JP4445463B2 - Video re-encoding method and apparatus - Google Patents

Description

  The present invention relates to a moving image re-encoding method and apparatus that decodes moving image encoded data and re-encodes the encoded data.

  In recent years, with the spread of digital broadcasting and the like, a moving image re-encoding device called a transcoder that converts moving image encoded data into another moving image encoded data by changing the bit rate, image size, or encoding method The need for is increasing.

  In the moving image re-encoding device, high-efficiency encoding can be performed by obtaining a more suitable motion vector, as in a normal moving image encoding device. In order to obtain a suitable motion vector, the amount of calculation generally increases and a large memory access bandwidth is required for reading reference image data from the reference image memory. Therefore, the burden on the processor increases to obtain a suitable motion vector.

Therefore, Xiaoan Lu, et al, “Fast Mode Decision and Motion Estimation for H.264 with a Focus on MPEG-2 / H.264 Transcoding ”, ISCAS2005, pp.1246-1249 (Non-Patent Document 1) proposes a method for searching a narrow range centering on the motion vector of the original moving image encoded data and the predicted motion vector. .
Xiaoan Lu, et al, “Fast Mode Decision and Motion Estimation for H.264 with a Focus on MPEG-2 / H.264 Transcoding”, ISCAS2005, pp.1246-1249

  The method of Non-Patent Document 1 evaluates the position on the reference image indicated by the original motion vector and the predicted motion vector added to the original moving image encoded data, and determines which position is the search center. Has been decided. For this reason, it is necessary to read the reference image data corresponding to both the original motion vector and the predicted motion vector from the reference image memory, which increases the memory access bandwidth. When the bandwidth of memory access increases, the burden on the processor increases, which causes a problem that the encoding efficiency is lowered.

  It is an object of the present invention to provide a moving image re-encoding method and apparatus that enable efficient encoding by obtaining a suitable motion vector without increasing the memory access bandwidth.

  According to one aspect of the present invention, the second moving image encoding is performed by re-encoding the first moving image encoded data in an encoding mode selected from the intra mode and the motion compensation prediction mode for each encoding target block. In the moving image re-encoding device for obtaining data, a moving image decoding unit for decoding the first moving image encoded data to generate a decoded image signal and extracting a first motion vector; and A motion vector prediction unit for obtaining a predicted motion vector between the encoding target block and the reference image signal; and a position indicated by the first motion vector or a position indicated by the predicted motion vector on the reference image signal. A search center setting unit for setting a search center, and a motion compensation prediction of the encoding target block from a search range set in the reference image signal according to the search center. To provide a moving picture re-encoding apparatus comprising a motion vector search unit for searching for a second motion vector for.

  The search center setting unit, for example, (a) when a reference block having the same spatial position as the encoding target block in the reference image signal is encoded in the intra mode, indicates the position indicated by the first motion vector. (B) When the reference block is encoded in the motion compensated prediction mode, the position indicated by the prediction vector is set as the search center.

  The search center setting unit, for example, (a) when a reference block having the same spatial position as the encoding target block in the reference image signal is encoded in the intra mode, or when the reference block is in the motion compensation prediction mode. If the sum of absolute values of the difference motion vectors between the reference block and its surrounding blocks exceeds a threshold value, the position indicated by the first motion vector is set as the search center, and (b) the reference block is When encoded in the motion compensation prediction mode and the absolute value sum is equal to or less than the threshold, the position indicated by the prediction vector is set as the search center.

  According to the present invention, the search center is set at the position on the reference image pointed to by either the original motion vector or the predicted motion vector, and the motion vector search is performed without increasing the memory access bandwidth. It is possible to obtain a suitable motion vector and perform efficient video re-encoding.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
(First embodiment)
As shown in FIG. 1, a moving image re-encoding device according to an embodiment of the present invention includes a moving image decoding unit 100 and a moving image encoding unit 200. The moving image signal decoding unit 100 is, for example, MPEG-2, MPEG-4 or H.264. H.264 / AVC and other moving image coding schemes are used to decode moving image encoded data (input code string) 10 obtained by encoding a moving image signal to generate a decoded image signal 11 and to extract side information 15 And output. The side information 15 is information obtained in the process in which the moving image decoding unit 100 encodes a moving image signal, and includes, for example, information on a prediction mode and a motion vector.

