JP2012191247A - Moving image encoding device and moving image encoding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a moving image encoding device with pipeline structure capable of reducing a generated code amount for each block to a specific maximum value or less without tracing the pipeline back.SOLUTION: A moving image encoding device 100 according to the invention allows in a header code string generation unit 107 to perform variable length encoding, at all time, on encoding information determined at the time of performing normal encoding processing on an encoding object block as header information, and in a coefficient code string generation unit 108 to perform processing by switching between a function of performing variable length encoding on residual coefficient information generated by normal encoding processing and a function of describing a reconstructed image in a code string without performing variable length encoding thereto.

Description

  The present invention relates to an image encoding device and an image encoding method for encoding an input moving image by dividing it into blocks.

  In recent years, with the development of multimedia applications, it has become common to handle all media information such as images, sounds and texts in a unified manner. Also, since a digitized image has a huge amount of data, an image information compression technique is indispensable for storage and transmission. On the other hand, in order to interoperate compressed image data, standardization of compression technology is also important. For example, as a standard for image compression technology, ITU-T (International Telecommunication Union Telecommunication Standardization Sector) 261, H.H. 263, H.M. H.264, ISO / IEC (International Organization for Standardization) MPEG-1, MPEG-3, MPEG-4, MPEG-4AVC, and the like. At present, standardization activities for a next-generation screen coding method called HEVC in cooperation with ITU-T and ISO / IEC are in progress.

  In such moving picture coding, each picture to be coded is divided into coding unit blocks, and the amount of information is compressed by reducing redundancy in the time direction and the spatial direction for each block. In inter-frame predictive coding for the purpose of reducing temporal redundancy, motion is detected and a predicted image is created in block units with reference to the front or rear picture, and the resulting predicted image and encoding target are obtained. The difference image with the block of is acquired. In addition, in the intra prediction encoding for the purpose of reducing spatial redundancy, a prediction image is generated from pixel information of surrounding encoded blocks, and the obtained prediction image and a block to be encoded are obtained. Get the difference image. Further, orthogonal transformation such as discrete cosine transformation and quantization are performed on the obtained difference image, and a code string is generated using variable length coding, thereby compressing the information amount.

  In decoding, prediction information and residual coefficient information are obtained by analyzing the code string generated by the encoding process, and prediction is performed by performing inter-screen prediction decoding and intra-screen prediction decoding using the prediction information. An image is generated, a difference image is generated by performing inverse quantization and inverse orthogonal transform on the residual coefficient information, and a final output image is restored by adding the obtained predicted image and the difference image.

  H. In H.264 (Non-Patent Document 1), in order to restrict the upper limit of the processing amount in block units, the maximum value of the generated code amount in block units is defined (specifically, 3200 bits). If the normal encoding process described above is performed, a code string exceeding the maximum value of the generated code amount may be generated depending on the properties of the input image and the conditions of the quantization process. By using the encoding mode, it is possible to always keep within the maximum value.

  Unlike the normal encoding mode, IPCM is a mode in which pixel values of an input image are described in a code string as they are as a bit string without generating a difference image by intra-screen / inter-screen prediction, or performing orthogonal transform / quantization. is there. Using this mode, for example, if the format of the input image is YUV 4: 2: 0 with 8 bits per pixel, the luminance component block is 16 × 16 pixels, and the two color difference component blocks are 8 × 8 pixels respectively. Therefore, the total is 384 bytes, and even if necessary information is included in the header, it can be surely stored within the maximum value of 3200 bits.

ITU-T H.264: Advanced video coding for generic audiovisual services (03/2010)

  Many moving image encoding / decoding devices realize encoding / decoding processing by an integrated circuit called LSI. Such an encoding / decoding device has a configuration that enables a parallel operation called a pipeline in order to increase processing speed. Specifically, the processing proceeds simultaneously by starting the processing of the next block before the processing of one block is completed.

  FIG. 3A shows an example of a pipeline in encoding. The processing of pixel reading, mode determination, inter-screen / intra prediction, transform / quantization, and variable-length encoding is sequentially applied to block 1, and the same processing is applied to block 2. . At this time, the block 2 starts the processing immediately after the pixel reading of the block 1 is completed, thereby performing the processing in parallel while delaying the processing timing by one step. H. In the H.264 and HEVC encoding / decoding processes, processing is performed with reference to information of blocks previously encoded / decoded, so that prediction information, pixel information, encoding determined in block 1 as shown in the figure It is necessary to perform processing while the block 2 refers to information and the like.

  However, since it is not possible to determine whether or not the generated code amount for each block is below the maximum value unless the code amount at the time of completion of variable-length coding is checked, it is determined that the maximum value is exceeded. In such a case, the code string must be regenerated by switching to the IPCM at that time.

  FIG. 3B shows an example of a pipeline when switching to IPCM occurs. Although it has been determined that switching to IPCM is performed in the variable length encoding process of block 1, the encoding process has already been performed for block 2 while referring to prediction information, pixel information, and the like when block 1 performs normal encoding. Since the process has progressed, it is necessary to return to the block 1 mode determination, update the information to be referred to by replacing the assumption that the block 1 is encoded with the IPCM, and repeat the process of the block 2.

  In this way, the control going back in the pipeline requires very complicated processing control, and when the number of IPCMs that occur in the target picture increases and the number of times that it goes back increases, it causes a processing speed delay. The encoding process cannot be completed.

