JP2012231295A - Encoder, encoding method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable encode processing to be executed in parallel by using a general encoder so as not to destroy a virtual decoder buffer.SOLUTION: Input image data is divided in picture units to separate it into the same number of image data as the number of encoders. Then, the separated plural image data respectively are encoded in plural encoders by intra-screen compression. After that, the plural image data which have respectively been encoded by the plural encoders are combined to generate combined image data, in which case a header containing frame_num information is detected from the combined image data thus combined on the basis of separation information, and frame_num of the detected header is corrected to generate final output image data. At this time, the level of a virtual buffer for the combined image data is also corrected to ensure that the virtual buffer will not be destroyed.

Description

  The present invention relates to an encoding device, an encoding method, and a program, and more particularly, to a technique suitable for use in performing an encoding process using a plurality of encoders.

  MPEG and H.264 In the moving image compression technology represented by H.264, encoding is performed by reducing the amount of data using temporal redundancy or spatial redundancy of moving images. Here, temporal redundancy refers to the similarity of the current image to be compressed (hereinafter referred to as the current picture) with the previous or subsequent image in terms of time. The amount of data is compressed by encoding the difference between. Such a method of encoding using temporal redundancy is generally called inter-screen predictive encoding.

  On the other hand, spatial redundancy is redundancy within one natural image, and compression is performed using this redundancy. Also, H.264, which is a newer technology than MPEG. In H.264, the compression efficiency is further increased by utilizing the similarity in the current picture. Such a method of encoding using spatial redundancy is generally called intra-screen predictive encoding.

  In inter-screen predictive coding, an original image (reference image) is necessary to obtain a difference between pixels, and it cannot be decoded only with an image that has undergone inter-screen predictive coding. On the other hand, in the intra prediction encoding, a difference between pixels in the screen is taken, so that an image subjected to intra prediction encoding can be decoded alone.

  In addition, the compression efficiency is generally higher in the inter-frame prediction encoding, and an image that has been inter-frame prediction encoded is immediately decoded when it becomes a reference image of the next image. However, there is a tendency that errors are accumulated and image quality is deteriorated by sequentially using images subjected to inter-picture prediction encoding as reference images. Therefore, in order to prevent the accumulation of coding errors, images to be subjected to intra prediction prediction are periodically inserted, and the insertion unit of images to be intra prediction prediction is generally called GOP (Group of Pictures). ing. This GOP unit is generally set to a certain time (about 0.5 seconds) or more in consideration of compression efficiency.

  Also, in order to make editing easier or to suppress the effects of the accumulation of errors as described above, encoding is performed with a low correlation (or no correlation) between GOPs. There are many. Therefore, when parallel processing is performed using a plurality of encoders in order to improve the processing speed, it is performed in units of GOP with low correlation before and after. At that time, the problem is the amount of data that must be retained internally and the method of connecting compressed data generated by parallel processing. In order to solve these problems, there is a method in which encoding is performed only by intra prediction encoding as in JPEG or as described in Patent Document 1.

JP 2008-31824 A

  However, in the prior art disclosed in Patent Document 1, H.264 is used. In order to perform parallel processing so that the virtual decoder buffer defined in H.264 does not fail, it is necessary to use a plurality of special encoders based on parallel processing. Therefore, there is a problem that the existing encoder cannot be used and the cost is increased.

  In view of the above-described problems, an object of the present invention is to enable an encoding process to be performed in parallel using a general encoder so as not to cause a virtual decoder buffer to fail.

  An encoding apparatus according to the present invention divides input image data in units of pictures and distributes the image data to a plurality of image data, and encodes the plurality of image data distributed by the distribution unit by intra-screen encoding. A plurality of encoding means; a combining means for combining the plurality of image data encoded by the plurality of encoding means to generate combined image data; and the combination based on information relating to distribution by the distribution means The plurality of image data before being combined by the means, or a header detecting means for detecting a predetermined header from the combined image data, and correcting the predetermined header detected by the header detecting means and correcting the predetermined header from the combined image data Header correcting means for generating output image data.

  According to the present invention, it is possible to perform parallel processing without using a special encoder and without destroying the virtual decoder buffer.