  The moving image encoding unit 200 converts the decoded image signal 11 from the moving image decoding unit 100 into the same or different moving image encoding method as that of the input code sequence 10, for example, MPEG-2, MPEG-4, or H.264. Encoding is performed by a moving image encoding method such as H.264 / AVC, and moving image encoded data (output code string) 14 is obtained. When the output code string 14 has the same moving picture coding method as that of the input code string 10, the moving picture coding unit 200 performs re-coding for conversion of the bit rate and the image size. The moving image coding unit 200 includes a decoded image memory 201, a subtracter 202, an orthogonal transform / quantization unit 203, an entropy coding unit 204, a motion compensation prediction unit 205, an inverse quantization / inverse orthogonal transform unit 206, an adder 207, and A reference image memory 208 is included.

  In the decoded image memory 201, the decoded image signal 11 to be encoded by the moving image encoding unit 200 is temporarily stored. The motion compensation prediction unit 205 obtains an optimal motion vector using the side information 15 from the video decoding unit 100, the decoded image signal read from the decoded image memory 11 and the reference image signal read from the reference image memory 208, and Motion compensation prediction is performed on the reference image signal using the optimum motion vector. That is, the motion compensation prediction unit 205 reads the reference image signal of the region indicated by the optimal motion vector from the reference image memory 208, thereby generating the predicted image signal 17 based on the motion compensation prediction. The motion compensation prediction unit 205 outputs motion vector information 16 indicating the optimal motion vector. The motion compensation prediction unit 205 will be described in detail later.

  The subtractor 202 performs subtraction between the decoded image signal read from the decoded image memory 201 and the predicted image signal 17 read from the reference image memory 208 by the motion compensation prediction unit 205 to obtain the difference between them, that is, the prediction residual signal 12. Generate. The prediction residual signal 12 is input to the orthogonal transform / quantization unit 203.

  The orthogonal transform / quantization unit 203 performs orthogonal transform and quantization on the prediction residual signal 12 to generate a quantized orthogonal transform coefficient 13. The quantized orthogonal transform coefficient 113 is input to the entropy encoding unit 104 and the inverse quantization / inverse orthogonal transform unit 206.

  The entropy encoding unit 104 performs entropy encoding on the quantized orthogonal transform coefficient 13 and the motion vector information 16 to generate an output code string 14.

  The inverse quantization / inverse orthogonal transform unit 206 performs inverse quantization and inverse orthogonal transform on the quantized orthogonal transform coefficient 13 from the orthogonal transform / quantization unit 203 to reproduce a prediction residual signal.

  The adder 207 adds the reproduced prediction residual signal and the predicted image signal 17 from the reference image memory 208 to generate a locally decoded image signal. The locally decoded image signal is stored in the reference image memory 208 as a reference image signal for the next motion compensation prediction.

  Next, the motion compensation prediction unit 205 will be described in detail with reference to FIG. The motion compensation prediction unit 205 illustrated in FIG. 2 includes a motion vector prediction unit 301, a search center setting unit 302, and a motion vector search unit 303, and predicts a motion vector used for motion compensation prediction of the decoded image signal 11 (predicted motion vector). Generation), setting of a search center used for motion vector search and search for an optimal motion vector.

  The motion vector prediction unit 301 is a block on the reference image signal in which the peripheral block of the encoding target block on the decoded image signal and the spatial position (xy coordinate position on the image plane) are the same as the encoding target block. A predicted motion vector is obtained using motion vector information (hereinafter referred to as a reference block having the same spatial position as that of the encoding target block).

  The search center setting unit 302 includes a position on the reference image signal indicated by the motion vector (referred to as an original motion vector) in the side information 15 and a position on the reference image signal indicated by the prediction motion vector obtained by the motion vector prediction unit 301. Is set as a search center used for motion vector search in the motion vector search unit 303.