  The present invention solves the above-mentioned problem, and in a coding apparatus having a pipeline structure, it is possible to keep the generated code amount in block units below a specific maximum value without going back through the pipeline. It is an object of the present invention to provide an encoding method that replaces the IPCM.

  The moving image encoding device according to the present invention is a moving image encoding device that encodes an input moving image in units of blocks, and includes a predicted image generation unit that generates a predicted image corresponding to an encoding target image; A subtractor that generates a difference image between the encoding target image and the generated predicted image, and a prediction residual code that generates an orthogonal coefficient by performing orthogonal transform processing and quantization processing on the output of the subtractor A prediction residual decoding unit that generates a residual decoded image by performing inverse quantization processing and inverse orthogonal transformation processing on the residual coefficient, and a prediction image generated by the prediction image generation unit And an adder that generates a reconstructed image by adding the residual decoded image generated by the prediction residual decoding unit, and header information including at least prediction information used when generating the predicted image Header code string generation to generate In the first mode, the residual coefficient generated by the prediction residual encoding unit is variable-length encoded to generate a first coefficient code string, and the coefficient code string is generated by the header information generation unit While the header information is associated, the coefficient code string and the header information are output, and in the second mode, the reconstructed image is generated without performing variable length coding on the reconstructed image generated by the adder. A coefficient code string as it is, and a coefficient code string generator that outputs the coefficient code string and the header information in a state where the coefficient code string is associated with the header information generated by the header information generator.

  Further, the moving image encoding apparatus according to the present invention is a moving image encoding apparatus that encodes an input moving image in units of blocks, and a predicted image generating unit that generates a predicted image corresponding to an encoding target image; A subtractor that generates a difference image between the encoding target image and the generated predicted image; and a prediction residual that generates a residual coefficient by performing orthogonal transform processing and quantization processing on the output of the subtractor An encoding unit, a prediction residual decoding unit that performs an inverse quantization process and an inverse orthogonal transform process on the residual coefficient to generate a residual decoded image, and a prediction generated by the prediction image generation unit An adder that generates a reconstructed image by adding the image and the residual decoded image generated by the prediction residual decoding unit, and a header that includes at least prediction information used when generating the predicted image Header code string that generates information In the generating unit and the first mode, the residual coefficient generated by the prediction residual encoding unit is variable-length encoded to generate a coefficient code string, and the coefficient code string is generated by the header information generating unit While the header information is associated, the coefficient code string and the header information are output, and in the second mode, the encoding target image is directly converted into the coefficient code string without performing variable length encoding on the encoding target image. And a coefficient code string generation unit that outputs the coefficient code string and the header information in a state in which the header information generated by the header information generation unit is associated with the coefficient code string.

  Preferably, the coefficient code string generation unit associates the coefficient code string and the header information with an identifier indicating whether or not the reconstructed image is directly used as a coefficient code string. The header information and the identifier are output.

  Preferably, the coefficient code string generation unit associates the coefficient code string and the header information with an identifier indicating whether or not the encoding target image is directly used as a coefficient code string. The column, the header information, and the identifier are output.

  Preferably, the coefficient code string generation unit performs processing before decoding the coefficient code string in the moving picture decoding apparatus that decodes the output code string, of the output code string that is the output result. In the above, the identifier is stored.

  Further preferably, the identification is an identifier used in common in the coefficient code string in the first mode and the coefficient code string in the second mode, and one of the identifier information is in the first mode. It indicates that it has been encoded as a coefficient code string, and another one indicates that it has been encoded as a coefficient code string in the second mode, and indicates whether or not there is an encoded residual coefficient. Is.

  Preferably, the coefficient code string generation unit outputs the code string using the second mode when there is a possibility that the code amount in the entire block of the code string to be output exceeds a specific size. .

  Preferably, the coefficient code string generation unit generates the code string in the first mode when the code amount in the entire block of the code string exceeds a specific size as a result of outputting the code string in the first mode. Instead of the code string, the code string generated using the second mode is output.

  Note that the present invention can be realized not only as such a moving image encoding / decoding device, but also as a program or an integrated process equivalent to each unit included in such a moving image encoding / decoding device. It can also be realized as a circuit.

  As described above, the moving image encoding apparatus according to the present invention is the case where the information referred to by the block subsequent to the encoding target block is always normally encoded even when the encoding method is realized using the pipeline structure. Since the information can be used, it is possible to continue the encoding process without going back in the pipeline, and the generated code amount for each block can be set to a specific maximum value without delaying the processing speed or increasing the processing amount. The following can be made.

  In the moving picture decoding apparatus according to the present invention, in the moving picture encoding apparatus that generates a corresponding code string, information referred to by a block subsequent to the encoding target block is always information obtained when normal encoding is performed. Because it can be used, it is possible to continue the encoding process without going back in the pipeline, and the generated code amount per block can be kept below the specified maximum value without increasing the processing speed or increasing the processing amount. It becomes possible to pay.

It is a block diagram which shows the moving image encoder which concerns on this Embodiment 1. It is a block diagram which shows the moving image decoding apparatus which concerns on this Embodiment 3. It is a conceptual diagram for demonstrating the pipeline control of the conventional moving image encoder. 3 is a flowchart of a code string generation process according to the first embodiment. It is a conceptual diagram for demonstrating an example of the syntax of the code sequence produced | generated by this embodiment. It is a conceptual diagram for demonstrating another example of the syntax of the code sequence produced | generated by this embodiment. It is a conceptual diagram for demonstrating another example of the syntax of the code sequence produced | generated by this embodiment. It is a conceptual diagram for demonstrating the pipeline control of the moving image encoder which concerns on this Embodiment 1. FIG. It is a block diagram which shows the moving image encoder which concerns on this Embodiment 2. 10 is a flowchart of a code string generation process according to the second embodiment. It is a conceptual diagram for demonstrating the pipeline control of the moving image encoder which concerns on this Embodiment 2. FIG. 10 is a flowchart of code string analysis processing according to the third embodiment.

