JP5938935B2 - Moving picture coding apparatus and moving picture coding method - Google Patents

Description

  The present invention relates to a moving picture coding apparatus and a moving picture coding method that perform motion vector detection in parallel.

  H. Video coding such as H.264 / AVC (Advanced Video Coding) (hereinafter also referred to as H.264) is a hybrid video coding method using motion compensation and orthogonal transformation / quantization processing. For the input image to be encoded, a similar image block (hereinafter also referred to as a reference block) is calculated from a reference image (hereinafter also referred to as a reference picture) by motion vector detection.

  H. In H.264 encoding, orthogonal transform / quantization processing is performed on a difference image between an input image and a reference block, and the obtained coefficient data is entropy-coded and output. Information on which reference picture used to identify the reference block, information on which block size the motion vector is calculated, and information on the motion vector are separately entropy-coded and output.

  H. The H.264 encoding technique includes P_Skip as a mode for encoding an input image. In this P_Skip mode, instead of encoding information on which reference picture was used, information on the block size from which the motion vector was calculated, and information on the motion vector, information indicating that the macroblock was encoded with P_Skip. Encode.

  The code indicating P_Skip has a smaller code amount than the information indicating which reference picture is selected and the information of the motion vector. For this reason, since the compression rate of moving images increases, it is desirable that many macroblocks of this type appear.

  Receiving that the macroblock type is P_Skip, the decoding apparatus restores the reference pictures and motion vectors of these macroblocks from the neighboring macroblock reference pictures and motion vectors in the macroblock. Therefore, the image can be restored without any problem. This restoration method is described in H.264. H.264.

  Motion vector detection is performed on image data in a rectangular area centered on a processing target macroblock from a reference picture that is a motion vector detection target. The reference picture is stored in an external memory such as SDRAM (Synchronous Dynamic Random Access Memory).

  The motion vector detection unit reads the image data from the SDRAM and performs motion vector detection. In order to reduce the amount of image data read at this time, there is a technique using a prefetch memory. In addition, there is a technique that includes a plurality of motion vector detection units in order to encode a high-resolution image in real time.

  There is a technique for determining whether or not a motion vector searched by a motion search unit is valid, and determining a motion vector search range based on a distribution of motion vectors determined to be valid.

JP 2008-141288 A JP 2008-5545 A JP 2009-296443 A

  First, P_Skip will be described. FIG. 1 is a diagram illustrating a relationship between a processing target block and adjacent blocks. The adjacent block A shown in FIG. 1 is the block adjacent to the left of the processing target block, the adjacent block B is the upper adjacent block of the processing target block, the adjacent block C is the upper right adjacent block, and the adjacent block D is the processing target block. It is the block on the upper left.

  In the case of MB (macroblock) at the top of the picture, there are no adjacent blocks B, C, and D. In the case of MB at the left end of the picture, there is no adjacent block A. In the case of MB at the right end of the picture, there is no adjacent block C.

FIG. 2 is a flowchart showing the calculation process of the motion vector of the P_Skip macroblock. In the calculation method shown in FIG. 2, the P_Skip vector is determined as follows.
(1) (2) In the case of an MB for which there is no adjacent A or adjacent B, the P_Skip vector is a 0 vector (S11-13, S18).
(3) When the reference picture index ref_idxL0A of adjacent A is 0 and the motion vector of adjacent A is a 0 vector, the P_Skip vector is 0 (S14, S18).
(4) When the reference picture index ref_idxL0B of adjacent B is 0 and the motion vector of adjacent B is a 0 vector, the P_Skip vector is 0 (S15, S18).
If none of (1) to (4) applies, the P_Skip vector is determined in the flowchart of FIG. 3 (S16, S17).

  FIG. 3 is a flowchart showing a normal P_Skip vector calculation process. As shown in FIG. 3, for each component of the motion vectors of adjacent blocks A, B, and C, the median value is a motion vector with P_Skip (S21 to S26). Median (X, Y, Z) is a function that returns an intermediate value among the three values X, Y, Z.

  Here, a conventional technique in which a plurality of motion vector detection units perform motion vector detection in parallel using a prefetch memory will be described. FIG. 4 is a diagram for explaining the operation of the prior art. The example shown in FIG. 4 shows an example in which two motion vector detection units detect a motion vector in parallel. Hereinafter, the two motion vector detection units are also referred to as a core unit 1 and a core unit 2.

  A core 1 illustrated in FIG. 4 indicates a processing target macroblock of the core unit 1, and a core 2 indicates a processing target macroblock of the core unit 2.

  H. H.264 requires information on how adjacent blocks A, B, C, and D are encoded, such as calculation of a P_Skip vector. Therefore, when two macroblocks are processed in parallel, as shown in FIG. 4, the core 1 and the core 2 are shifted in vertical position by 1 MB and shifted in the horizontal direction by 2 MB or more.

  With the relationship of the macroblocks shown in FIG. 4, it can be guaranteed that the encoding has already been completed for the adjacent blocks A, B, C, and D of the macroblock processed by each motion vector detection unit.

  Among the reference pictures, the motion vector detection unit (core unit 1) uses a broken line range rg11 as a search range, and uses an image in the region within the search range rg11 for motion vector detection. The motion vector detection unit (core unit 2) uses the range rg12 of the alternate long and short dash line as a search range, and uses an image in the region within the search range rg12 for motion vector detection.

  When the core unit 1 and the core unit 2 detect a motion vector for the next right macroblock, the region of the search range rg11 and the region of the search range rg12 are both regions shifted by 1 MB to the right. Of the reference pictures of the current motion vector detection MB and the next reference vector of the motion vector detection MB, the overlapping area is maintained in the prefetch memory, and a newly required area ar12 is read from the SDRAM into the prefetch memory. .

  The read area ar12 is overwritten and saved in the area where the unnecessary area ar13 is saved in the prefetch memory. An area ar11 illustrated in FIG. 4 represents an image area stored in the prefetch memory.

  In motion vector detection, it is preferable to detect an optimal motion vector from a wider reference picture area because a reference block more similar to the processing target macroblock may be found.

  However, in the conventional technique, the range (search range rg11 and search range rg12) that is a motion vector detection target is smaller than the area ar11 that is stored and stored in the prefetch memory. Therefore, in the prior art, the motion vector is not detected by efficiently using the area in the prefetch memory.