  The motion vector search unit 303 receives the decoded image signal data of the encoding target block from the decoded image memory 201 and sets the motion vector search range set around the search center obtained by the search center setting unit 302 from the reference image memory 208. The reference image signal data is received and a motion vector search is performed to obtain an optimal motion vector. That is, among the reference image signals in the motion vector search range, an area most similar to the image signal of the encoding target block is obtained by known block matching, and a motion vector between the area and the encoding target block is determined as the optimum motion vector. And

Next, a detailed processing procedure of the motion compensation prediction unit 205 shown in FIG. 2 will be described using the flowchart of FIG.
First, the decoded image signal of the encoding target block is read from the decoded image memory 201 (step S101). Next, the motion vector search center is set using the side information 15 and the predicted motion vector (step S102). More specifically, the search center is selected by selecting either the position on the reference image signal indicated by the original motion vector included in the side information 15 or the position on the reference image signal indicated by the predicted motion vector as the search center. Set up.

  Next, the reference image signal data in the motion vector search range set around the search center obtained in step S102 is read from the reference image memory 208 (step S103), and the motion vector is searched for this range (step S104). ). Note that the process of step S101 may be performed anywhere before step S104, and may be performed simultaneously with the process of other steps depending on circumstances.

  As described above, in the first embodiment, one of the position indicated by the original motion vector and the position indicated by the predicted motion vector on the reference image signal is set as the search center for the motion vector. Only the reference image signal data in the motion vector search range set around the search center from 208 may be read. Therefore, the reference image corresponding to both the original motion vector and the predicted motion vector from the reference image memory in order to evaluate the positions indicated by the original motion vector and the predicted motion vector on the reference image signal as disclosed in Non-Patent Document 1, respectively. Compared with the method of reading the data, the memory access bandwidth when reading the data of the reference image signal from the reference image memory 208 is greatly reduced. Therefore, the processing load of a processor such as a CPU is reduced, and as a result, it is possible to contribute to improvement of encoding efficiency.

(Second Embodiment)
Next, a second specific example of the motion compensation prediction unit 205 will be described with reference to FIG. In the motion compensation prediction unit 205 illustrated in FIG. 4, a path from the entropy encoding unit 204 to the search center setting unit 302 is added to the motion compensation prediction unit of the first specific example illustrated in FIG. 2. Encoding result information from the entropy encoding unit 104 is output to this path.

  Similar to the first embodiment, the motion vector prediction unit 301 includes information on motion vectors of the peripheral blocks of the block to be encoded on the decoded image signal and the reference block having the same spatial position as the block to be encoded on the reference image signal. Is used to obtain a predicted motion vector.

  The search center setting unit 302 encodes information on the encoding result of the reference block having the same spatial position as the encoding target block from the entropy encoding unit 204 (for example, the difference between the generated code amount of the reference block and the motion vector of the reference block) The position indicated by the original motion vector in the side information 15 on the reference image and the position indicated by the predicted motion vector obtained by the motion vector prediction unit 301 are used in the motion vector search based on the sum of absolute values). Set to search center.

  The motion vector search unit 303 receives the decoded image signal data of the encoding target block from the decoded image memory 201 and sets the motion vector search range set around the search center obtained by the search center setting unit 302 from the reference image memory 208. The reference image signal data is received and a motion vector search is performed to obtain an optimal motion vector.

  Next, the processing procedure of the motion compensation prediction unit 205 shown in FIG. 4 will be described using the flowchart of FIG.

  First, the decoded image signal of the encoding target block is read from the decoded image memory 201 (step S101). Next, information indicating the encoding result of the reference block having the same spatial position as the encoding target block on the reference image is read from the entropy encoding unit 204 (step S111). The information indicating the coding result of the reference block is, for example, information indicating the coding mode of the reference block or motion vector information. As a reference block having the same spatial position as the encoding target block, for example, H.264 as shown in FIG. A co-located block in H.264 / AVC is exemplified.

  FIG. 2 shows a motion-compensated prediction structure in temporal direct mode in H.264 / AVC, where Current is a current encoding target image to be encoded, and List 0 Reference and List 1 Reference are temporal encoding target images. Reference images before and after, current block is a block to be encoded, and co-located block is a reference block having the same spatial position as the block to be encoded in the reference image List 1 Reference.