(Embodiment 1)
Hereinafter, the first embodiment will be described with reference to the drawings.

  FIG. 1 is a block diagram of a video encoding apparatus 100 according to this embodiment. The moving image encoding apparatus 100 divides a moving image input in units of pictures into blocks, performs an encoding process in units of blocks, and generates a code string.

  The moving picture encoding apparatus 100 includes a picture memory 101, a prediction residual encoding unit 102, a prediction residual decoding unit 103, a local buffer 104, a prediction encoding unit 105, and a quantization value determination unit 106. A header code string generation unit 107 and a coefficient code string generation unit 108.

  The picture memory 101 stores the input image signal 151 input in units of pictures in the order of display by rearranging the pictures in the order of encoding. Next, when the picture memory 101 receives a read command from the difference calculation unit 109 or the predictive coding unit 105, the picture memory 101 outputs an image signal related to the read command. At this time, each picture is divided into coding units composed of a plurality of pixels called coding units (CUs). This CU includes, for example, a horizontal 64 × vertical 64 pixel block, a horizontal 32 × vertical 32 pixel block, and a horizontal 16 × vertical 16 pixel block. In the moving picture encoding apparatus 100 according to the present embodiment, the subsequent processing is performed in units of CUs.

  The prediction residual encoding unit 102 performs orthogonal transformation on the difference image signal 152 output from the difference calculation unit 109. Further, image information is compressed by performing quantization on the obtained orthogonal transform coefficient of each frequency component, and a residual encoded signal 153 is generated. The prediction residual encoding unit 102 outputs the generated residual encoded signal 153 to the prediction residual decoding unit 103 and the coefficient code string generation unit 108. Note that the prediction residual encoding unit 102 quantizes the orthogonal transform coefficient using the quantized value signal 158 determined by the quantized value determining unit 106.

  The prediction residual decoding unit 103 restores the difference image information by performing inverse quantization and inverse orthogonal transformation on the residual encoded signal 153 output from the prediction residual encoding unit 102. Then, the generated residual decoded signal 154 is output to the addition operation unit 110.

  The local buffer 104 stores the reconstructed image signal 155 output from the addition operation unit 110. The reconstructed image signal 155 is used for predictive coding processing in coding of a picture subsequent to a picture that is currently being coded. That is, the reconstructed image signal 155 is referred to as pixel data when a picture after the current picture to be encoded is encoded. In response to the read command from the predictive encoding unit 105, the local buffer 104 outputs the stored reconstructed image signal 155 as pixel data to the predictive encoding unit 105.

  The predictive coding unit 105 generates a predicted image signal 156 using intra prediction or inter prediction based on the image signal output from the picture memory 101. Then, the predictive coding unit 105 outputs the generated predicted image signal 156 to the difference calculation unit 109 and the addition calculation unit 110. Note that the prediction encoding unit 105 uses the reconstructed image signal 155 of a past picture that has already been encoded and stored in the local buffer 104 when using inter-screen prediction. Further, when using intra prediction, the predictive encoding unit 105 uses the reconstructed image signal 155 of the current picture of an already encoded CU adjacent to the encoding target CU. The mode determination method of using intra-screen prediction or inter-screen prediction is performed by predicting which prediction method can reduce the information amount of the residual signal.

  Based on the picture stored in the picture memory 101, the quantized value determining unit 106 sets a quantized value used when the prediction residual encoding unit 102 quantizes the difference image signal 152. The quantized value determination unit 106 outputs the set quantized value to the prediction residual encoding unit 102 and the header code string generation unit 107. Note that the quantization value setting method in the quantization value determination 106 is such that the quantization value is set so that the bit rate of the code string signal 159 approaches the target bit rate, that is, the quantization value based on so-called rate control. You may use the setting method.

  The header code string generation unit 107 is a variable-length code for the prediction information signal 157 output from the prediction encoding unit 105, the quantization value signal 158 output from the quantization value determination unit 106, and other control information related to encoding control. To generate a code string. Note that the prediction information includes, for example, information indicating an intra-screen prediction mode, an inter-screen prediction mode, a motion vector, a reference picture, and the like. The control information is information that can be acquired before the processing in the coefficient code string generation unit 108, and is information that indicates the encoding condition applied when the CU is encoded. For example, block coding type, block division information, and the like are included.

  The coefficient code string generation unit 108 follows the code string generated by the header code string generation unit 107 after the code string generated by variable-length encoding the residual encoded signal 153 output from the prediction residual encoding unit 102. A first mode for generating a final code string signal 159 by additionally writing is provided. Further, the coefficient code string generation unit 108 continues the code string obtained without performing variable length coding on the reconstructed image signal 155 output from the addition calculation unit 110, following the code string generated by the header code string generation unit 107. And a second mode for generating a final code string signal 159 by additionally writing.

  The difference calculation unit 109 generates a difference image signal 152 that is a difference value between the image signal read from the picture memory 101 and the prediction image signal 156 that is the output of the prediction encoding unit 105, and performs prediction residual encoding. Output to the unit 102.