  Here, when the maximum area in the vertical direction among the areas in the prefetch memory is used as the search range for motion vector detection, there are the following problems. FIG. 5 is a diagram for explaining a problem when the search range is expanded.

  As shown in FIG. 5, for example, in the core unit 1, it is assumed that motion vector detection is performed in the search range rg21, and the motion vectors in the macroblock B and the macroblock C are mv_B and mv_C.

  At this time, the motion vectors of the upper adjacent block B and the upper right adjacent block C of the MB (core 2) processed by the core unit 2 are mv_B and mv_C. Here, if the motion vector mv_A of the left adjacent block A of the core 2 is a 0 vector, the P_Skip vector mv_skip is a 0 vector according to the flowchart shown in FIG.

  However, when the motion vector mv_A of the adjacent block A is not a 0 vector, mv_skip is calculated by the median of mv_A, mv_B, and mv_C. At this time, as shown in FIG. 5, when both mv_B and mv_C are large vectors in the downward direction, mv_skip has a median value, so mv_skip is also a large vector in the downward direction.

  In this case, the reference block at the position of mv_skip for the core 2 is outside the prefetch memory holding range. Therefore, the macro block processed by the core unit 2 cannot calculate the similarity (sum of absolute differences) between the reference block at the position of P_Skip and the processing target macro block. That is, it becomes impossible to determine whether or not the MB can be encoded with P_Skip, and P_Skip with high encoding efficiency is not selected.

  Therefore, an object of the disclosed technique is to provide a moving picture coding apparatus and a moving picture coding method that can improve coding efficiency by appropriately expanding a search range.

A moving image encoding device according to an aspect of the disclosure is a moving image encoding device that performs motion vector detection in parallel for each block line and encodes a moving image, and is a reference image used in each motion vector detection A storage unit that stores an area including the search range of the block, and based on a motion vector of a left adjacent block with respect to the processing target block, the search range in the region stored in the storage unit is made variable, and the motion of the processing target block A plurality of motion vector detection units that perform vector detection, and the storage unit is used in each motion vector detection unit, and is a reference predetermined region in which a predetermined number of pixels are set in each of the horizontal direction and the vertical direction from the processing target block A region including a reference search range indicating, the plurality of motion vector detection units, the motion vector of the left adjacent block is the reference search range A motion vector is detected for the search range in which the vertical component of the region stored in the storage unit is maximized, and the motion vector of the left adjacent block is the vertical of the reference search range. When a region exceeding the component is indicated, motion vector detection is performed on the processing target block using the reference search range .

  According to the disclosed technique, it is possible to improve the encoding efficiency by appropriately widening the search range.

The figure which shows the relationship between a process target block and an adjacent block. The flowchart which shows the calculation process of the motion vector of a P_Skip macroblock. The flowchart which shows a normal P_Skip vector calculation process. The figure for demonstrating operation | movement of a prior art. The figure for demonstrating the problem at the time of extending a search range. The block diagram which shows an example of a structure of the moving image encoder in an Example. The figure which shows an example of the encoding cost calculated in each motion vector detection part. The figure which shows the example of a transition of the area | region of the reference picture hold | maintained. The figure for demonstrating the motion vector of a left adjacent block. The figure which shows an example of the left adjacent block which has the motion vector used as a candidate vector by P_Skip. The figure which shows an example of a search range table (the 1). The figure which shows an example of a search range table (the 2). The figure which shows an example of the initial stage search range in case a process target block is 16x16. The figure which shows an example of the search range restricted when a process target block is 16x16. The figure which shows an example of the initial search range in case a process target block is upper 16 * 8. The figure which shows an example of the initial search range in case a process target block is lower 16x8. The figure which shows an example of the initial search range in case a process target block is lower 16x8. The figure which shows an example of the initial search range in case a process target block is 8 * 16 left. The figure which shows an example of the limited search range in case a process target block is 8 * 16 left. The figure which shows an example of the initial stage search range in case a process target block is right 8x16. The figure which shows an example of the initial search range in case a process target block is 8 * 8 on the upper left. The figure which shows an example of the initial search range in case a process target block is upper right 8x8 The figure which shows an example of the initial search range in case a process target block is lower left 8x8 The figure which shows an example of the limited search range in case a process target block is lower left 8x8. The figure which shows an example of the initial search range in case a process target block is lower right 8x8. The figure which shows an example of the initial stage search range in case a process target block is 16x16. The figure which shows an example of the search range restricted when a process target block is 16x16. The figure which shows an example of the initial search range in case a process target block is upper 16 * 8. The figure which shows an example of the initial search range in case a process target block is lower 16x8. The figure which shows an example of the initial search range in case a process target block is lower 16x8. The figure which shows an example of the initial search range in case a process target block is 8 * 16 left. The figure which shows an example of the limited search range in case a process target block is 8 * 16 left. The figure which shows an example of the initial stage search range in case a process target block is right 8x16. The figure which shows an example of the initial search range in case a process target block is 8 * 8 on the upper left. The figure which shows an example of the initial search range in case a process target block is upper right 8x8 The figure which shows an example of the initial search range in case a process target block is lower left 8x8 The figure which shows an example of the limited search range in case a process target block is lower left 8x8. The figure which shows an example of the initial search range in case a process target block is lower right 8x8. The flowchart which shows an example of the control processing of the moving image encoding in a whole control part. The flowchart which shows an example of a motion vector detection process. The flowchart which shows an example of the determination process (the 1) of a search range. The flowchart which shows an example of the determination process (the 2) of a search range. The block diagram which shows an example of the hardware of a moving image encoder.

  Hereinafter, embodiments will be described in detail with reference to the accompanying drawings.

[Example]
<Configuration>
FIG. 6 is a block diagram illustrating an example of a configuration of the video encoding device 1 in the embodiment. The moving image encoding apparatus 1 in the embodiment includes a moving image encoding unit 10 and a first storage unit 20.

  The moving image encoding apparatus 1 is realized by, for example, a dedicated circuit configured by hard-wired logic or an FPGA (Field Programmable Gate Array) whose circuit configuration is changed by mapping information. In addition, the moving image encoding apparatus 1 is realized by operating a computer including a processor, a DSP, a memory, and the like with an installed program, for example.

  The moving image encoding unit 10 is realized by, for example, LSI (Large Scale Integration), and the first storage unit 20 can be realized as an SDRAM as an external memory.