  In FIG. 7, temporal direct MV0 is a motion vector between the current block to be encoded and the reference image List 0 Reference, temporal direct MV1 is a motion vector between the current block to be encoded and the reference image List 1 Reference, The co-located MV represents a motion vector between the co-located block and the List 0 Reference on the reference image List 1 Reference.

  Next, the motion vector search center is set using the side information 15 and the predicted motion vector (step S102). Specifically, the search center is selected by selecting either one of the position on the reference image signal indicated by the original motion vector included in the side information 15 and the position on the reference image signal indicated by the predicted motion vector as the search center. Set.

  Next, the reference image signal data in the motion vector search range set around the search center obtained in step S102 is read from the reference image memory 208 (step S103), and the motion vector is searched (step S104). Note that the process in step S101 may be performed anywhere before step S104, and in some cases, the process in step S101 may be performed simultaneously with the processes in other steps.

  As described above, in the second embodiment, only the reference image signal data in the motion vector search range set around the search center may be read from the reference image memory 208 when searching for a motion vector, as in the first embodiment. Therefore, the memory access bandwidth when reading the reference image signal data is greatly reduced.

  Furthermore, in the second embodiment, as a motion vector predictor, a motion vector of a type that does not require the periphery of the encoding target block, for example, H.264, etc. When the temporal direct mode motion vector in H.264 / AVC (temporal direct MV0, temporal direct MV1 in FIG. 7) is used, information necessary for determining the search center is obtained only from the side information 15 and the reference image coding result. Therefore, the processing of the motion compensation prediction unit 205 can be easily performed at high speed by pipelining or parallelizing the processing.

(Third embodiment)
Next, a third embodiment of the present invention will be described. In this embodiment, the configuration and schematic processing procedure of the motion compensation prediction unit 205 are the same as those of the second embodiment, and the processing of the search center setting unit 302 is different from that of the second embodiment. Hereinafter, a specific processing procedure of the search center setting unit 302 in the present embodiment will be described using the flowchart of FIG.

  First, information indicating the encoding mode is received from the entropy encoding unit 204 indicating information indicating the encoding result of the reference block having the same spatial position as the block to be encoded on the reference image signal, and the reference block is encoded in the intra mode. Whether the reference block is an intra block (step S201). If the reference block is an intra block, the motion vector search center is set to the position indicated by the original motion vector in the side information 15 (step S202).

  On the other hand, when the reference block having the same spatial position as the encoding target block is not an intra block, that is, when the reference block is an inter block (also referred to as a motion compensated prediction block) encoded in a motion compensation prediction mode that is an inter mode. Then, the motion vector search center is set to the position indicated by the motion vector predictor obtained by the motion vector predictor 301 (step S203).

  Thus, in the third embodiment, it is possible to select a more appropriate motion vector search center by using the encoding result of the reference block having the same spatial position as the target block in the reference image signal. . Therefore, efficient coding with a suitable motion vector is possible even with a small memory access bandwidth.

  Now, the encoding format of the moving image encoding unit 200 is H.264. In H.264 / AVC, a case where an encoding target image is a B picture and a temporal direct mode motion vector is used as a predicted motion vector is considered. The motion vector in the temporal direct mode becomes a zero vector when the co-located block is an intra block. In a video with actual motion, a suitable motion vector cannot be obtained even if the position indicated by the zero vector is the search center.

  Therefore, in such a case, that is, when the reference block (co-located block) is determined to be an intra block in step S201, the position indicated by the original motion vector included in the side information 15 in the input code string 10 is the search center. Is desirable. Thus, after setting the search center of the motion vector in step S102, it is possible to search for a suitable motion vector by performing the processing of steps S103 to S104 and perform highly efficient encoding.

(Fourth embodiment)
Next, a fourth embodiment of the present invention will be described. In this embodiment, the configuration and schematic processing procedure of the motion compensation prediction unit 205 are the same as those of the second embodiment, and the processing of the search center setting unit 302 is different from that of the second embodiment. Hereinafter, a specific processing procedure of the search center setting unit 302 in the present embodiment will be described using the flowchart of FIG.