  The addition operation unit 110 adds the residual decoded signal 154 output from the prediction residual decoding unit 103 and the predicted image signal 156 output from the prediction encoding unit 105, thereby adding the reconstructed image signal 155. Generate and output to the local buffer 104.

  Here, a method for generating a code string signal in the header code string generation unit 107 and the coefficient code string generation unit 108 will be specifically described with reference to the flowchart of FIG.

  First, the header code string generation unit 107 performs variable length encoding with the prediction information signal 157, the quantized value signal 158, and other encoding control information generated as a result of performing the normal encoding process described above as inputs. Thus, a code string of header information is generated (S401).

  Next, in step S402, the coefficient code string generation unit 108 uses the input residual encoded signal 153 to determine whether or not the generated code amount of the encoding target CU may exceed the specified value. To do.

  When it is determined in step S402 that there is no possibility of exceeding, an identifier indicating that the coefficient is encoded as the Residual mode is encoded (S403). Subsequently, a code string is generated by variable-length encoding the input residual encoded signal 153 (Residual mode) in the same manner as in the conventional encoding (S404).

  On the other hand, if it is determined in step S402 that there is a possibility of exceeding, an identifier indicating that the coefficient is encoded as the PCM mode is encoded (S405). Subsequently, the input reconstructed image signal 155 is not subjected to variable length coding, and the bit of the pixel is added to the code string as it is (PCM mode) to generate a code string (S406).

  Here, in step S402, it is determined whether or not the generated code amount of the CU to be encoded may exceed the specified value by using the input residual encoded signal 153. A method may be used to determine whether or not the generated code amount may exceed a specified value. For example, there is a method for determining whether the code amount exceeds a predetermined value based on the code string signal 159. In this case, since the code string has already been output from the coefficient code string generation unit 108 at the time of determination, in the code string, the residual encoded signal 153 is obtained by variable-length coding and replaced with the code string. Then, the processing is performed by directly replacing the pixel bit information indicated by the input reconstructed image signal 155.

Further, instead of performing the determination in units of CUs, the determination may be performed in units of a plurality of CUs or another block.
(Description of syntax)
FIG. 5 is a diagram illustrating an example of the syntax of the CU unit of the code string generated by the present embodiment: coding_unit ().

  Information such as prediction mode: pred_mode, prediction information: prediction_unit (), quantization value: qp_value, and the like generated by the header code string generation unit 107 are encoded at the beginning of the syntax.

  Subsequently, the identifier pcm_flag described in FIG. 4 is encoded. When this identifier is 0, it indicates that the coefficient code string is described in Residual_data () in the Residual mode. Further, when the identifier is 1, it indicates that the coefficient code string is described in pcm_data () in the PCM mode.

  FIG. 6 is a diagram showing another example of the syntax of the CU unit: coding_unit () of the code string generated according to the present embodiment. The only difference from the syntax described in FIG. 5 is that cbp_yuv_root is used instead of pcm_flag as an identifier.

  This identifier is used to indicate whether or not there is a residual encoded signal for each luminance component and chrominance component in the conventional encoding, and has conventional information about 0 to 7, The Residual mode indicates that the coefficient code string is described in Residual_data (), and 8 indicates that the coefficient code string is described in pcm_data () in the PCM mode. That is, the eighth information is extended to the information from 0 to 7 that has been conventionally used.

  As a result, it is possible to share and use a conventional signal as an identifier, and to suppress an increase in code amount due to the addition of a new identifier.

  FIG. 7 is a diagram showing still another example of the CU unit syntax: coding_unit () of the code string generated according to the present embodiment. The only difference from the syntax described in FIG. 5 is that residual_data_flag is used as an identifier instead of pcm_flag.

  This identifier is used to indicate whether or not there is a residual encoded signal in the target block in another conventional encoding. That is, when this identifier is 0, it indicates that there is no coefficient information as usual. Further, when the identifier is 1, it indicates that the coefficient code string is described in Residual_data () in the Residual mode because there is coefficient information as usual. Furthermore, when the identifier is 2, it indicates that the coefficient code string is described in pcm_data () in the PCM mode.

  As a result, it is possible to share and use a conventional signal as an identifier, and to suppress an increase in code amount due to the addition of a new identifier.

  Note that the syntax and identifier values described in FIGS. 5, 6, and 7 are examples for explaining the present embodiment, and syntax and identifier values different from those described here. A similar function may be realized by assigning.

  Note that the specified value in step S402 in FIG. 4 is an encoding of the code amount necessary for describing the pixel value of the previous reconstructed image as it is as a code string and all the information to be described in the header code string. This is a code amount obtained by adding a margin amount to a code amount that is combined with the maximum code amount that is necessary for the above. For example, when the format of the image is YUV 4: 2: 0 with 8 bits for each pixel and the size of the encoding target CU is 32 × 32 pixels, the pixel value of the previous reconstructed image is described as a code string as it is. The code amount required for 1 is 1536 bytes, and a margin amount is added to the code amount combined with the maximum code amount required for encoding all the information to be described in the header code string, A value such as 13000 bits can be considered as the specified value.

  Here, an example of a pipeline of the moving picture coding apparatus according to the present embodiment will be described with reference to FIG.

  FIG. 8A shows pipeline control when the coefficient code string is generated in the Residual mode as a result of the determination in step S402 of FIG. Processing is performed in exactly the same flow as in the conventional control described with reference to FIG.

