KR20100100540A - Moving pictures encoder/decoder and method for determining block mode for encoding/decoding moving pictures in the moving pictures encoder/decoder - Google Patents

Abstract

PURPOSE: Moving pictures encoder/decoder and a method for determining a block mode for encoding/decoding moving pictures in the moving pictures encoder/decoder are provided to efficiently reduce the operation amount of the encoder/decoder by increasing the speed of a motion prediction operation. CONSTITUTION: When a video frame is inputted to a video encoder, a motion estimator performs the motion prediction of 4×4, 8×8 and 16×16 blocks(S70,S72). The video encoder determines candidate modes(S74), and calculates a rate-distortion value for each candidate mode(S76). In addition video encoder determines an optimal block mode(S78), and encodes a video frame(S80).

Description

MOVING PICTURES ENCODER / DECODER AND METHOD FOR DETERMINING BLOCK MODE FOR ENCODING / DECODING MOVING PICTURES IN THE MOVING PICTURES ENCODER / DECODER}

The present invention relates to a method of determining a block mode for video encoding / decoding and a video encoder / decoder for the same. More particularly, the amount of computation can be effectively reduced by speeding up a motion prediction operation that requires a large amount of computation during video encoding / decoding. The present invention relates to a block mode determination method for encoding a moving picture and a video encoder therefor.

ITU-T and Moving Picture Experts Group (MPEG) have developed H.264 / AVC, a new standard that provides superior and superior video image compression performance compared to earlier developed MPEG4 and H.263. Because of the potential impact of AVC, H.264 / AVC is expected to replace the MPEG2 system, which dominates today's digital video systems. In addition, H.264 / AVC's video compression standard guarantees high compression rate and high picture quality, so it is expected to be applied as a very effective video compression standard even in high bandwidth system such as HDTV system or future interactive TV.

On the other hand, H.264 / AVC has a high compression efficiency but a very high complexity. Especially, a transcoder that performs video transcoding includes a decoder and an encoder. As we increase the complexity.

In general, the amount of computation of an encoder has an absolute influence on the amount of computation required for the entire encoding. In particular, in the case of H.264 / AVC using various block modes and multiple reference images, the complexity of motion prediction is very high. For example, when the motion prediction is performed on one or more forward and backward reference frames such as the H.264 / AVC compression coding scheme, the complexity is very high. In the case of 264 / AVC, the hierarchical scalability is required to have temporal scalability. Red B-pictures can be used, and the complexity increases exponentially when motion prediction is performed using various blocks for hierarchical B-pictures.

1 is a flowchart illustrating a method of determining a fast motion mode of a video encoder according to the prior art.

Referring to FIG. 1, the video encoder first inputs an input image step S10 of using the reference frame as the input image with respect to the current frame of the input image, skips (SKIP) / direct (Direct) and 16 × 16 block mode motion prediction, and Compensation step S12, skip / direct mode and 16x16 mode rate-distortion (RD) cost calculation step (S14), 16x16 block mode hypothesis verification step (S16), hypothesis Determining whether there is a confidence interval of the verification statistics (S18), 16x8, 8x16, and 8x8 block mode motion prediction and compensation step (S20), and 16x8, 8x16, and 8x8 block modes Whether the rate-distortion (RD) cost calculation step (S122), the minimum determination of the 8 × 8 block mode rate-distortion cost, that is, the 8 × 8 block mode hypothesis verification step (S24), and the confidence interval of the hypothesis test statistics Determination step S26, 8x4, 4x8 and 4x4 block mode motion prediction and compensation step S28, and 8x4, 4x8 and 4x4 block mode rate-distortion (RD) cost Calculation step (S30), Minimum rate-distortion (RD) cost determination and final motion block mode determination steps S32 are continuously performed in conjunction.

As described above, in the determination of the motion mode, when performing the hypothesis verification on the 16 × 16 macroblock (S18), if the verification statistic value exists within a given confidence interval, the final motion mode is determined as the 16 × 16 block mode. The mode prediction and compensation process can be completed early. In this case, the motion prediction and compensation process for the lower 16x8, 8x16, and 8x8 block modes may not be performed.

If the verification statistic value does not exist within the given confidence interval, after performing the motion prediction and compensation process for the 16x8, 8x16, and 8x8 block modes (S20), the rate-distortion cost is calculated ( S22). Here, when the 16x8 or 8x16 block mode generates the minimum rate-distortion cost value, the final motion block mode is determined as the 16x8 or 8x16 block mode (S24). However, if the 8 × 8 block mode generates the minimum rate-distortion cost value, the hypothesis verification process is performed (S26). After the determination, the motion prediction and compensation process ends (S32).