  First, it is checked whether or not a reference block having the same spatial position as the encoding target block from the entropy encoding unit 204 is an intra block (step S201). If the reference block having the same spatial position as the encoding target block is an intra block, the motion vector search center is set to the position indicated by the original motion vector in the side information 15 (step S202).

  On the other hand, if the reference block having the same spatial position as the encoding target block is not an intra block, that is, if it is an inter block, the absolute value sum of the difference motion vectors included in the encoding result of the reference block from the entropy encoding unit 204 It is checked whether or not a certain threshold value or less (step S204). The difference motion vector is a difference vector between the motion vector of the reference block and the predicted motion vector of the surrounding blocks.

  When the sum of absolute values of the difference motion vectors between the reference block and its surrounding blocks is larger than the threshold, the motion vector search center is set to the position indicated by the original motion vector in the side information 15 (step S202). When the absolute value sum of the difference motion vectors of the reference block is equal to or smaller than the threshold value, the motion vector search center is set to the position indicated by the motion vector predictor obtained by the motion vector prediction unit 301 (step S203).

  As described above, in the fourth embodiment, as in the third embodiment, by using the encoding result of the reference block whose spatial position is the same as that of the target block in the reference image signal, the center of a more appropriate motion vector search is used. Can be selected. Therefore, efficient coding with a suitable motion vector is possible even with a small memory access bandwidth.

  The encoding format of the moving image encoding unit 200 is H.264. In H.264 / AVC, when an encoding target image is a B picture and a motion vector in temporal direct mode is used as a predicted motion vector, when a reference block (co-located block) is determined to be an intra block, As in the third embodiment, a more suitable motion vector can be searched by setting the position indicated by the original motion vector included in the side information 15 in the input code string 10 as the search center.

  As shown in FIG. 7, the temporal direct mode motion vectors (temporal direct MV0, temporal direct MV1) are obtained by internally dividing (scaling) the motion vector (co-located MV) of the co-located block. This is particularly effective when the motion is constant over time and the block to be encoded is moving in the same manner as the surrounding blocks. Furthermore, the fact that the sum of the absolute value of the difference motion vector between the motion vector of the co-located block and the predicted motion vector of the surrounding block is small indicates that the co-located block is moving similarly to the surrounding block. Therefore, it can be said that the motion vector in the temporal direct mode is highly reliable.

  Therefore, when the absolute value sum of the difference motion vectors between the reference block and the surrounding blocks is equal to or less than the threshold value in step S204, the position on the reference image indicated by the prediction motion vector is determined using the temporal direct mode motion vector as the prediction motion vector. By using as a motion vector search center, it is possible to search for a more suitable motion vector and perform highly efficient encoding.

  On the other hand, when the absolute value sum of the difference vector between the motion vector of the co-located block and the prediction vector of the surrounding block is large, it indicates that the co-located block is moving differently from the surrounding block. It can be said that the reliability of the motion vector in the temporal direct mode is low. Therefore, when the absolute value sum of the difference motion vectors between the reference block and the surrounding blocks exceeds the threshold in step S204, the position indicated by the original motion vector included in the side information 15 in the input code string 10 is set as the search center. By doing so, a suitable motion vector can be searched.

  In step S204, the threshold value may be a fixed value or may be changed. When changing the threshold value, for example, the threshold value is decreased as the generated code amount of the reference block having the same spatial position as the encoding target block from the entropy encoding unit 204 increases. That is, when the generated code amount of the reference block is large, the position indicated by the original motion vector included in the side information 15 in the input code string 10 is easily selected as the search center. This is because the amount of generated code is large because the amount of code used to encode a motion vector also has a margin, and as a result, a motion vector with a motion vector generated around a predicted motion vector with a small amount of generated motion vector as a search range. The need to search for is reduced.

  In this way, it is possible to set a more appropriate search center by changing the threshold value in step S204 depending on the generated code amount of the reference block having the same spatial position as the encoding target block, and the memory access bandwidth is small However, it is possible to search for a more suitable motion vector and perform highly efficient encoding.