  On the other hand, FIG. 8B shows pipeline control when the coefficient code string is generated in the PCM mode as a result of the determination in step S402 of FIG. Even when the block 1 is switched to the PCM mode, the prediction information described in the header code string and the pixel information of the finally generated reconstructed image are not changed. For this reason, there is no influence on the processing of block 2 in which the encoding processing is proceeding while referring to them. Therefore, it is possible to encode only the coefficient code string in the PCM mode without going back in the pipeline.

As described above, since the moving picture encoding apparatus according to the present embodiment can perform encoding by switching to the PCM mode without going back in the pipeline, the block can be achieved without increasing the processing speed or increasing the processing amount. It is possible to keep the generated code amount of a unit below a specific maximum value.
(Summary)
A moving image encoding apparatus 100 according to the present embodiment is a moving image encoding apparatus 100 that encodes an input moving image in units of blocks, and generates a prediction image corresponding to an encoding target image. 105, a difference calculation unit 109 that generates a difference image between the encoding target image and the generated predicted image, an orthogonal transform process and a quantization process are performed on the output of the difference calculation unit 109, and a residual A prediction residual encoding unit that generates a coefficient, a prediction residual decoding unit that performs an inverse quantization process and an inverse orthogonal transform process on the residual coefficient, and generates a residual decoded image; and the prediction An addition operation unit 110 that generates a reconstructed image by adding the prediction image generated by the encoding unit 105 and the residual decoded image generated by the prediction residual decoding unit, and generates the prediction image Used when A header code string generation unit 107 that generates header information including at least measurement information; and in the first mode, the residual coefficient generated by the prediction residual encoding unit 102 is variable-length encoded to generate a first coefficient code string. Generating and outputting the coefficient code string and the header information in a state in which the header information generated by the header information generation unit is associated with the coefficient code string, while in the second mode, the addition calculation unit 110 Without generating a variable-length code on the generated reconstructed image, the reconstructed image is directly used as a coefficient code string, and the coefficient code string is associated with the header information generated by the header information generating unit. A coefficient code string generation unit 108 that outputs a code string and the header information.

  Preferably, the coefficient code string generation unit 108 associates the coefficient code string and the header information with an identifier indicating whether or not the reconstructed image is directly used as a coefficient code string. The column, the header information, and the identifier are output.

  Also preferably, the coefficient code string generation unit 108 performs processing before decoding the coefficient code string in a moving picture decoding apparatus that decodes the output code string of output code strings that are output results. The identifier is stored in the position.

  Preferably, the identifier is an identifier used in common for the coefficient code string in the first mode and the coefficient code string in the second mode, and one of the identifier information is a coefficient in the first mode. It indicates that it has been encoded as a code string, and another one indicates that it has been encoded as a coefficient code string in the second mode, and indicates whether there is an encoded residual coefficient. It is.

  Here, the identifier corresponds to cbp_yuv_root shown in FIG. 6 or residual_data_flag shown in FIG. 7 in the present embodiment.

  Preferably, the coefficient code string generation unit 108 outputs a code string using the second mode when there is a possibility that the code amount in the entire block of the code string to be output exceeds a specific size. To do.

The coefficient code string generation unit 108 outputs a code string in the first mode, and as a result, when the code amount in the entire block of the code string exceeds a specific size, the code string generated in the first mode Instead, a code string generated using the second mode is output.
(Embodiment 2)
Next, the second embodiment will be described with reference to the drawings.

  FIG. 9 is a block diagram of the video encoding device 100-1 according to the present embodiment. The moving image encoding apparatus 100-1 divides a moving image input in units of pictures into blocks, performs encoding processing in units of blocks, and generates a code string.

  The moving image encoding apparatus 100-1 is substantially the same as the moving image encoding apparatus 100 described with reference to FIG. 1 in the first embodiment, but a coefficient code string generation unit is used instead of the coefficient code string generation unit 108. The only difference is that 108-1 is provided.

  Hereinafter, for convenience of description, detailed description of the same configuration as that of the first embodiment will be omitted. Further, in FIG. 9, blocks having the same functions as those in FIG.

  The coefficient code string generation unit 108-1 converts the code string obtained by variable-length coding the residual encoded signal 153 output from the prediction residual coding unit 102 into the code string generated by the header code string generation unit 107. It has the 1st mode which generates final code string signal 159-1 by adding continuously. Further, the coefficient code string generation unit 108-1 converts the code string obtained without performing variable-length coding on the block-unit input image output from the picture memory 101 into the code string generated by the header code string generation unit 107. It has the 2nd mode which generates final code sequence signal 159 by appending continuously.

  FIG. 10 is a flowchart illustrating a method of generating a code string signal in the header code string generation unit 107 and the coefficient code string generation unit 108-1.

  The process is substantially the same as that in the flowchart described with reference to FIG. 4 in the first embodiment, except that the process of step S406-1 is performed instead of step S406.

  In step S402, when it is determined that the generated code amount of the encoding target CU may exceed the specified value, an identifier indicating that the coefficient is encoded as the PCM mode is encoded (S405). . Subsequently, the input image signal is not subjected to variable length coding, and the bit of the pixel is added to the code string as it is (PCM mode) to generate a code string (S406-1).

  Further, in the first embodiment, an example of the CU unit syntax of the generated code string and the identifier encoded in S405 has been described with reference to FIGS. The syntax in this embodiment and the identifier encoded in S405 are the same as those in the first embodiment.