It is a block diagram which shows the schematic structural example of the encoding apparatus which concerns on 1st Embodiment. It is a figure explaining the outline | summary of the operation | movement of the encoding process in 1st Embodiment. It is a figure explaining the outline | summary of the operation | movement of the encoding process in 2nd Embodiment. It is a figure explaining the outline | summary of the operation | movement of the encoding process in 3rd Embodiment. It is a block diagram which shows the schematic structural example of the encoding apparatus which concerns on 4th Embodiment. H. 2 is a diagram illustrating a relationship between a level defined in H.264 and a buffer size of a stream. It is a block diagram which shows the schematic structural example of the encoding apparatus which concerns on 5th Embodiment. It is a figure explaining the outline | summary of the operation | movement of the encoding process in 5th Embodiment.

(First embodiment)
Hereinafter, an example of improving the performance of an encoding process using a plurality of encoders in parallel according to the first embodiment of the present invention will be described with reference to the drawings. In this embodiment, a parallel operation of two encoders will be described. The encoding method is H.264. It is based on H.264, and it is assumed that intra-screen coding that can be independently coded independently of previous and subsequent images is used.

FIG. 1 is a block diagram illustrating a schematic configuration example of an encoding apparatus 100 according to the present embodiment.
In FIG. 1, an input terminal 101 inputs an input parameter that instructs the compression condition of the encoded stream from the outside. The encoding parameter setting unit 102 sets an input parameter input from the input terminal 101.

  The rate control unit 103 controls the code amounts of a first encoder 106 and a second encoder 107 described later. The memory 104 is a memory that stores input image data to be encoded. The image distribution unit 105 divides the input image data stored in the memory 104 in units of pictures and distributes the input image data to the first encoder 106 or the second encoder 107 at a predetermined timing.

  The encoding unit 108 includes a plurality of encoders that generate an encoded stream. In the present embodiment, the encoding unit 108 includes a first encoder 106 and a second encoder 107. The stream combining unit 109 is for combining two streams output from the first encoder 106 or the second encoder 107. The header detection unit 110 is for detecting a predetermined header from the stream (combined image data) combined by the stream combining unit 109. The header correction unit 111 is for rewriting the detected header to a predetermined value. The output terminal 112 is for outputting the finally obtained composite stream (output image data) to the outside.

FIG. 2 is a diagram for explaining the outline of the operation of the encoding process in the present embodiment.
An input moving image 200 shown in FIG. 2 is expressed in units of pictures with the horizontal axis as a time axis, and a dummy picture 201 is added to the head of the input moving image 200. A stream 202 indicates a stream generated by the encoding process of the first encoder 106. A stream 203 indicates a stream generated by the encoding process of the second encoder 107. The dummy picture 204 is an encoded picture in which the dummy picture 201 is encoded by the second encoder 107. A combined stream 205 is a combination of the generated streams 202 and 203.

Next, the flow of processing in this embodiment will be described with reference to FIG.
First, an input parameter serving as a composite stream generation condition is input from the input terminal 101. Here, the input parameters are specifically parameters that need to be instructed from the outside, such as setting the image size and the target bit rate. This input parameter is set when the encoding process is started. For example, when used with a camera or the like, the input parameters are changed according to the shooting mode of the camera and the performance of the recording medium.

  Next, the encoding parameter setting unit 102 converts the instruction content of the input parameter into a format that can be set to the first encoder 106 and the second encoder 107 in the encoding unit 108 and transmits the converted format. Of the input parameters, the final target bit rate is set in the rate control unit 103.

  When the setting for encoding is completed by the above-described procedure, the input image data stored in the memory 104 is input to the image distribution unit 105. The image distribution unit 105 distributes the input image data that has been input to the first encoder 106 and the second encoder 107 alternately for each picture, and outputs the result. Further, the image distribution unit 105 sends information indicating how the input image data has been distributed to the header detection unit 110. The information relating to the distribution is information indicating, for example, which of the first encoder 106 and the second encoder 107 each picture is allocated to.

  The first encoder 106 and the second encoder 107 are input with data in which pictures are skipped one by one in time. The first encoder 106 and the second encoder 107 recognize these data as continuous moving images and perform intra-screen encoding processing according to the encoding parameters. Note that the first encoder 106 and the second encoder 107 do not need to have a special configuration for parallel processing, and each output is independent of H.264. H.264.

  The stream combining unit 109 combines the two encoded streams output from the encoding unit 108 alternately in units of pictures and combines them into one stream. In this embodiment, since only intra-picture encoding is used, the stream combination order is the same as the input order. Then, the stream combining unit 109 sends information on the bit amount of the combined stream to the rate control unit 103. The rate control unit 103 controls the first encoder 106 and the second encoder 107 from the target bit rate specified by the input parameter and the actually generated bit amount, and approaches the desired bit rate. Like that.