  The moving image encoding unit 10 illustrated in FIG. 6 performs encoding conforming to the encoding standard. The moving image encoding unit 10 is, for example, H.264. H.264 encoder. An encoding standard applicable to the embodiment is H.264. In addition to H.264, any encoding standard including processing similar to the P_Skip mode may be used.

  The first storage unit 20 accumulates and stores, for example, an image to be encoded captured by a camera, a reference picture used for motion vector detection, encoded data obtained after the encoding process is completed, and the like.

  Next, the moving image encoding unit 10 will be described in detail. The moving image encoding unit 10 includes an overall control unit 100, a second storage unit 200, a first motion vector detection unit 300, a second motion vector detection unit 400, a first encoding unit 500, and a second code. And a conversion unit 600.

  The overall control unit 100 controls the second storage unit 200, the first motion vector detection unit 300, and the second motion vector detection unit 400. For example, the overall control unit 100 controls the first motion vector detection unit 300 and the second motion vector detection unit 400 to simultaneously detect the motion vectors of two macroblocks (MB) in parallel.

  The second storage unit 200 stores a region including a reference picture search range used by the first motion vector detection unit 300 and the second motion vector detection unit 400. For example, the second storage unit 200 stores the holding area ar11 illustrated in FIG. 4 and FIG. 5 and repeats discarding and updating of the holding area each time MB processing proceeds.

  The second storage unit 200 is, for example, a prefetch RAM. The area of the reference picture stored in the second storage unit 200 is read by the first motion vector detection unit 300 and the second motion vector detection unit 400.

  The first motion vector detection unit 300 and the second motion vector detection unit 400 change the search range in the region stored in the second storage unit 200 based on the motion vector of the left adjacent block with respect to the processing target block, The motion vector of the processing target block is detected. The processing target block is, for example, a processing target macro block or a sub macro block.

  Hereinafter, the configuration of the first motion vector detection unit 300 will be described, but the second motion vector detection unit 400 has the same configuration as the first motion vector detection unit 300.

  The first motion vector detection unit 300 includes a control unit 301, an adjacent block vector storage memory 302, a difference absolute value sum calculation unit 303, an original picture memory 304, a comparison unit 305, a result storage memory 306, and a determination unit 307. And have.

  The control unit 301 receives a signal from the overall control unit 100 and controls the first motion vector detection unit 300 as a whole. In addition, the control unit 301 calculates the vector cost of the motion vector to be evaluated, and notifies the comparison unit 305 of the vector cost.

  The control unit 301 determines the search range of the processing target block based on the motion vector of the left adjacent block with respect to the processing target block, and notifies the difference absolute value sum calculation unit 303 of it. In addition, the control unit 301 may determine the search range based on the block size and position of the processing target block. A method for determining the search range will be described later.

  The adjacent block vector storage memory 302 stores a motion vector and a reference frame number as an adjacent block of an MB for which motion vector detection will be performed later based on mb_type and a motion vector output from the determination unit 307 described later.

  The difference absolute value sum calculation unit 303 obtains the data of the processing target MB and the data of the reference block from the original image memory 304 and the second storage unit 200, respectively, and calculates the difference absolute value sum. The difference absolute value sum calculation unit 303 acquires the reference block of the search range notified from the control unit 301 from the second storage unit 200.

  The comparison unit 305 compares the value obtained by adding the difference absolute value sum acquired from the difference absolute value sum calculation unit 303 and the vector cost acquired from the control unit 301 with the minimum cost stored in the result storage memory 306. To do.

  When the comparison unit 305 determines that the added “difference absolute value sum + vector cost (total cost)” is smaller, the content stored in the result storage memory 306 is added to the added “difference absolute value sum”. Overwrite with “+ vector cost”. As a result, the minimum “sum of absolute values + vector cost” is stored in the result storage memory 306 for one processing target block.

  The result storage memory 306 stores “difference absolute value sum + vector cost” output from the comparison unit 305 for each processing target block.

  The determination unit 307 reads the content stored in the result storage memory 306, selects a mode with a small total cost as mb_type, and outputs a motion vector corresponding to mb_type and mb_type to the first encoding unit 500.

  The first motion vector detection unit 300 and the second motion vector detection unit 400 obtain the mb_type and motion vector of the other motion vector detection unit from each other, and store them in their adjacent block vector storage memory.

  Second motion vector detection section 400 performs the same processing as described above, and outputs mb_type and a motion vector corresponding to mb_type to second encoding section 600.

  The first encoding unit 500 or the second encoding unit 600 encodes the processing target MB based on the mb_type and the motion vector acquired from the corresponding motion vector detection unit, and generates a bit stream that is encoded data. .

  Further, the first encoding unit 500 or the second encoding unit 600 writes the generated bit stream in the first storage unit 20. At this time, the first encoding unit 500 or the second encoding unit 600 reads the reference block specified by the motion vector from the second storage unit 200 and uses it. Also, the first encoding unit 500 or the second encoding unit 600 writes the reconstructed image to the first storage unit 20 in order to use it as a reference picture for pictures to be processed after the next. The reconstructed image is also called a local decoded image.

(Encoding cost)
FIG. 7 is a diagram illustrating an example of the coding cost calculated by each motion vector detection unit. In the example illustrated in FIG. 7, the processing target blocks include 16 × 16, 16 × 8, 8 × 16, and 8 × 8 macroblocks or P_Skip. The processing target block may be a sub-block divided up to 4 × 4. Below, 8x8 is demonstrated as the minimum block size.

  The difference absolute value sum calculation unit 303 calculates the difference absolute value sum for each block size, and the control unit 301 calculates the motion vector cost for each block size. The determination unit 307 outputs mb_type indicating the block size that minimizes the coding cost (total cost) at each block size and its motion vector.

(Transition of the area held in the second storage unit)
Next, how the reference picture area stored in the second storage unit 200 changes each time 1 MB processing is described. FIG. 8 is a diagram illustrating a transition example of the area of the reference picture to be held. In the example shown in FIG. 8, for example, it is assumed that motion vector detection is performed on regions of ± 32 pixels in the horizontal direction and ± 16 pixels in the vertical direction.

  Here, the horizontal direction ± 32 and the vertical direction ± 16 are referred to as a reference search range. The reference search range is a limit range when the range is expanded symmetrically in the horizontal direction and the vertical direction from the processing target block within the area stored in the second storage unit.