  Note that the specified value in step S402 in FIG. 10 encodes the amount of code necessary for describing the pixel value of the previous input image as a code string and all the information to be described in the header code string. This is a code amount obtained by adding a margin amount to the code amount combined with the maximum code amount required at the time. For example, when the image format is YUV 4: 2: 0 with 8 bits for each pixel and the size of the encoding target CU is 32 × 32 pixels, the pixel value of the previous input image is described as a code string as it is. The required code amount is 1536 bytes, and a margin amount is added to the code amount that is combined with the maximum code amount required when encoding all the information to be described in the header code string. A value such as 13000 bits can be considered as the specified value.

  Here, an example of a pipeline of the moving picture coding apparatus according to the present embodiment will be described with reference to FIG.

  FIG. 11A shows the pipeline control when the coefficient code string is generated in the Residual mode as a result of the determination in step S402 of FIG. Processing is performed in exactly the same flow as in the conventional control described with reference to FIG.

  On the other hand, FIG. 11B shows pipeline control when a coefficient code string is generated in the PCM mode as a result of the determination in step S402 of FIG. When the block 1 is switched to the PCM mode, the prediction information described in the header code string is not changed. However, finally generated pixel information is changed to an input image instead of a reconstructed image. On the other hand, since the encoding process of block 2 referring to the pixel information of block 1 is in progress, it is necessary to go back to the inter-screen / intra prediction process of block 1 in order to replace the pixel information and start the process again. However, compared with the conventional control described with reference to FIG.

As described above, since the moving picture encoding apparatus according to the present embodiment can reduce the amount of processing going back in the pipeline as compared with the prior art, it suppresses a delay in processing speed or an increase in the processing amount, while suppressing a block unit. In this embodiment, unlike the first embodiment, the generated code amount can be kept below a certain maximum value. Encode as a coefficient code string in PCM mode. Therefore, it is possible to improve the image quality of the image decoded by the corresponding video decoding device.
(Summary)
The moving image encoding apparatus according to the present embodiment is a moving image encoding apparatus 100-1 that encodes an input moving image in units of blocks, and generates a predicted image corresponding to an encoding target image. Unit 105, a difference calculation unit 109 that generates a difference image between the encoding target image and the generated predicted image, and an orthogonal transformation process and a quantization process on the output of the difference calculation unit 109, A prediction residual encoding unit 102 that generates a difference coefficient, a prediction residual decoding unit that performs an inverse quantization process and an inverse orthogonal transform process on the residual coefficient, and generates a residual decoded image; An addition operation unit 110 that generates a reconstructed image by adding the prediction image generated by the prediction encoding unit 105 and the residual decoded image generated by the prediction residual decoding unit; The schedule used when generating A header code string generation unit 107 that generates header information including at least information, and in the first mode, a coefficient code string is generated by variable-length coding the residual coefficient generated by the prediction residual coding unit 102; The coefficient code string and the header information are output in a state in which the header information generated by the header information generation unit is associated with the coefficient code string, while the encoding target image is a variable length code in the second mode. Without encoding, the coefficient code string and the header information are output in a state where the encoding target image is used as is as a coefficient code string, and the header information generated by the header information generation unit is associated with the coefficient code string. A coefficient code string generation unit 108.

  Preferably, the coefficient code string generation unit 108-1 associates an identifier indicating whether or not the encoding target image is directly used as a coefficient code string with the coefficient code string and the header information, The coefficient code string, the header information, and the identifier are output.

  Preferably, the coefficient code string generation unit 108-1 performs the decoding process on the coefficient code string in the moving picture decoding apparatus that decodes the output code string among the output code strings that are output results. The identifier is stored at a position to be processed.

  Preferably, the identifier is an identifier used in common for the coefficient code string in the first mode and the coefficient code string in the second mode, and one of the identifier information is a coefficient in the first mode. It indicates that it has been encoded as a code string, and another one indicates that it has been encoded as a coefficient code string in the second mode, and indicates whether there is an encoded residual coefficient. It is.

  Here, the identifier corresponds to cbp_yuv_root shown in FIG. 6 or residual_data_flag shown in FIG. 7 in the present embodiment.

  Preferably, the coefficient code string generation unit 108-1 uses the second mode when the code amount in the entire block of the code string to be output may exceed a specific size. Is output.

Preferably, the coefficient code string generation unit 108-1 outputs the code string in the first mode, and as a result, when the code amount in the entire block of the code string exceeds a specific size, Instead of the code string generated in the mode, the code string generated using the second mode is output.
(Embodiment 3)
Hereinafter, the third embodiment will be described with reference to the drawings.

  FIG. 2 is a block diagram of the video decoding apparatus 200 according to the present embodiment. The moving picture decoding apparatus 200 decodes the code string generated by the moving picture coding apparatus described in the first embodiment or the second embodiment in units of blocks called coding units (CUs), and outputs the decoded data. Generate an image.

  The moving picture decoding apparatus 200 includes a header code string analysis unit 201, a coefficient code string analysis unit 202, a prediction residual recovery unit 203, a picture memory 204, a prediction decoding unit 205, and a quantization value determination. Part 206.

  The header code string analysis unit 201 analyzes header information by performing variable length decoding on the header area of the input code string signal 251 in block units. The header code string analysis unit 201 outputs the prediction information signal 256 obtained by the analysis to the prediction decoding unit 205. Further, the header code string analysis unit 201 outputs the quantized value information obtained by the analysis to the quantized value determination unit 206.