  Note that the combined stream at this stage is simply combined. H.264 standard is not satisfied. Therefore, in the present embodiment, the header detection unit 110 detects the position of a predetermined header in the combined stream combined by the stream combining unit 109 based on the image distribution information. In this embodiment, the header is detected for the stream after combining, but it may be performed for each stream before combining. The header information detected by the header detection unit 110 is sent to the header correction unit 111 together with the combined stream, and the header correction unit 111 rewrites a specific header part. The header correction unit 111 corrects the header information in the combined stream in which a contradiction has occurred due to the combination, so H.264 is generated.

Next, a detailed operation will be described with reference to FIG.
First, the image distribution unit 105 adds the dummy picture 201 to the head of the input moving image stored in the memory 104 to generate the input moving image 200 shown in FIG. Then, the image distribution unit 105 distributes even-numbered pictures to the first encoder 106 and distributes odd-numbered pictures to the second encoder 107 as shown in FIG.

  Each picture is encoded by the first encoder 106 and the second encoder 107, but since it is processed in parallel as shown in FIG. Double the processing time. That is, even with an encoder that has about half the processing capability of the desired picture rate, encoding can be realized at the desired picture rate.

  H. In H.264, a special picture called an IDR (Instantaneous Decoding Refresh) picture (hereinafter referred to as IDR) is always required at the beginning of encoding. This IDR is a special picture among the I pictures in the intra-picture encoding, and the head of the picture processed by almost all the encoders is the IDR.

  The IDR can appear at an arbitrary position even in the middle of the stream, and an identifier between IDRs is added to the header. When an IDR appears in one encoder, identifiers are always added so that the identifiers are different. However, when a plurality of encoders are used, identifiers may overlap. Note that even if the setting is made so that the IDR is not used in one encoder, the IDR must be put at the head, so that identifiers may overlap at that time. As a result, if two streams are combined as they are, identifiers may be duplicated.

  Therefore, a method of forcibly changing the identifier in the header correcting unit 111 after combining the streams in the stream combining unit 109 is also conceivable. However, H.C. In H.264, IDR identifiers are subjected to variable-length coding in which the code length changes depending on the data before and after Golomb coding. Therefore, it is very difficult to correct the identifier after coding. .

  Therefore, in this embodiment, as shown in FIG. 2, the image sorting unit 105 adds a dummy picture 201 that is not input data to the head in advance. Then, the first encoder 106 or the second encoder 107 encodes the leading dummy picture by intra-screen encoding, but the encoded dummy picture 204 is combined by the stream combining unit 109. Control not to include in the stream. In this way, it is not necessary to acquire IDR identifiers that are difficult to modify. As shown in FIG. 2, in the data sequence in the composite stream 205, the leading dummy picture 204 is not included because it is dummy data.

  Further, in this embodiment, all the pictures of intra-frame coding performed by the first encoder 106 are IDR, and all the pictures of intra-frame coding performed by the second encoder 107 are I pictures other than the head. . The difference between the IDR and the I picture is not only the identifier described above, but also a parameter indicating the picture number called frame_num written in the header in the picture. H. H.264 stipulates that frame_num is counted up with an I picture and must be reset when an IDR is received.

  Thus, the frame_num of the headers in all IDRs output from the first encoder 106 is 0. On the other hand, the frame_num of the header in all I pictures other than the head output from the second encoder 107 is counted up from 0 because there is no IDR other than the head. Therefore, at the stage where streams are combined by the stream combining unit 109, frame_num is counted up despite the presence of IDR between I pictures, resulting in an error stream state.

  Therefore, in the present embodiment, after the two streams are combined by the stream combination unit 109, the header detection unit 110 detects the header including frame_num. At this time, since the frame_num of the IDR is always 0, there is no need to detect it, and the header including the frame_num of the I picture generated by the second encoder 107 is detected using the information output from the image sorting unit 105. To work.

  Then, the header detection unit 110 sends the detected header information to the header correction unit 111, and the header correction unit 111 corrects the value of frame_num of the I picture in the composite stream. As described above, the frame_num of IDR is 0, and the I_frame and IDR are nested as shown in the composite stream 205 in FIG. In this way, it is not necessary to have a counter in order to change the value of frame_num of the I picture to achieve matching, and it is only necessary to always replace it with 0. Note that frame_num is not Golomb encoded, and there is no inconvenience caused by replacing it with 0.