  An area ar101 shown in (1) of FIG. 8 indicates an initial reading area. The area ar101 is a rectangular area having 48 (16 + 32) pixels in the horizontal direction and 32 (16 + 16) pixels in the vertical direction. A block b101 indicates a 16 × 16 macroblock processed by the second motion vector detection unit 400 (core unit 2), for example. When the motion vector detection of the block b101 is completed, the reference picture area stored in the second storage unit 200 is updated.

  An area ar102 illustrated in (2) of FIG. 8 indicates an area that is read from the first storage unit 20 when the processing of the block b101 ends and the area of the reference picture is updated. The second motion vector detection unit 400 moves the block to the right by one and performs motion vector detection for the block b102.

  An area ar103 illustrated in (3) of FIG. 8 indicates an area that is read from the first storage unit 20 when the processing of the block b102 ends and the area of the reference picture is updated. The second motion vector detection unit 400 moves the block to the right by one and performs motion vector detection for the block b103.

  A block b151 illustrated in (3) of FIG. 8 represents a 16 × 16 macroblock processed by the first motion vector detection unit 300 (core unit 1), for example.

  An area ar104 illustrated in (4) of FIG. 8 indicates an area read from the first storage unit 20 when the processing of the block b103 is completed and the area of the reference picture is updated. The second motion vector detection unit 400 moves the block to the right by one and performs motion vector detection for the block b104. The first motion vector detection unit 300 moves the block to the right by one and performs motion vector detection for the block b152.

  An area ar105 illustrated in (5) of FIG. 8 indicates an area that is read from the first storage unit 20 when the processing of the block b104 is completed and the area of the reference picture is updated. The second motion vector detection unit 400 moves the block to the right by one and performs motion vector detection for the block b105. The first motion vector detection unit 300 moves a block to the right and performs motion vector detection on the block b153.

  An area ar106 illustrated in (6) of FIG. 8 indicates an area that is read from the first storage unit 20 when the processing of the block b105 is completed and the area of the reference picture is updated. An area ar151 illustrated in (6) of FIG. 8 indicates an area that is deleted (discarded) as an unnecessary area when the area ar106 is updated from the area stored in the second storage unit 200.

  The second motion vector detection unit 400 moves the block to the right by one and performs motion vector detection for the block b106. The first motion vector detection unit 300 moves a block to the right and performs motion vector detection on the block b154.

  An area ar107 illustrated in (7) of FIG. 8 indicates an area that is read from the first storage unit 20 when the processing of the block to the left of the block b107 is completed and the reference picture area is updated. An area ar152 illustrated in (7) of FIG. 8 indicates an area that is deleted (discarded) as an unnecessary area when the area ar107 is updated from the area stored in the second storage unit 200.

  The second motion vector detection unit 400 moves the block to the right by one and performs motion vector detection for the block b107. The first motion vector detection unit 300 moves the block to the right by one and performs motion vector detection for the block b155.

  An area ar108 illustrated in (8) of FIG. 8 indicates an area that is read from the first storage unit 20 when the processing of the block b107 is completed and the area of the reference picture is updated. An area ar153 illustrated in (8) of FIG. 8 indicates an area that is deleted (discarded) as an unnecessary area when the area ar108 is updated from the area stored in the second storage unit 200.

  The second motion vector detection unit 400 moves the block to the right by one and performs motion vector detection for the block b108. The first motion vector detection unit 300 moves the block to the right by one and performs motion vector detection for the block b156.

  An area ar109 illustrated in (9) of FIG. 8 indicates an area that is read from the first storage unit 20 when the processing of the block b108 ends and the area of the reference picture is updated. An area ar154 illustrated in (9) of FIG. 8 indicates an area that is deleted (discarded) as an unnecessary area when the area ar109 is updated from the area stored in the second storage unit 200.

  The second motion vector detection unit 400 moves the processing block to the left end two blocks below, and performs motion vector detection on the block b109. The first motion vector detection unit 300 moves the block to the right by one and performs motion vector detection for the block b157.

  By repeating the above processing, the area of the reference picture stored in the second storage unit 200 changes. Each motion vector detection unit in the embodiment determines a search range as wide as possible from the reference picture area stored in the second storage unit 200 and detects a motion vector. At this time, each motion vector detection unit determines the search range without reducing the probability of occurrence of P_Skip.

<Determination of search range>
Next, search range determination processing will be described. First, the motion vector of the left adjacent block used for determining the search range of the processing target block will be described.

  FIG. 9 is a diagram for explaining the motion vector of the left adjacent block. A block b201 shown in FIG. 9 is a processing block of the maximum size, and a block b202 is an encoded left adjacent block. The maximum size processing target block is, for example, a 16 × 16 macroblock.

  As illustrated in FIG. 9, the left adjacent block for the processing target block is a block including the lower left pixel among the maximum size blocks adjacent to the left of the maximum size processing target block. In the example shown in FIG. 9, when the motion vector of the block including the lower left pixel of the left adjacent block b202 is processed in the next lower block line, H.264 is processed. This is a candidate vector of an adjacent block used in H.264 P_Skip. A candidate vector refers to a motion vector that is a target of media processing of P_Skip.

  FIG. 10 is a diagram illustrating an example of a left adjacent block having a motion vector that is a candidate vector in P_Skip. In the example shown in FIG. 10A, when the mb_type of the left adjacent block is 16 × 16, the motion vector of the 16 × 16 block b301 is used as a candidate vector in P_Skip when processing the next lower block line. .

  In the example shown in FIG. 10B, when the mb_type of the left adjacent block is 16 × 8, the motion vector of the lower 16 × 8 block b302 is used as a candidate vector in P_Skip when processing the next lower block line. It is done.

  In the example shown in FIG. 10C, when the mb_type of the left adjacent block is 8 × 16, the motion vector of the left 8 × 16 block b303 is used as a candidate vector in P_Skip when processing the next lower block line. It is done.

  In the example illustrated in FIG. 10D, when the mb_type of the left adjacent block is 8 × 8, the motion vector of the lower left 8 × 8 block b304 is used as a candidate vector in P_Skip when processing the next lower block line. It is done. The blocks b301 to b304 located at the block size and position shown in FIG. 10 are hereinafter also referred to as candidate blocks.

  As shown in FIG. 10, upper 16 × 8, right 8 × 16, upper left 8 × 8, upper right 8 × 8, lower right, which do not become P_Skip candidate blocks as upper adjacent blocks when processing the next lower block line For an 8 × 8 block, the search range may be wider than the reference search range. Hereinafter, these blocks are also referred to as non-candidate blocks.

