding device 500 Video processing device 110 Coding processing unit 120 Code amount control unit 130 Group determination unit 140 Decoding time determination unit 142 Group decoding delay determination unit 143 Group decoding delay information Adder 150 Output time determiner 152 Group output delay determiner 153 Group output delay information adder 210 Block decoder 212 Group output unit 221 Group decode delay information extractor 222 Group decode time calculator 231 Group output delay information extractor 232 Group Output time calculation unit 324 Filler addition unit 413 Display control unit

Claims (1)

A moving picture decoding method for decoding an input stream indicating encoded data of each block obtained by dividing each picture included in moving picture data,
Obtain and decode data of each block encoded from the input stream,
Storing the decoded block of each block in a frame memory;
Extracting from the input stream group information indicating which group of blocks each of which constitutes a picture belongs to a group obtained by dividing a picture into predetermined units,
Extracting group output delay information including the output delay of the group associated with the division shape when the picture is divided into predetermined units from the input stream;
Calculating the output time of the first group of the picture based on the group information and the group output delay information;
When the output time is reached, the decoded data of each block of the group belonging to the decoded picture to be output is output in order,
At the time of decoding, occurrence of underflow is detected for each group, and if underflow is detected in a group constituting the picture at the output time of the first group of the picture, A video decoding method in which a computer executes a process of outputting another decoding block stored in the frame memory.

Images