If the verification statistic value does not exist within the given confidence interval, after performing the motion prediction and compensation process for 8x4, 4x8 and 4x4 block modes (S28), the rate-distortion cost value is calculated. (S30). Here, the block mode that generates the minimum rate-distortion cost value is determined as the final motion mode.

Such a conventional computation method has a lot of obstacles in implementing a video encoder that operates in real time due to its high complexity, and accordingly, there is a problem that the computation depends only on hardware, not software, when predicting motion. Therefore, there is a need for a method capable of effectively reducing the amount of computation by speeding up a motion prediction operation that requires a large amount of computation during video encoding.

An object of the present invention is to provide a block mode determination method for video encoding and a video encoder therefor, which can effectively reduce the amount of computation by speeding up a motion prediction operation that requires a large amount of computation during video encoding.

According to an embodiment of the present invention, a method of determining a block mode for encoding a video may include: performing motion prediction on a macroblock of a previously stored size in units of macroblocks when a video to be encoded is input, and according to the motion prediction. Selecting at least one of the block modes of the macroblock as a candidate mode, calculating a rate-distortion value corresponding to each of the at least one candidate mode, and calculating the rate-distortion value according to the calculation result of the rate-distortion value Determining one of the at least one candidate mode as an optimal block mode.

The video encoder for determining a block mode according to the present invention, when a video to be encoded is input, performs a motion prediction for each macroblock of a stored size of the video for each macroblock unit, and has the macroblock according to the motion prediction. Selecting at least one candidate mode among block modes included in the macroblock when the at least one candidate mode is selected, a motion estimator for calculating a rate-distortion value corresponding to each of the at least one candidate mode; And a macroblock mode determiner for determining any one of the at least one candidate mode as an optimal block mode according to the calculation result of the rate-distortion value.

The method for determining a block mode for video decoding according to the present invention includes, when a bitstream to be decoded is input, performing motion prediction on a macroblock of a previously stored size in units of macroblocks; Selecting at least one of the block modes of the macroblock as a candidate mode, calculating a rate-distortion value corresponding to each of the at least one candidate mode, and calculating a rate-distortion value Accordingly, the method includes determining any one of the at least one candidate mode as an optimal block mode.

When a bitstream to be decoded is input, the video decoder for determining a block mode according to the present invention performs motion prediction for each macroblock unit of a macroblock having a previously stored size among the bitstreams, and performs the macroblock according to the motion prediction. A motion estimator for calculating a rate-distortion value corresponding to each of the at least one candidate mode when at least one candidate mode among the block modes is selected, and at least one candidate mode among the block modes of the macroblock; And a macroblock mode determiner for determining any one of the at least one candidate mode as an optimal block mode according to the calculation result of the rate-distortion value.

According to the present invention, there is provided a video encoder capable of effectively reducing the computation amount by speeding up a motion prediction operation that requires a large amount of computation during video encoding / decoding, and a method of determining a high speed motion mode of the video encoder / decoder.

In addition, according to the present invention, a video encoder / decoder and a fast motion mode of the video encoder / decoder capable of effectively shortening the image quality and minimizing the amount of bits generated while performing motion prediction using various motion blocks. Provided are a determination method and apparatus.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be noted that the same elements in the figures are denoted by the same reference numerals wherever possible. In the following description and the annexed drawings, detailed descriptions of well-known functions and configurations that may unnecessarily obscure the subject matter of the present invention will be omitted.

2 is a block diagram illustrating a schematic configuration of a video encoder according to an embodiment of the present invention.

The video encoder 40 according to the present embodiment includes a discrete cosine transform (DCT) operator 44, a quantizer 46, an inverse quantizer 50, and an inverse discrete cosine transform (IDCT). ) An operator 52, a frame memory 60, a motion compensator 58, a subtractor 42, an adder 56, an entropy encoder 48, a motion estimator 62, and a macroblock mode determiner 54. .

It is assumed that the video encoder 40 according to the present embodiment is an H.264 / AVC scheme encoding apparatus, and will be described below based on the H.264 / AVC scheme encoding apparatus.

The image frame input in units of macroblocks is converted into units of macroblocks by the DCT operator 52 and quantized by the quantizer 46. The quantized video frame output from the quantizer 46 is decoded through the inverse quantizer 50 and the IDCT operator 52, and the decoded data is input to the adder 56. The input image data is reconstructed and stored in the frame memory 60, and then provided as a reference image for motion estimation of the input image frame.