  That is, when the processing target block is a non-candidate block, each motion vector detection unit performs motion vector detection in a range wider than the reference search range. The range wider than the reference search range is, for example, a search range in which the vertical component of the region stored in the second storage unit 200 is maximized.

  Further, even if the processing target block is a candidate block, if the motion vector of the processed candidate block on the left is a motion vector indicating a region within the reference search range, each motion vector detection unit Motion vector detection is performed in a wider range than the reference search range.

  Further, when the processing target block is a candidate block, if the motion vector of the processed candidate block on the left is a motion vector indicating a region within a range wider than the reference search range, each motion vector detection unit, Motion vector detection is performed within the reference search range.

  This is because, as shown in FIG. 5, when a motion vector indicating an area within a range wider than the reference search range is obtained between blocks adjacent in the horizontal direction, the motion vector of P_Skip is processed during processing of the next lower block line. This is because it indicates the outside of the area stored in the second storage unit 200.

  Therefore, when a large motion vector is obtained so that a large motion vector is not continuously obtained in the horizontal direction, the search range of the next block is limited to the reference search range. Accordingly, it can be ensured that the reference block at the position indicated by the motion vector of P_Skip is within the area stored in the second storage unit 200. A large motion vector refers to a motion vector indicating a reference block within a range wider than the standard search range.

  Further, each motion vector detection unit may perform processing so that a large motion vector is not continuously obtained in a horizontal block regardless of the block size and position of the processing target block. Even with this processing alone, it is possible to improve the coding efficiency by expanding the search range without reducing the probability of occurrence of P_Skip.

(Search range table)
Next, an initial search range determined by each motion vector detection unit will be described. FIG. 11 is a diagram illustrating an example of the search range table (part 1). The search range table illustrated in FIG. 11 is a search range table used by the first motion vector detection unit 300 that processes even-numbered macroblocks, for example.

  This search range table is held by the control unit 301, for example. When notified by the overall control unit 100 of the block size and position of the processing target block, the control unit 301 determines an initial search range using the search range table and notifies the difference absolute value sum calculation unit 303 of the initial search range. .

  The reason for calling the initial search range is that when the motion vector of the left adjacent block is a large motion vector, the search range is limited and the search range changes.

  FIG. 12 is a diagram illustrating an example of the search range table (part 2). The search range table shown in FIG. 12 is a search range table used by the second motion vector detection unit 400 that processes, for example, odd-numbered macroblocks. This search range table is held by the control unit of the second motion vector detection unit 400, for example. The determination of the initial search range is the same as the processing of the first motion vector detection unit 300.

  The examples shown in FIGS. 11 and 12 are examples in which the reference search range is horizontal ± 32 and vertical ± 16.

(Specific example of search range)
Next, a specific example of the search range determined by each motion vector detection unit will be described. First, the search range determined by the first motion vector detection unit 300 will be described. In the following drawings, the direction of the arrow outside the picture in both horizontal and vertical directions is +.

  FIG. 13A is a diagram illustrating an example of an initial search range when the processing target block is 16 × 16. As illustrated in FIG. 13A, using the search range table illustrated in FIG. 11, the control unit 301 determines horizontal ± 32, vertical +32, and −16 as initial search ranges.

  FIG. 13B is a diagram illustrating an example of a limited search range when the processing target block is 16 × 16. As illustrated in FIG. 13B, when the motion vector of the candidate block on the left (see FIG. 10) is a large motion vector, the control unit 301 limits the search range to the reference search range (horizontal ± 32, vertical ± 16). To do. This is to ensure that the motion vector of P_Skip refers to the position in the area stored in the second storage unit 200 when processing the next lower block line (when processing by the core unit 2). .

  FIG. 14A is a diagram illustrating an example of an initial search range when the processing target block is the upper 16 × 8. As illustrated in FIG. 14A, using the search range table illustrated in FIG. 11, the control unit 301 determines horizontal ± 32, vertical +40, and −16 as initial search ranges. Since the upper 16 × 8 block is a non-candidate block, this initial search range is used for motion vector detection.

  FIG. 14B is a diagram illustrating an example of an initial search range when the processing target block is lower 16 × 8. As illustrated in FIG. 14B, using the search range table illustrated in FIG. 11, the control unit 301 determines horizontal ± 32, vertical +32, and −24 as initial search ranges.

  FIG. 14C is a diagram illustrating an example of a limited search range when the processing target block is lower 16 × 8. As illustrated in FIG. 14C, when the motion vector of the candidate block on the left is a large motion vector, the control unit 301 restricts the search range to the reference search range (horizontal ± 32, vertical ± 16).

  FIG. 15A is a diagram illustrating an example of an initial search range when the processing target block is 8 × 16 left. As illustrated in FIG. 15A, using the search range table illustrated in FIG. 11, the control unit 301 determines horizontal +40, −32, vertical +32, and −16 as initial search ranges.

  FIG. 15B is a diagram illustrating an example of a limited search range when the processing target block is 8 × 16 left. As illustrated in FIG. 15B, when the motion vector of the candidate block on the left is a large motion vector, the control unit 301 restricts the search range to the reference search range (horizontal ± 32, vertical ± 16).

  FIG. 15C is a diagram illustrating an example of an initial search range when the processing target block is 8 × 16 on the right. As illustrated in FIG. 15C, using the search range table illustrated in FIG. 11, the control unit 301 determines horizontal +32, −40, vertical +32, and −16 as initial search ranges. Since the right 8 × 16 block is a non-candidate block, this initial search range is used for motion vector detection.

  FIG. 16A is a diagram illustrating an example of an initial search range when the processing target block is the upper left 8 × 8. As illustrated in FIG. 16A, using the search range table illustrated in FIG. 11, the control unit 301 determines horizontal +40, −32, vertical +40, and −16 as the initial search range. Since the upper left 8 × 8 block is a non-candidate block, this initial search range is used for motion vector detection.

  FIG. 16B is a diagram illustrating an example of an initial search range when the processing target block is an upper right 8 × 8. As illustrated in FIG. 16B, using the search range table illustrated in FIG. 11, the control unit 301 determines horizontal +32, −40, vertical +40, and −16 as the initial search range. Since the upper right 8 × 8 block is a non-candidate block, this initial search range is used for motion vector detection.