  H. In order to obtain a composite stream compliant with H.264, it is necessary to manage a stream buffer (virtual buffer) built in the virtual decoder. This stream buffer is defined by the level at which the stream is generated, and differs depending on the level as shown in FIG. Note that the capacity of the stream buffer varies depending on other conditions, but the ratio between the levels is always constant. In the present embodiment, even if the capacity of the stream buffer changes due to other conditions, the processing itself is not affected. Therefore, a detailed description of the capacity change due to other conditions is omitted.

  As with a general encoder, the first encoder 106 and the second encoder 107 of the present embodiment, when a certain level is set, do not break down the stream buffer defined by that level. Create a stream on For example, when both encoders generate a stream at level 4, the buffer size is 25000 as shown in FIG.

  However, even if each generated stream satisfies level 4, the combined stream is a combination of the two streams, and may not satisfy the standard. That is, when two streams are combined into one, the required amount of stream buffer may be doubled. Therefore, the header detection unit 110 detects the header specifying the level, and the header correction unit 111 rewrites the stream buffer of the combined stream to a level at which the stream buffer does not fail.

  For example, when streams generated at level 4 are combined, they are rewritten to level 4.1 or higher having a stream buffer twice or more. Since the level information is not Golomb encoded, there is no inconvenience due to rewriting. In the example of the level shown in FIG. 6, the stream buffer has been described. However, the level defines not only the stream buffer but also the picture rate. The picture rate also needs to be changed to a level that satisfies the double rate of the composite stream, as with the stream buffer.

  As described above, according to the present embodiment, it is possible to parallelize encoding processes without preparing a special encoder.

(Second Embodiment)
Hereinafter, a second embodiment of the present invention will be described with reference to FIG. Note that the configuration of the encoding apparatus according to the present embodiment is the same as that of the first embodiment, and a description thereof will be omitted.

FIG. 3 is a diagram for explaining the outline of the operation of the encoding process in the present embodiment. In addition, the same code | symbol is attached | subjected about the thing similar to FIG.
A stream 300 in FIG. 3 indicates a stream generated by the encoding process of the first encoder 106. The combined stream 301 is a stream obtained by combining the stream 300 generated by the first encoder 106 and the stream 203 generated by the second encoder 107.

  In the present embodiment, as in the second encoder 107, the first encoder 106 encodes only the first picture of the input image as IDR and the subsequent pictures as I pictures. The synthesis procedure is the same as that in the first embodiment. As a result, the composite stream 301 has a format of IDR only at the head.

  Also, frame_num is incremented individually in the first encoder 106 and the second encoder 107. Therefore, the composite stream 301 has a picture having the same frame_num, and is in an error stream state.

  According to the present embodiment, as shown in FIG. 3, the frame_num of the picture starting from the even number and the next odd-numbered picture, such as I4 and I5, I6 and I7, and I8 and I9, coincide. In the first embodiment, every other picture becomes an IDR picture, so the value of frame_num of the I picture only needs to be 0. In this case, since the IDR picture is only the top, if the value of frame_num is 0, the error stream End up.

  Therefore, in the present embodiment, the header correction unit 111 is provided with a counter function to correct the frame_num value of the composite stream 301 to a value incremented in the encoding order. As described above, it is possible to parallelize the encoding process without preparing a special encoder.

(Third embodiment)
The third embodiment of the present invention will be described below with reference to FIG. Note that the configuration of the encoding apparatus according to the present embodiment is the same as that of the first embodiment, and a description thereof will be omitted.

FIG. 4 is a diagram for explaining the outline of the operation of the encoding process in the present embodiment. In addition, the same code | symbol is attached | subjected about the thing similar to FIG.
In FIG. 4, a moving image 400 is expressed in units of pictures with the horizontal axis as a time axis. In the present embodiment, the leading picture 401 of the moving image 400 is not a dummy picture, and no dummy picture is added to the beginning.

  A stream 402 indicates a stream generated by the encoding process of the second encoder 107. An encoded picture 403 is a leading picture encoded by the second encoder 107. The combined stream 404 is a stream obtained by combining the stream 202 generated by the first encoder 106 and the stream 402 generated by the second encoder 107.

  In the present embodiment, as shown in FIG. 4, both the first encoder 106 and the second encoder 107 encode all input images as IDR pictures. Further, unlike the first and second embodiments, the first picture 401 of the input image does not have to be a dummy picture as described above. Therefore, the composite stream 404 is composed of all encoded pictures including the first picture 401 of the input image.