The motion estimator 62 performs motion estimation by receiving at least one of the reference images stored in the frame memory 60 with respect to the image frame in the macroblock unit currently input, and performs a motion vector and a reference image. Outputs the index data indicated.

In the MPEG series of video compression standards, 'motion estimation' refers to a process of obtaining the most similar portion of a previous frame in encoding a current frame. In general, the previous frame and the current frame are very similar in a video. In more detail, the motion estimation includes searching for a macroblock currently encoded in a current frame (hereinafter referred to as a 'current reference block') based on a macroblock having the same position as the current reference block in a previous frame. By comparing the number of macroblocks. Then, the macroblocks currently encoded in the current frame, that is, macroblocks similar to the previous reference block are found and the displacement is extracted. When the data blocks most similar to the current reference block in the previous frame are referred to as 'previous reference blocks', the 'displacement' may be a direction and a distance in which the previous reference block is moved to the current reference block. For example, if the previous reference block moved 1 pixel diagonally in the current frame, the movement corresponds to 'displacement'. Hereinafter, the displacement will be referred to as a 'motion vector'. The reason why the motion vector is important is that the amount of information to be transmitted can be reduced by transmitting the motion vector instead of the information about the current macroblock, that is, the current reference block of the current frame. That is, it is possible to calculate where the previous reference block of the previous frame exists in the current frame by using the motion vector, thus inferring the current reference block in the current frame.

In addition, a criterion of how similar the current reference block and the previous reference block is is a distortion, and a sum absolute difference (SAD) is generally used.

The macroblock mode determiner 54 determines and outputs an optimal encoding mode, that is, a 'block mode' of the macroblock currently input. If the macroblock mode determiner 54 determines the block mode, it means that the motion estimator 62 determines the block unit for estimating the motion. For example, if the macroblock mode determiner 54 determines the 16x16 macroblock as the block mode, the motion estimator 62 calculates a motion vector using a block having a size of 16x16.

When determining the block size, the macroblock mode determiner 54 according to the present embodiment obtains a rate-distortion (RD) value for a block having a block size of all sizes, and among them, the smallest rate- The block size having the distortion value is determined as the block mode.

The motion compensator 58 extracts and outputs image data corresponding to the currently input macroblock from the reference image stored in the frame memory 60 according to the motion data input from the motion estimator 62. The data output from the motion compensator 58 is input to the adder 56 to restore the decoded data input at the other end of the adder 56 to an actual image and transmit the decoded data to the frame memory 60.

The subtractor 42 receives image data in the reference image corresponding to the input macroblock from the motion compensator 58 when interframe prediction coding is performed on the macroblock input to the video encoder 40. A residual signal is output by performing a difference operation with the input macroblock. The output residual signal is again transformed and quantized by the DCT operator 44 and the quantizer 46 and entropy-encoded by the entropy encoder 48, and is output in the form of, for example, network abstraction layer (NAL) unit data. . The image data output from the entropy encoder 48 may be temporarily stored in a buffer (not shown) and then output to adjust the bit rate.

In FIG. 2, only the video encoder is described, but the fast motion mode determination method according to the present embodiment may be applied to a video decoder (not shown).

In the case of the video decoder, similar to the video encoding of FIG. 2, the video decoder includes a motion estimator 62, a macroblock mode determiner 54, and the like, and receives a bitstream. Determine their block mode and perform motion prediction. The method of determining the fast motion mode according to the video decoder is basically the same as the method of determining the fast motion mode of the video encoder. Instead of inputting an input video frame to the video encoder 40, a bitstream is input to the video decoder. 3 to 6 below are also applicable to a video decoder.

3 is a flowchart illustrating a method of determining a fast motion mode of a video encoder according to an embodiment of the present invention.

When an image frame is input to the video encoder 40 (S70), the motion estimator 62 performs motion prediction of 4 × 4, 8 × 8, and 16 × 16 blocks (S72). The video encoder 40 determines a candidate mode based on the motion prediction result performed in step S72 (S74), and calculates a rate-distortion value for each candidate mode (S76). The calculation of the rate-distortion value is preferably calculated by the motion estimator 62 of the video encoder 40. At this time, the motion estimator 62 calculates the rate-distortion value for block modes having sizes of 4 × 4, 8 × 8, and 16 × 16, but 8 for block modes having sizes of 8 × 16 or 16 × 8. Rate-distortion values are estimated using either x8 or 16x16. The motion estimator 62 multiplies the rate-distortion value for block mode with 8 × 8 by 2 or divides the rate-distortion value for block mode with 16 × 16 by 2 by 8 × 8. The rate-distortion value for a block mode having a size of 16 or 16 × 8 can be obtained. As a result, the video encoder 40 may reduce the complexity caused by calculating the rate-distortion value according to each block mode.