  FIG. 16C is a diagram illustrating an example of the initial search range when the processing target block is the lower left 8 × 8. As illustrated in FIG. 16C, using the search range table illustrated in FIG. 11, the control unit 301 determines horizontal +40, −32, vertical +32, and −24 as the initial search range.

  FIG. 16D is a diagram illustrating an example of a limited search range when the processing target block is the lower left 8 × 8. As illustrated in FIG. 16D, when the motion vector of the candidate block on the left is a large motion vector, the control unit 301 limits the search range to the reference search range (horizontal ± 32, vertical ± 16).

  FIG. 16E is a diagram illustrating an example of the initial search range when the processing target block is the lower right 8 × 8. As illustrated in FIG. 16E, the control unit 301 determines horizontal +32, −40, vertical +32, and −24 as initial search ranges using the search range table illustrated in FIG. Since the lower right 8 × 8 block is a non-candidate block, this initial search range is used for motion vector detection.

  Next, the search range determined by the second motion vector detection unit 400 will be described. FIG. 17A is a diagram illustrating an example of an initial search range when the processing target block is 16 × 16. As illustrated in FIG. 17A, using the search range table illustrated in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal ± 32, vertical +16, and −32 as initial search ranges.

  FIG. 17B is a diagram illustrating an example of a limited search range when the processing target block is 16 × 16. As illustrated in FIG. 17B, the control unit of the second motion vector detection unit 400 sets the search range as the reference search range (horizontal ± 32) when the motion vector of the candidate block on the left (see FIG. 10) is a large motion vector. , Vertical ± 16). This is to ensure that the motion vector of P_Skip refers to the position in the area stored in the second storage unit 200 when processing the next lower block line (when processing by the core unit 1). .

  FIG. 18A is a diagram illustrating an example of the initial search range when the processing target block is the upper 16 × 8. As shown in FIG. 18A, using the search range table shown in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal ± 32, vertical +24, and −32 as initial search ranges. Since the upper 16 × 8 block is a non-candidate block, this initial search range is used for motion vector detection.

  FIG. 18B is a diagram illustrating an example of an initial search range when the processing target block is lower 16 × 8. As illustrated in FIG. 18B, using the search range table illustrated in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal ± 32, vertical +16, and −40 as the initial search range.

  FIG. 18C is a diagram illustrating an example of a limited search range when the processing target block is lower 16 × 8. As illustrated in FIG. 18C, the control unit of the second motion vector detection unit 400 determines that the search range is the reference search range (horizontal ± 32, vertical ± 16) when the motion vector of the candidate block on the left is a large motion vector. Limit to.

  FIG. 19A is a diagram illustrating an example of an initial search range when the processing target block is 8 × 16 left. As illustrated in FIG. 19A, using the search range table illustrated in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal +40, −32, vertical +16, and −32 as initial search ranges.

  FIG. 19B is a diagram illustrating an example of a limited search range when the processing target block is 8 × 16 left. As illustrated in FIG. 19B, the control unit of the second motion vector detection unit 400 determines that the search range is a reference search range (horizontal ± 32, vertical ± 16) when the motion vector of the candidate block on the left is a large motion vector. Limit to.

  FIG. 19C is a diagram illustrating an example of an initial search range when the processing target block is 8 × 16 on the right. As illustrated in FIG. 19C, using the search range table illustrated in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal +32, −40, vertical +16, and −32 as the initial search range. Since the right 8 × 16 block is a non-candidate block, this initial search range is used for motion vector detection.

  FIG. 20A is a diagram illustrating an example of an initial search range when the processing target block is the upper left 8 × 8. As illustrated in FIG. 20A, using the search range table illustrated in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal +40, −32, vertical +24, and −32 as initial search ranges. Since the upper left 8 × 8 block is a non-candidate block, this initial search range is used for motion vector detection.

  FIG. 20B is a diagram illustrating an example of an initial search range when the processing target block is 8 × 8 in the upper right. As illustrated in FIG. 20B, using the search range table illustrated in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal +32, −40, vertical +24, and −32 as initial search ranges. Since the upper right 8 × 8 block is a non-candidate block, this initial search range is used for motion vector detection.

  FIG. 20C is a diagram illustrating an example of the initial search range when the processing target block is the lower left 8 × 8. As illustrated in FIG. 20C, using the search range table illustrated in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal +40, −32, vertical +16, and −40 as the initial search range.

  FIG. 20D is a diagram illustrating an example of a limited search range when the processing target block is the lower left 8 × 8. As illustrated in FIG. 20D, the control unit of the second motion vector detection unit 400 determines that the search range is the reference search range (horizontal ± 32, vertical ± 16) when the motion vector of the candidate block on the left is a large motion vector. Limit to.

  FIG. 20E is a diagram illustrating an example of the initial search range when the processing target block is the lower right 8 × 8. As shown in FIG. 20E, using the search range table shown in FIG. 12, the control unit of the second motion vector detection unit 400 determines horizontal +32, −40, vertical +16, and −40 as the initial search range. Since the lower right 8 × 8 block is a non-candidate block, this initial search range is used for motion vector detection.

  As described above, each motion vector detection unit performs motion vector detection by changing the search range based on the block size and position of the processing target block and the motion vector of the left adjacent block. Note that each motion vector detection unit may vary the search range based only on the motion vector of the block adjacent to the left of the processing target block.

<Operation>
Next, the operation of the video encoding device 1 will be described. FIG. 21 is a flowchart illustrating an example of a moving image coding control process in the overall control unit 100. The process shown in FIG. 21 is a process for one picture.

  In step S101 illustrated in FIG. 21, the overall control unit 100 instructs the second storage unit 200 to read an initial area. The processing area is, for example, an area ar101 shown in (1) of FIG.

  In step S102, the overall control unit 100 instructs each motion vector detection unit to start a 1 MB motion vector detection process. The motion vector detection process will be described later with reference to FIG.

  In step S <b> 103, upon receiving notification of completion of motion vector detection from each motion vector detection unit, the overall control unit 100 determines whether or not an unprocessed MB remains. If the process is completed for all MBs (step S103—YES), this process ends. If there is an unprocessed MB (step S103—NO), the process proceeds to step S104.

  In step S104, the overall control unit 100 instructs the second storage unit 200 to update the area of the held reference picture in order to perform motion vector detection for the next MB. See FIG. 8 for an example of this update. When the process in step S104 is completed, the process proceeds to step S02 in order to perform a motion vector detection process in the next MB.