  Here, identification IDs are added to the headers of all IDR pictures of the composite stream 404. As described above, unlike the frame_num modified in the first and second embodiments, the identification ID is subjected to Golomb coding, which is a variable-length coding scheme. If IDR pictures having the same identification ID exist in the same stream, This violates the H.264 standard and is difficult to correct because the identification ID is variable-length encoded.

  Therefore, in this embodiment, IDR picture identification IDs are set in the first encoder 106 and the second encoder 107 in accordance with the input parameters given from the encoding parameter setting unit 102. In this case, the input parameter includes identification ID distribution information.

  The value of the identification ID itself is H.264. It is not specified by H.264. Therefore, even when an arbitrary identification ID is added to the IDR, the respective streams output from the first encoder 106 and the second encoder 107 are the same as in the first and second embodiments. H. H.264. Regarding frame_num, it is specified that frame_num of IDR is always 0, so there is no need to change it. In the case of this embodiment, the header to be corrected by the combined stream may be only the level portion. As described above, according to the present embodiment, it is possible to parallelize encoding processes without preparing a special encoder.

(Fourth embodiment)
The fourth embodiment of the present invention will be described below with reference to FIGS.
FIG. 5 is a block diagram illustrating a schematic configuration example of the encoding apparatus 500 according to the present embodiment. In addition, about the structure similar to FIG. 1, the same number is attached | subjected and the description is abbreviate | omitted.
In FIG. 5, the first encoder 501 and the second encoder 502 constituting the encoding unit 503 have a function of controlling the stream buffer. The encoding apparatus 500 according to the present embodiment is different from the first to third embodiments in that a rate control unit is unnecessary.

  Similar to the first to third embodiments, when generating a composite stream, it is necessary to consider the possibility of the stream buffer failing. Therefore, there is a limit to the level set for the first encoder and the second encoder. For example, when the final result composite stream is to be set to level 4.1, the maximum value set in the first encoder and the second encoder is level 4 as shown in FIG.

  However, in practice, level 4.1 allows a stream buffer capacity of up to 62500. On the other hand, the first encoder and the second encoder can use a buffer up to half of 31250, but at level 4, the capacity of the stream buffer is 25000. Thus, even if two encoders are added, only 50,000 are used, and a loss occurs in efficiency.

  Therefore, in this embodiment, the encoding parameter setting unit 102 sets the capacity of the stream buffers of the first encoder 501 and the second encoder 502 according to the input parameters. For example, as described above, when the composite stream is to be set to level 4.1, in this embodiment, the first encoder 501 and the second encoder 502 are also set to level 4.1. In the first encoder 501 and the second encoder 502, the capacity of the stream buffer to be used is set to 31250 which is half of 62500.

  As a result, the stream buffer can be used up to the maximum value 62500 defined in level 4.1, and a combined stream satisfying level 4.1 can be generated. Furthermore, since the level of the stream buffer generated by the first encoder 501 and the second encoder 502 is 4.1, the header correction unit 111 needs to correct the level for the combined stream. Disappear. Other than that, it is the same as the first to third embodiments. As described above, according to the present embodiment, it is possible to parallelize encoding processes without preparing a special encoder.

(Fifth embodiment)
The fifth embodiment of the present invention will be described below with reference to FIGS.
FIG. 7 is a block diagram illustrating a schematic configuration example of the encoding apparatus 700 according to the present embodiment. In addition, about the structure similar to FIG. 1, the same number is attached | subjected and the description is abbreviate | omitted.
In FIG. 7, an encoding unit 701 is composed of n encoders for parallel processing.

FIG. 8 is a diagram for explaining the outline of the operation of the encoding process in the present embodiment.
The input moving image 800 in FIG. 8 is expressed in units of pictures with the horizontal axis as a time axis, and n−1 dummy pictures 801 are added to the head of the input moving image 800. A stream 802 indicates a stream generated by the encoding process in the first encoder. The dummy picture 803 is an encoded picture that is positioned at the head as a result of the first dummy picture of the input moving image 800 being encoded by the second encoder. The dummy picture 804 is an encoded picture that is positioned at the head as a result of encoding the second dummy picture of the input moving image 800 by the third encoder. Furthermore, the dummy picture 805 is an encoded picture that is positioned at the head as a result of the n−1th dummy picture of the input moving image 800 being encoded by the nth encoder. The combined stream 806 is a stream obtained by combining all the streams generated by each encoder.