Also, the video encoder 40 determines an optimal block mode based on the result of calculating the rate-distortion value for each candidate mode in step S76 (S78). The video encoder 40 encodes an image frame using the block mode determined in step S78 (S80).

4 is a diagram illustrating basic processing block units of a video encoder according to an embodiment of the present invention.

In the present embodiment, after the video encoder 40 performs the motion prediction for the 8x8 and 16x16 candidate modes corresponding to the unit block, the candidate mode that may be additionally displayed together with the unit block is shown. When the candidate mode is determined as described above, the video encoder 40 calculates a rate-distortion for each of the determined candidate modes. In this case, the video encoder 40 according to the present embodiment uses the motion vector obtained through the motion prediction of the 8x8 or 16x16 candidate mode for the 16x8 or 8x16 candidate mode in which the motion prediction is not performed. For example, a 16 × 8 or 8 × 16 motion vector may be twice the value of the motion vector of the 8 × 8 candidate mode or half of the motion vector of the 16 × 16 candidate mode. have.

(A), (b), (c), and (d) of FIG. 4 illustrate candidate modes that a current reference block may have according to block modes of adjacent blocks. Referring to FIG. 4, when the block modes of adjacent blocks are 16 × 16 A, 16 × 8 B, D, and E as shown in (a), candidate modes C and E may be determined as 16 × 8. . In this case, the rate-distortion value in C, E, B, and D may be a value obtained by dividing the rate-distortion value obtained by A of 16 × 16 by 2.

In (b), the block modes adjacent to I and L corresponding to the current reference block are 16x16 F and 8x8 G, H, J and K. Accordingly, candidate modes I and L can be determined to be 16x8. In addition, the rate-distortion values in I and L may be divided by the rate-distortion value of F by 2 or by multiplying the rate-distortion values of G, H, J, and K by two.

(c) shows that 8x8 M, N, O, P, and 16x16 Q adjacent R and S have 8x16 candidate modes, and (d) 8x8 T, It is shown that U, V, W, and 16x16 X are adjacent to Y and Z have an 8x16 candidate mode. In (c), the rate-distortion values of M, N, O, and P of 8 × 8 are multiplied by two, or the rate-distortion values of Q of 16 × 16 are divided by two to obtain the rate-distortion values of R and S. Can be. Similarly, in (d), the rate-distortion value of 8, 8 T, U, V, and W is multiplied by 2, or the rate-distortion value of X of 16 × 16 is divided by 2 and the rate-distortion value of Y and Z is divided by 2. Can be obtained.

Table 1 below shows a new context when a candidate mode is determined by using information on a neighboring macroblock, which is a macroblock located adjacent to a macroblock to be encoded, and new context information is determined through the determined candidate mode. An example of determining information is shown in a table.

Conventional H.264 / AVC H.264 / AVC according to the present invention Condition 1 Condition 2 Mb_type Name of mb_type Mb_type Name of mb_type Name of mb_type 0 P_L0_16 × 16 0 P_L0_16 × 16 P_L0_16 × 16 One P_L0_L0_16 × 8 One P_L0_L0_16 × 8
or
P_L0_L0_8 × 16 P_8 × 8 2 P_L0_L0_8 × 16 2 P_8 × 8 P_8 × 8ref0 3 P_8 × 8 3 P_8 × ref0 4 inferred P_skip inferred P_skip

Referring to Table 1, the left side of Table 1 shows context information on Mb_type in conventional H.264 / AVC, and the right side shows context information on Mb_type in H.264 / AVC according to the present invention. In addition, Table 1 assumes that context information (Syntax) for Mb_type is context information that can be included when the current reference block is 8 × 8. At this time, the 'current reference block' refers to the number of blocks that can be processed at one time, that is, the basic processing block, in encoding the plurality of macroblocks constituting the input video frame.