  Next, motion vector detection processing for the processing target block will be described. FIG. 22 is a flowchart illustrating an example of a motion vector detection process. In the example illustrated in FIG. 22, the process performed by the first motion vector detection unit 300 for 1 MB is described, but the same process is performed by the second motion vector detection unit.

  In step S201, the control unit 301 selects processing target blocks having different block sizes and positions of the processing target blocks in the processing target MB. For example, the control unit 301 has 16 × 16, upper 16 × 8, lower 16 × 8, left 8 × 16, right 8 × 16, upper left 8 × 8, upper right 8 × 8, lower left 8 × 8, lower right 8 ×. Select in order of 8.

  In step S202, the control unit 301 determines a search range based on the motion vector of the left adjacent block with respect to the processing target block, or based on the motion vector of the left adjacent block, the block size and the position of the processing target block. This determination process will be described later with reference to FIG.

  In step S203, the difference absolute value sum calculation unit 303 detects a motion vector having the minimum difference absolute value sum in the determined search range. When a motion vector having a smaller difference absolute value sum is found, the comparison unit 305 stores the motion vector in the result storage memory 306 and overwrites the temporary motion vector and the minimum difference absolute value sum.

  In step S204, the control unit 301 determines whether the processing has been completed for all the processing target blocks. If the processing is completed for all the processing target blocks (step S204—YES), the process proceeds to step S205. If the processing is not completed for all the processing target blocks (step S204—NO), the process returns to step S201.

  In step S205, the difference absolute value sum calculation unit 303 calculates the difference absolute value sum at the position of the P_Skip vector with a block size of 16 × 16. The comparison unit 305 writes the P_Skip vector and the sum of absolute differences in the result storage memory 306. At the end of step S205, the contents as shown in FIG.

  In step S206, the determination unit 307 reads the contents of the result storage memory 306, calculates the total cost for the block size 16 × 16, 16 × 8, 8 × 16, 8 × 8, P_Skip, and the total cost is the minimum. The block size mb_type is determined.

  In step S <b> 207, the determination unit 307 writes the determined mb_type and the motion vector corresponding to the mb_type in the adjacent block vector storage memory 302.

  In step S208, the determination unit 307 outputs the determined mb_type and the motion vector corresponding to the mb_type to the first encoding unit 500 and the second motion vector detection unit 400.

  Next, search range determination processing will be described. FIG. 23 is a flowchart illustrating an example of search range determination processing (part 1). The process illustrated in FIG. 23 is a search range determination process performed for each processing target block in the first motion vector detection unit 300.

  In step S301, the control unit 301 determines an initial search range with reference to FIG. 11 based on the block size and position of the processing target block.

  In step S302, the control unit 301 determines whether the processing target block is a candidate block and the MV vertical component of the left adjacent MB exceeds the vertical component of the reference search range. Specifically, the candidate blocks are, for example, 16 × 16, lower 16 × 8, left 8 × 16, and lower left 8 × 8, and the vertical component of the reference search range is, for example, +16.

  If the condition of step S302 is satisfied (step S302—YES), the process proceeds to step S303. If this condition is not satisfied (step S302—NO), the determination process ends.

  In step S303, the control unit 301 changes the determined initial search range to the reference search range. The reference search range is, for example, horizontal ± 32 and vertical ± 16.

  FIG. 24 is a flowchart illustrating an example of search range determination processing (part 2). The processing illustrated in FIG. 24 is search range determination processing performed for each processing target block in the second motion vector detection unit 400.

  In step S401, the control unit of the second motion vector detection unit 400 determines an initial search range with reference to FIG. 12 based on the block size and position of the processing target block.

  In step S402, the control unit of the second motion vector detection unit 400 determines whether the processing target block is a candidate block and the vertical component of the MV of the left adjacent MB exceeds the vertical component of the reference search range. Specifically, the candidate blocks are, for example, 16 × 16, lower 16 × 8, left 8 × 16, and lower left 8 × 8, and the vertical component of the reference search range is, for example, +16.

  If the condition of step S402 is satisfied (step S402-YES), the process proceeds to step S403, and if this condition is not satisfied (step S402-NO), the determination process ends.

  In step S403, the control unit of the second motion vector detection unit 400 changes the determined initial search range to the reference search range. The reference search range is, for example, horizontal ± 32 and vertical ± 16.

  As described above, by determining the search range, it can be guaranteed that the P_Skip vector points to an area in the second storage unit 200 in all MBs.

  As described above, according to the embodiment, it is possible to improve the encoding efficiency by appropriately expanding the search range. In addition, according to the embodiment, the search range can be expanded without reducing the occurrence probability of P_Skip, and it can be ensured that the area pointed to by the MV of P_Skip is in the second storage unit 200.

[Modification]
In the above embodiment, the description has been made using two motion vector detection units for the sake of simplification. However, even if there are three or more motion vector detection units, if the processing described in the embodiment is extended, the implementation can be performed. As in the example, the search range can be made variable.

  In addition, the moving image coding apparatus represents the moving image coding unit 10, and the first storage unit 20 may be represented as an external memory because it exists outside the moving image coding unit 10. That is, the moving image encoding device may represent an LSI part of the moving image encoding unit 10 or an apparatus including the moving image encoding unit 10 and the first storage unit 20.

  FIG. 25 is a block diagram illustrating an example of hardware of the video encoding device 2. In the example shown in FIG. 25, the processing of the above embodiment is a moving image coding program, and the moving image coding program is read by the control unit 501 of the moving image coding apparatus 2 to execute the processing of the above embodiment.

  The moving picture encoding apparatus 2 illustrated in FIG. 25 includes a control unit 701, a main storage unit 702, an auxiliary storage unit 703, a drive device 704, a network I / F unit 706, an input unit 707, and a display unit 708. Have These components are connected to each other via a bus so as to be able to transmit and receive data.

  The control unit 701 is a CPU that controls each device, calculates data, and processes in a computer. The control unit 701 is an arithmetic device that executes a moving image encoding program stored in the main storage unit 702 or the auxiliary storage unit 703. The control unit 701 receives data from the input unit 707 or the storage device, calculates and processes the data. The data is output to the display unit 708, a storage device, or the like.

  The main storage unit 702 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 701. It is.

  The second storage unit 200 in the embodiment is, for example, a prefetch RAM, each memory of the motion vector detection unit is, for example, a RAM, and the first storage unit 20 is, for example, an SDRAM.

  The auxiliary storage unit 703 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 704 reads the program from the recording medium 705, for example, a flexible disk, and installs it in the storage device.