  The coefficient code string analysis unit 202 analyzes the coefficient code string encoded following the header information analyzed by the header code string analysis unit 201. At this time, if the coefficient code sequence is a residual encoded signal as a result of the analysis, the coefficient code sequence analyzing unit 202 outputs the residual encoded signal 252 to the prediction residual decoding unit 203. On the other hand, when the coefficient code string is a reconstructed image signal as a result of the analysis, the coefficient code string analyzer 202 replaces the reconstructed image signal with the reconstructed image signal 255. At this time, when processing is performed in a mode that replaces the reconstructed image signal 255, a process of generating a residual decoded signal 253 by the prediction residual decoding unit 203 described below, and a process of generating the prediction image signal 254 by the prediction decoding unit 205 will be described. The generation process may not be performed.

  The prediction residual decoding unit 203 decodes the difference image information by performing inverse quantization and inverse orthogonal transform on the residual encoded signal 252 input from the coefficient code string analysis unit 202. Then, the prediction residual decoding unit 203 outputs the generated residual decoding signal 253 to the addition operation unit 207. At this time, the prediction residual decoding unit 203 dequantizes the residual encoded signal 252 using the quantized value signal 257 determined by the quantized value determining unit 206.

  The picture memory 204 rearranges and accumulates the signals input in order from the addition operation unit 207 or the coefficient code string analysis unit 202 in the order of display in the order of display. Then, the picture memory 204 outputs the accumulated signal as an output image signal 258 to the outside.

  The predictive decoding unit 205 generates a predicted image signal 254 using intra prediction or inter prediction based on the prediction information signal 256 output from the header code string analysis unit 201. Then, the predictive decoding unit 205 outputs the generated predicted image signal 254 to the addition operation unit 207. Note that the predictive decoding unit 205 uses the reconstructed image signal 255 of a past picture that has already been decoded and stored in the picture memory 204 when using inter-screen prediction. Further, when using intra prediction, the predictive decoding unit 205 uses the reconstructed image signal 255 of the current picture of a CU that has already been decoded and is adjacent to the decoding target CU. The determination of whether to use intra prediction or inter prediction is performed according to the input prediction information signal 256.

  The addition operation unit 207 generates a reconstructed image signal 255 by adding the residual decoded signal 253 output from the prediction residual decoding unit 203 and the predicted image signal 254 output from the prediction decoding unit 205. . The generated reconstructed image signal 255 is stored in the picture memory 204 and is finally output to the display device as an output image signal 258 in units of pictures.

  Here, a method for analyzing a code string in the header code string analysis unit 201 and the coefficient code string analysis unit 202 will be specifically described with reference to a flowchart of FIG.

  First, the header code string analysis unit 201 analyzes header information by performing variable-length decoding on the header area of the input code string, and generates a generated prediction information signal 256, quantized value information, and the like. Is output to each processing block described with reference to FIG. 2 (S1201).

  Next, the coefficient code string analysis unit 202 first analyzes the identifier in step S1202, and in step S1203, the analyzed identifier indicates that the coefficient is encoded as the PCM mode, or It is determined whether the coefficient is encoded as the Residual mode.

  If it is determined in step S1203 that the coefficient has been encoded as the Residual mode, the residual encoded signal 252 obtained by performing variable-length decoding on the input coefficient code string in the same manner as in the past. Is output to the prediction residual decoding unit 203 (S1204).

  On the other hand, if it is determined in step S1203 that the coefficient is encoded as the PCM mode, the obtained bit string is directly subjected to the reconstructed image signal without performing variable length decoding on the input coefficient code string. As a result, the processing is replaced with the reconstructed image signal 255 output from the addition operation unit 207 (S1205).

  In addition, although the processing method with respect to the code sequence produced | generated by the moving image encoder 100 demonstrated in Embodiment 1 was demonstrated here, in the moving image encoder 100-1 demonstrated in Embodiment 2, it is. The generated code string can be decoded by the same processing method. At this time, the information acquired in step S1205 is replaced with the input image in the corresponding encoding device, but the decoding process can always be processed as the reconstructed image signal 255 without distinction.

  Further, in the first embodiment, an example of the syntax and identifier of the CU unit of the generated code sequence has been described with reference to FIGS. 5 to 7. The identifiers analyzed in the tax and S1202 are exactly the same.

  Thus, by using the moving picture decoding apparatus according to the present embodiment, the moving picture encoding apparatus that generates a code string corresponding to the moving picture decoding apparatus can take the configuration described in the first embodiment. Thus, as shown in FIG. 8B, since encoding can be performed by switching to the PCM mode without going back in the pipeline, the generated code amount in block units can be reduced without increasing the processing speed or increasing the processing amount. It becomes possible to keep it below a certain maximum value.