  As described above, the encoding format of the moving image encoding unit 200 is H.264. In H.264 / AVC, when the encoding target image is a B picture and a temporal direct mode motion vector is used as a predicted motion vector, the temporal direct mode motion vector is a zero vector when the co-located block is an intra block. become. In an image with actual motion, a suitable motion vector cannot be obtained even if the position indicated by the zero vector is the search center, but the zero vector may be a more suitable motion vector regardless of the motion of the image.

  Therefore, in addition to the motion vector search from the search range around the search center set in step S202 or step S203, the motion vector is also searched for the position of the zero vector in the reference image signal. The position at which the reference image signal is read from the reference image memory 208 is determined in correspondence with the position of the zero vector and is in a narrow range, so the influence of the memory access on the bandwidth is small. Therefore, by adding the position of the zero vector to the search range, it is possible to search for a more suitable motion vector and perform highly efficient encoding without increasing the memory access bandwidth.

  The moving image re-encoding process described in each embodiment described above may be performed by dedicated hardware, or the moving image re-encoding process (FIG. 3, FIG. (See FIGS. 5, 6, and 8). Further, a program for causing the computer to execute such processing may be provided to the user via a communication line such as the Internet, or the program may be provided as a CD-ROM (Compact Disc-Read Only Memory). It may be recorded on a computer-readable recording medium and provided to the user.

  That is, according to the present invention, the second moving image encoded data is obtained by re-encoding the first moving image encoded data for each encoding target block in the encoding mode selected from the intra mode and the motion compensation prediction mode. A moving picture decoding unit for causing a computer to function as a moving picture re-encoding device, generating a decoded picture signal by decoding the first moving picture coded data and extracting a first motion vector A motion vector prediction unit that obtains a predicted motion vector between a coding target block of the decoded image signal and a reference image signal, a position indicated by the first motion vector on the reference image signal, and the predicted motion vector A search center setting unit that sets a search center at one of the indicated positions, and a search set in the reference image signal according to the search center A program for causing the computer to function as a moving image re-encoding device comprising a motion vector search unit for searching for a second motion vector for motion compensated prediction of the encoding target block from a box, or recording the program It is also possible to provide such a recording medium.

The block diagram which shows the structure of the moving image re-encoding apparatus which concerns on one Embodiment of this invention. The block diagram which shows the 1st specific example of the motion compensation estimation part in FIG. The flowchart which shows the process sequence of the motion compensation estimation part of FIG. The block diagram which shows the 2nd specific example of the motion compensation estimation part in FIG. The flowchart which shows the process sequence of the motion compensation estimation part of FIG. The flowchart which shows the 1st specific process sequence of the search center setting part in FIG. H. The figure which shows the motion compensation prediction structure of the temporal direct (temporal direct) mode in H.264 / AVC The flowchart which shows the 2nd specific process sequence of the search center setting part in FIG.

Explanation of symbols

10: Input code string (first moving image encoded data)
DESCRIPTION OF SYMBOLS 11 ... Decoded image signal 12 ... Prediction residual signal 13 ... Quantization orthogonal transformation coefficient 14 ... Output code sequence (2nd moving image coding data)
DESCRIPTION OF SYMBOLS 15 ... Side information 16 ... Motion vector information 17 ... Reference image signal 100 ... Moving image signal decoding part 200 ... Moving image encoding part 201 ... Decoded image memory 202 ... Subtractor 203 ... Orthogonal transformation / quantization part 204 ... Entropy encoding Reference numeral 205: Motion compensation prediction unit 206: Inverse quantization / inverse orthogonal transformation unit 207: Adder 208 ... Reference image memory 301 ... Motion vector prediction unit 302 ... Search center setting unit 303 ... Motion vector search unit

Claims (9)