Referring to Table 1, it is assumed that the video encoder 40 corresponds to Case 1 when the mode information on the adjacent macroblocks of the macroblocks to be encoded satisfies FIG. 4 and otherwise corresponds to Case 2. do. That is, in case 1, if (a), 1 of Mb_type becomes P_L0_L0_16 × 8, and in case of (b), P_L0_L0_8 × 16. Such context information may be defined as a protocol of an encoder or a decoder, and according to another embodiment of the present invention, the video encoder 40 may adaptively change the context information without additional information.

Table 2 below shows another example of determining a candidate mode using information of an adjacent macroblock according to an embodiment of the present invention and determining new context information when the new context information is determined through the determined candidate mode. Table shown. Table 2 also shows the case where the unit block is set to 8x8 as in Table 1.

Conventional H.264 / AVC H.264 / AVC according to the present invention Mb_type Name of mb_type Mb_type Name of mb_type 0 P_L0_16 × 16 0 P_L0_16 × 16 One P_L0_L0_16 × 8 One P_8 × 8 2 P_L0_L0_8 × 16 2 P_8 × 8ref0 3 P_8 × 8 4 P_8 × 8ref0 inferred P_skip inferred P_skip

Table 2 shows the context information for 16 × 8 or 8 × 16, which is an additional mode by using only the values for macroblocks 16 × 16 and 8 × 8, in the context information for Mb_type through flag bits.

5 is a diagram illustrating the structure of a bitstream according to FIG. 4 and Table 2 above.

Referring to FIG. 5, Case 1 illustrates a case in which Mb_type has a form suggested in Table 2 because it does not satisfy the above FIG. 4. In addition, Case 2 and Case 3 in FIG. 5 is a structure of a bitstream that may be indicated when the above FIG. 4 is satisfied.

If the mode information of the neighboring macroblock of the current macroblock, which is the macroblock to be encoded currently, satisfies FIG. 3, the video encoder 40 corresponds to Case 2 or Case 3 when Mb_type is 8 × 8. When the mode of the current macroblock becomes 8 × 8 through the rate-distortion calculation and the determination of the optimal block mode for the candidate mode according to the present embodiment, the value of the flag indicating the extended 8 × 8 mode is 0. Becomes In this case, in the case of the extended 8x8, after the candidate mode is determined as the extended 16x8 or 8x16, the flag indicating the extended 8x8 mode is 1.

In the present embodiment, when the Mb_type is 8x8 while the mode information of the neighboring macroblock adjacent to the macroblock to be encoded currently satisfies the case of FIG. 4, the bitstream structure of Case 2 or Case 3 may be provided. In the present embodiment, when the block mode of the current macroblock becomes 8 × 8 through the rate-distortion calculation and the determination of the optimal block mode for the candidate mode, the flag indicating the extended 8 × 8 mode becomes 0, and the block mode is At 16x8 or 8x16, the flag indicating the expanded 8x8 mode is one. At this time, the context information for the reference picture index and the motion vector also varies according to the value of the flag indicating the extended 8x8 mode.

FIG. 6 is a diagram illustrating a (2 × N) × N or 8 × (2 × N) block that can be generated using motion vectors and reference image information of an N × N macroblock according to an embodiment of the present invention. .

First, in FIG. 6A, a block indicated by a dotted line using information on macroblocks A 122 and B 124 having an N × N size (hereinafter, referred to as a “dotted block”) AB 120 When generating information about, the video encoder 40 generates a motion vector as follows.

1) The average vector of the motion vectors of the macroblocks A 122 and B 124 is regarded as the motion vector of the dotted block AB 120.

2) A value of one of the motion vectors of the macroblocks A 122 and B 124 is selected to determine the motion vector of the dotted block AB 120. In this case, the value determined as the motion vector of the AB may be arbitrarily selected or an optimized value through a method such as SAD.

3) The motion vector of the macroblocks A 122 and B 124 and the average vector of the macro blocks A 122 and B 124 are selected to determine the motion vector of the dotted block AB 120. . At this time, the motion vector of the dotted block AB 120 is arbitrarily selected or a function such as sum absolute difference (SAD), sum of absolute transformed difference (STAD), sum of squared difference (SSD), mean of absolute difference (MAD), or the like. It can be determined through the optimized value.

In FIG. 6B, when information on the dotted block CD 130 is generated using information on macroblocks C 130 and D 134 having an N × N size, the same as in FIG. 6A. The motion vector can be determined by

The video encoder 40 determines the average vector of the motion vectors of the macroblocks C 132 and D 134 as the motion vector of the dashed block 303 or one of the motion vectors of the C 132 and D 134. Select the value of to determine the motion vector of the dotted block CD 130, or select the motion vector of the C (132), D (134) or the average vector of the two macroblocks to move the motion of the dotted block CD (130) Can be determined by a vector. In this case, the optimized value may be arbitrarily selected by a person skilled in the art for carrying out the present invention or determined through a method such as SAD.