  In addition, a predetermined program is stored in the recording medium 705, and the program stored in the recording medium 705 is installed in the moving image encoding apparatus 30 via the drive device 704. The installed predetermined program can be executed by the moving image encoding device 30.

  The network I / F unit 706 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 encoding apparatus 2.

  The input unit 707 includes a keyboard having cursor keys, numeric input, various function keys, and the like, and a mouse and a slice pad for performing key selection on the display screen of the display unit 708. The input unit 707 is a user interface for a user to give an operation instruction to the control unit 701 and input data.

  The display unit 708 includes a CRT (Cathode Ray Tube), an LCD (Liquid Crystal Display), and the like, and performs display according to display data input from the control unit 701.

  For example, the overall control unit 100, each motion vector detection unit, and each encoding unit can be realized by the control unit 701, the main storage unit 702 as a work memory, or the like.

  Thus, the moving image encoding process described in the above embodiment 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 can be realized. For example, the motion vector detection process described above may be realized as a program.

  It is also possible to record the moving picture coding program on the recording medium 705 and cause the computer or portable terminal to read the recording medium 705 on which the program is recorded, thereby realizing the moving picture coding process described above. .

  The recording medium 705 is a recording medium that records information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, a magneto-optical disk, etc., or information electrically, such as a ROM, flash memory, or the like. Various types of recording media such as a semiconductor memory for recording can be used. In addition, each processing unit described in the above-described embodiments may be implemented by mounting one or more processing units on one or more integrated circuits.

  Although the embodiment has been described in detail above, the present invention is not limited to this embodiment, and various modifications and changes other than the above-described embodiment and modifications can be made within the scope described in the claims. is there.

In addition, the following additional remarks are disclosed regarding the above Example.
(Appendix 1)
A moving image encoding apparatus that performs motion vector detection in parallel for each block line and encodes a moving image,
A storage unit for storing a region including a search range of a reference image used in each motion vector detection;
A plurality of motion vector detection units for varying the search range in the region stored in the storage unit based on the motion vector of the left adjacent block with respect to the processing target block, and detecting a motion vector of the processing target block;
A video encoding device comprising:
(Appendix 2)
The storage unit
Storing a region including a reference search range indicating a predetermined reference region used in each motion vector detection unit;
The plurality of motion vector detection units,
When a motion vector is detected for a search range in which the vertical component of the region stored in the storage unit is maximized, and the motion vector of the left adjacent block indicates a region exceeding the vertical component of the reference search range The moving picture coding apparatus according to supplementary note 1, wherein motion vector detection is performed on the processing target block using the reference search range.
(Appendix 3)
The plurality of motion vector detection units,
The moving image encoding apparatus according to appendix 1 or 2, wherein the search range is made variable based further on the size and position of the processing target block.
(Appendix 4)
The plurality of motion vector detection units,
The processing target block is H.264. The moving picture coding apparatus according to supplementary note 3, wherein a motion vector is detected for the reference search range if the block is an adjacent block used in H.264 P_Skip.
(Appendix 5)
When the encoding process of the processing target block of each block size is completed, a part of the area that is the search range in the next processing target block is read from the image storage unit that stores the reference image, and is stored in the storage unit The moving picture encoding device according to any one of supplementary notes 1 to 4, further comprising an overall control unit that controls updating of the video.
(Appendix 6)
A moving image encoding method executed by a moving image encoding apparatus that performs motion vector detection in parallel for each block line by a plurality of motion vector detection units and encodes a moving image,
Storing a region including a search range of a reference image used in each motion vector detection in a storage unit;
In each motion vector detection unit, based on the motion vector of the left adjacent block with respect to the processing target block, the search range in the region stored in the storage unit is made variable, and the motion vector detection of the processing target block is performed A moving image encoding method.

1 and 2 Moving picture coding apparatus 10 Moving picture coding unit 20 First storage unit 100 Overall control unit 200 Second storage unit 300 First motion vector detection unit 301 Control unit 302 Adjacent block vector storage memory 303 Difference absolute value sum calculation Unit 304 original image memory 305 comparison unit 306 result storage memory 307 determination unit 400 second motion vector detection unit 500 first encoding unit 600 second encoding unit 701 control unit 702 main storage unit

Claims (3)

A moving image encoding apparatus that performs motion vector detection in parallel for each block line and encodes a moving image,
A storage unit for storing a region including a search range of a reference image used in each motion vector detection;
A plurality of motion vector detection units for varying the search range in the region stored in the storage unit based on the motion vector of the left adjacent block with respect to the processing target block, and detecting a motion vector of the processing target block;
With
The storage unit
Used in each motion vector detection unit, storing an area including a reference search range indicating a predetermined reference area in which a predetermined number of pixels is set in each of the horizontal direction and the vertical direction from the processing target block;
The plurality of motion vector detection units,
When the motion vector of the left adjacent block indicates an area within the reference search range, motion vector detection is performed on the search range that maximizes the vertical component of the area stored in the storage unit, and the left A moving picture encoding apparatus that performs motion vector detection using the reference search range for the processing target block when a motion vector of an adjacent block indicates an area exceeding a vertical component of the reference search range. The plurality of motion vector detection units,
The processing target block is H.264. If a block that do not neighbor blocks used in the 264 P_Skip the video encoding apparatus according to claim 1, wherein performing the motion vector detection for the search range to maximize the vertical component of the stored in the storage area . A moving image encoding method executed by a moving image encoding apparatus that performs motion vector detection in parallel for each block line by a plurality of motion vector detection units and encodes a moving image,
Storing a region including a search range of a reference image used in each motion vector detection in a storage unit;
In each motion vector detection unit, based on the motion vector of the left adjacent block with respect to the processing target block, the search range in the region stored in the storage unit is made variable, and the motion vector detection of the processing target block is performed Have
The storage unit
Used in each motion vector detection unit, storing an area including a reference search range indicating a predetermined reference area in which a predetermined number of pixels is set in each of the horizontal direction and the vertical direction from the processing target block;
The plurality of motion vector detection units,
When the motion vector of the left adjacent block indicates an area within the reference search range, motion vector detection is performed on the search range that maximizes the vertical component of the area stored in the storage unit, and the left A moving picture coding method for performing motion vector detection using the reference search range for the processing target block when a motion vector of an adjacent block indicates an area exceeding a vertical component of the reference search range.

Images