The detailed operation will be described below with reference to FIG. The basic processing procedure is the same as in the first embodiment.
In the present embodiment, the number of dummy pictures added to the input image by the image sorting unit 105 is one less than the number of encoders to be processed in parallel. As a result, the first encoder processes from the nth picture that is not a dummy picture. In the first embodiment, the processing time can be shortened to ½ because of two parallel processing, but in this embodiment, the processing time can be shortened to 1 / n.

  The second to nth encoders start the encoding process from the dummy picture. As a result, since the dummy picture encoded by each encoder is unnecessary, it is removed when combining, and a composite stream 806 starting from the nth picture is generated.

  As described above, according to this embodiment, the present invention can be applied even when n parallel processes are performed. In the present embodiment, an example in which the same processing as that of the first embodiment is performed by n parallel processes has been described, but the processing described in the second to fourth embodiments can be similarly applied.

(Other embodiments)
In all the embodiments described above, the unit of an image is expressed as a picture, but this unit of picture is applicable to both a frame in a progressive video and a field in an interlaced video.

  The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed.

105 Image Sorting Unit 106 First Encoder 107 Second Encoder 109 Stream Combining Unit 110 Header Detection Unit 111 Header Correction Unit

Claims (10)

A distribution unit that divides input image data in units of pictures and distributes the input image data to a plurality of image data;
A plurality of encoding means for encoding each of the plurality of image data distributed by the distribution means by intra-screen encoding;
Combining means for combining the plurality of image data encoded by the plurality of encoding means to generate combined image data;
Based on information relating to distribution by the distribution means, the plurality of image data before being combined by the combining means, or header detection means for detecting a predetermined header from the combined image data;
An encoding apparatus comprising: a header correction unit that corrects a predetermined header detected by the header detection unit and generates output image data from the combined image data. The header detection means detects a header indicating the order of pictures,
The encoding apparatus according to claim 1, wherein the header correcting unit corrects a header indicating the order of the pictures so as to match the order of the pictures of the output image data. The header detection means detects a header indicating the level of the virtual buffer,
The encoding apparatus according to claim 1, wherein the header correcting unit corrects a header indicating a level of the virtual buffer so as to be matched with a virtual buffer level of the output image data.   The encoding apparatus according to claim 1, further comprising a setting unit configured to set a capacity of a virtual buffer of the plurality of encoding units based on an input parameter.   2. The encoding according to claim 1, further comprising setting means for setting a distribution of identification IDs of a plurality of image data encoded by the plurality of encoding means based on input parameters. apparatus. A plurality of encoding means comprising a first encoder and a second encoder;
The distribution unit distributes the input image data to the first encoder and the second encoder alternately on a picture-by-picture basis, and assigns each picture to the first code. The encoding apparatus according to claim 1 or 2, wherein information indicating which of the encoder and the second encoder is allocated is output. The distribution means adds a dummy picture to the head of the input image data and outputs it to one of the first encoder and the second encoder,
The first encoder or the second encoder controls the dummy picture to be not included in the combined image data although the dummy picture is encoded by intra-picture encoding. The encoding apparatus according to claim 6.   The plurality of image data and the output image data are H.264. The encoding apparatus according to any one of claims 1 to 7, wherein the encoding apparatus conforms to H.264. A process of dividing the input image data in units of pictures and allocating it to a plurality of image data;
An encoding step of encoding a plurality of image data distributed in the distribution step by using a plurality of encoding means by intra-screen encoding,
A combining step of combining a plurality of image data encoded by the plurality of encoding means to generate combined image data;
Based on information relating to sorting in the sorting step, the plurality of image data before being combined in the combining step, or a header detecting step for detecting a predetermined header from the combined image data;
And a header correction step of correcting the predetermined header detected in the header detection step to generate output image data from the combined image data. A process of dividing the input image data in units of pictures and allocating it to a plurality of image data;
An encoding step of encoding a plurality of image data distributed in the distribution step by using a plurality of encoding means by intra-screen encoding,
A combining step of combining a plurality of image data encoded by the plurality of encoding means to generate combined image data;
Based on information relating to sorting in the sorting step, the plurality of image data before being combined in the combining step, or a header detecting step for detecting a predetermined header from the combined image data;
A program for causing a computer to execute a header correcting step of correcting a predetermined header detected in the header detecting step and generating output image data from the combined image data.

Images