1 is a flowchart illustrating a method of determining a fast motion mode of a video encoder according to the prior art.

2 is a block diagram illustrating a schematic configuration of a video encoder according to an embodiment of the present invention.

3 is a flowchart illustrating a method of determining a fast motion mode of a video encoder according to an embodiment of the present invention.

4 is a diagram illustrating basic processing block units of a video encoder according to an embodiment of the present invention.

5 is a diagram illustrating the structure of a bitstream according to FIG. 4 and Table 2 above.

FIG. 6 is a diagram illustrating a (2 × N) × N or 8 × (2 × N) block that can be generated using motion vectors and reference image information of an N × N macroblock according to an embodiment of the present invention. .

Claims (16)

In the block mode determination method for video encoding, When a video to be encoded is input, performing motion prediction on a macroblock of a previously stored size for each macroblock unit; Selecting at least one of the block modes of the macroblock as a candidate mode according to the motion prediction; Calculating a rate-distortion value corresponding to each of the at least one candidate mode; And determining one of the at least one candidate mode as an optimal block mode according to the calculation result of the rate-distortion value. The method of claim 1, The process of performing motion prediction for each macroblock unit may include: The method of claim 1, wherein the motion prediction is performed on the macroblock having the macroblock unit of at least one of 4x8, 8x8 and 16x16. The method of claim 1, wherein the macroblock unit is A block mode determination method for encoding a video, characterized in that N × N. The method of claim 1, If the optimal block mode is determined, further comprising the step of performing the motion prediction for the next video input by using the optimal block mode. In the video encoder for determining the block mode, When a video to be encoded is input, motion prediction is performed for each macroblock of the video stored in units of macroblocks, and at least one candidate mode among the block modes of the macroblock is selected according to the motion prediction. A motion estimator for calculating a rate-distortion value corresponding to each of the at least one candidate mode; A macroblock mode determiner for selecting at least one candidate mode among block modes of the macroblock and determining one of the at least one candidate mode as an optimal block mode according to a result of calculating the rate-distortion value And a video encoder. The method of claim 5, wherein the motion estimator, And the motion prediction is performed on a macroblock in which the macroblock unit is at least one of 4 × 4, 8 × 8, and 16 × 16. The method of claim 5, wherein the macroblock unit is Video encoder, characterized in that N × N. The method of claim 5, wherein the motion estimator, And when the optimal block mode is determined, the motion prediction is performed on the next input video using the optimal block mode. In the block mode determination method for video decoding, When a bitstream to be decoded is input, performing motion prediction on a macroblock of a previously stored size in units of macroblocks; Selecting at least one of the block modes of the macroblock as a candidate mode according to the motion prediction; Calculating a rate-distortion value corresponding to each of the at least one candidate mode; And determining one of the at least one candidate mode as an optimal block mode according to a result of the calculation of the rate-distortion value. 10. The method of claim 9, The process of performing motion prediction for each macroblock unit may include: The method of claim 1, wherein the motion prediction is performed on a macroblock having at least one of 4x4, 8x8, and 16x16. The method of claim 9, wherein the macroblock unit is Block mode determination method for video decoding, characterized in that N × N. 10. The method of claim 9, And if the optimal block mode is determined, further performing the motion prediction on the input bitstream using the optimal block mode. In the video decoder to determine the block mode, When a bitstream to be decoded is input, motion prediction is performed for each macroblock of the bitstream of the bitstream in units of macroblocks, and at least one candidate mode among the block modes of the macroblock is determined according to the motion prediction. A motion estimator for calculating a rate-distortion value corresponding to each of the at least one candidate mode, when selected; A macroblock mode determiner for selecting at least one candidate mode among block modes of the macroblock and determining one of the at least one candidate mode as an optimal block mode according to a result of calculating the rate-distortion value And a video decoder. The method of claim 13, wherein the motion estimator, And the motion prediction is performed on a macroblock in which the macroblock unit is at least one of 4x4, 8x8, and 16x16. The method of claim 13, wherein the macroblock unit is A video decoder, characterized in that N × N. The method of claim 13, wherein the motion estimator, And when the optimal block mode is determined, the motion prediction is performed on the input bitstream using the optimal block mode.

Images