A moving image re-encoding device that re-encodes the first moving image encoded data in an encoding mode selected from the intra mode and the motion compensated prediction mode for each encoding target block to obtain second moving image encoded data In
A moving image decoding unit that decodes the first moving image encoded data to generate a decoded image signal and extracts a first motion vector;
A motion vector prediction unit for obtaining a prediction motion vector between a block to be encoded of the decoded image signal and a reference image signal that has already been re-encoded ;
The second moving image encoded data of the reference image signal is referred to, and the code in the reference image signal is referred to.
A reference block with the same spatial position as the current block is re-encoded in intra mode
The position indicated by the first motion vector is set as the search center, and the reference block
Is re-encoded in motion compensated prediction mode, the position indicated by the prediction vector
A search center setting unit for setting the search center;
A motion image re-encoding device comprising: a motion vector search unit that searches for a second motion vector for motion compensation prediction of the encoding target block from a search range set in the reference image signal according to the search center.   In the search center setting unit, the reference block is re-encoded in a motion compensation prediction mode.
The absolute value sum of the difference motion vectors between the reference block and its surrounding blocks is calculated.
When the sum of absolute values of the difference vectors exceeds a threshold, the position indicated by the first motion vector
Is set at the search center, and when the sum of absolute values of the difference vectors is less than or equal to a threshold value, the prediction vector is set.
The moving image re-encoding device according to claim 1, wherein a position indicated by a kuttle is set at the search center. The moving image re-encoding device according to claim 2, wherein the search center setting unit decreases the threshold value as the generated code amount of the reference block in the moving image encoding unit increases. The motion vector search unit, any one of claims 1 to 3 to search for the second motion vector from the position of the zero vector of the reference search range and in the reference image signal is set in the image signal according to the search center The moving image re-encoding device according to claim 1. The first moving image coded data, to claim 1 which is encoded by MPEG-2
5. The moving image re-encoding device according to any one of 4 above. The moving image re-encoding device is the H.264 standard. 264 / AVC by the decoded image signal coded by the moving picture re-encoding apparatus according to any one of the second video encoding claim configured data to obtain 1 to 5. The moving image re-encoding device is the H.264 standard. 264 / the decoded image signal is encoded by AVC is configured to obtain the second encoded video data stream, according to claim 1, the motion vector of the motion vector prediction unit time direct mode and the predicted motion vector The moving image re-encoding device according to any one of claims 1 to 6 .   Intra mode and motion compensated prediction of first moving image encoded data for each encoding target block
The second moving image encoded data is obtained by re-encoding according to the encoding mode selected from the modes.
In the moving image re-encoding method,
  The first moving image encoded data is decoded to generate a decoded image signal, and the first motion vector is generated.
A video decoding step for extracting a toll;
  A block to be encoded of the decoded image signal and a reference image signal that has already been re-encoded.
A motion vector prediction step for obtaining a motion vector predictor between,
  The second moving image encoded data of the reference image signal is referred to, and the code in the reference image signal is referred to.
A reference block with the same spatial position as the current block is re-encoded in intra mode
The position indicated by the first motion vector is set as the search center, and the reference block
Is re-encoded in motion compensated prediction mode, the position indicated by the prediction vector
A search center setting step for setting the search center;
  From the search range set in the reference image signal according to the search center,
A motion vector search step for searching for a second motion vector for motion compensation prediction of the lock;
A moving image re-encoding method comprising:   Intra mode and motion compensated prediction of first moving image encoded data for each encoding target block
The second moving image encoded data is obtained by re-encoding according to the encoding mode selected from the modes.
In the video re-encoding program,
  The first moving image encoded data is decoded to generate a decoded image signal, and the first motion vector is generated.
A video decoding function that extracts
  A block to be encoded of the decoded image signal and a reference image signal that has already been re-encoded.
A motion vector prediction function for obtaining a predicted motion vector between
  The second moving image encoded data of the reference image signal is referred to, and the code in the reference image signal is
A reference block with the same spatial position as the current block is re-encoded in intra mode
The position indicated by the first motion vector is set as the search center, and the reference block
Is re-encoded in motion compensated prediction mode, the position indicated by the prediction vector
Search center setting function to set the search center,
  From the search range set in the reference image signal according to the search center, the encoding target block is set.
A motion vector search function for searching for a second motion vector for motion compensation prediction of the lock;
A moving image re-encoding program for causing a computer to realize the above.

Images