Similarly, by using the moving picture decoding apparatus according to the present embodiment, the moving picture encoding apparatus that generates a code string corresponding to the moving picture decoding apparatus can have the configuration described in the second embodiment. As shown in FIG. 11B, since the amount of processing going back in the pipeline can be reduced as compared with the conventional method, the generated code amount for each block can be specified while suppressing a delay in processing speed or an increase in processing amount. It can be stored below the maximum value, and the image quality of the decoded image can be improved.
(Summary)
The moving picture decoding apparatus 200 according to the present invention decodes a code string 251 that is encoded by dividing a moving picture in units of pictures into blocks and includes a header code string and a coefficient code string. A decoding apparatus 200, a header code string analysis unit 201 that acquires the code string, and a prediction image that is used when variable length decoding is performed on the header code string and at least the code string is encoded A header code string analysis unit 201 that obtains prediction information regarding, and an identifier indicating whether or not the coefficient code string is information of a reconstructed image obtained in the process of encoding the code string from the acquired code string A coefficient code string analysis unit 202 that acquires and outputs data related to the coefficient code string based on the acquired identifier, and inverse quantization and inverse orthogonal transform for the output data And the prediction residual decoding unit 203 that generates a residual decoded image, and is included in at least one of an already decoded region and a different decoded image in the image to which the acquired code string belongs Based on the pixel data, a predictive decoding unit 205 that generates a predicted image corresponding to the acquired code string, and generates and outputs a reconstructed image by adding the residual decoded image and the predicted image And the coefficient code sequence analysis unit 202 performs variable length decoding on the coefficient code sequence when determining that the coefficient code sequence is not information of the reconstructed image based on the identifier. When determining that the coefficient code string is information of the reconstructed image based on the first mode for outputting the residual coefficient obtained by converting to the prediction residual decoding unit 203 and the identifier, Instead of the output of root operations unit 207, and a second mode for outputting the coefficient code string as the reconstructed image.

  Also preferably, the video decoding apparatus 200 decodes a code sequence that is encoded by dividing a moving image in units of pictures into blocks, and includes a header code sequence and a coefficient code sequence, A header code string analysis unit 201 that obtains the code string, and a header code that performs variable-length decoding on the header code string and obtains at least prediction information related to a prediction image used when the code string is encoded From the sequence analysis unit 201 and the acquired code sequence, an identifier indicating whether or not the coefficient code sequence is information of an input image before encoding the code sequence is acquired, and the code sequence based on the acquired identifier A coefficient code string analysis unit 202 that outputs data relating to, a prediction residual decoding unit 203 that performs inverse quantization and inverse orthogonal transform on the output data, and generates a residual decoded image; Based on pixel data included in at least one of an already decoded region and a different decoded image in an image to which the acquired code sequence belongs, a predicted image corresponding to the acquired code sequence is generated A prediction decoding unit 205; and an addition operation unit 207 that generates and outputs a reconstructed image by adding the residual decoded image and the predicted image, and the coefficient code string analysis unit 202 includes: When it is determined that the coefficient code string is information of the input image based on the identifier, a residual coefficient obtained by variable-length decoding the coefficient code string is output to the prediction residual decoding unit 203. When it is determined that the coefficient code string is information of the input image based on the first mode and the identifier, the coefficient code string is used as a reconstructed image instead of the output of the addition operation unit 207. And a second mode for outputting Te.

  Preferably, when the coefficient code string analysis unit 202 performs processing as the second mode based on the identifier, the prediction residual decoding unit 203 and the prediction decoding unit 205 respectively The difference decoded image and the predicted image are not generated.

Preferably, the identifier is an identifier used in common in the first mode and the second mode, and one of the identifier information indicates that decoding is performed as the second mode, Another one indicates that decoding is performed as the first mode, and whether or not there is a residual coefficient to be decoded.
(Other embodiments)
Further, by recording a program having the same function as each unit included in the moving image encoding device and the moving image decoding device described in the above embodiment, on a recording medium such as a flexible disk, the above The processing shown in the embodiment can be easily performed in an independent computer system. The recording medium is not limited to a flexible disk, and can be similarly implemented as long as it can record a program, such as an optical disk, an IC card, and a ROM cassette.

  In addition, functions equivalent to the units included in the moving picture coding apparatus and the moving picture decoding apparatus described in the above embodiments may be realized as an LSI that is an integrated circuit. These may be integrated into one chip so as to include a part or all of them. An LSI may also be called an IC, a system LSI, a super LSI, or an ultra LSI depending on the degree of integration.

  Further, the method of circuit integration is not limited to LSI's, and implementation using dedicated circuitry or general purpose processors is also possible. An FPGA (Field Programmable Gate Array) that can be programmed after manufacturing the LSI, or a reconfigurable processor that can reconfigure the connection and setting of circuit cells inside the LSI may be used.

  Furthermore, if integrated circuit technology that replaces LSI or the like appears as a result of progress in semiconductor technology or other derived technology, it is naturally also possible to carry out function block integration using this technology.

  Further, the present invention is applied to a broadcast wave recording apparatus such as a DVD recorder or a BD recorder that compresses and records a broadcast wave broadcast from a broadcast station, including the above-described moving picture encoding apparatus and moving picture decoding apparatus. It doesn't matter.

  Moreover, you may combine at least one part among the functions of the moving image encoder and moving image decoder which concern on the said embodiment, or its modification.

  The present invention, for example, in a video camera, a digital camera, a video recorder, a mobile phone, a personal computer, etc., a moving image encoding device that encodes each picture constituting an input image and outputs it as moving image encoded data, The present invention is useful as a moving picture decoding apparatus that generates a decoded picture by decoding the moving picture encoded data.

DESCRIPTION OF SYMBOLS 100 Moving image encoding apparatus 101 Picture memory 102 Prediction residual encoding part 103 Prediction residual decoding part 104 Local buffer 105 Prediction encoding part 106 Quantization value determination part 107 Header code sequence generation part 108 Coefficient code sequence generation part 109 Difference calculation unit 110 Addition calculation unit 151 Input image signal 152 Difference image signal 153 Residual encoded signal 154 Residual decoded signal 155 Reconstructed image signal 156 Predicted image signal 157 Predicted information signal 158 Quantized value signal 159 Code sequence signal 200 Moving picture decoding apparatus 201 Header code sequence analysis unit 202 Coefficient code sequence analysis unit 203 Prediction residual decoding unit 204 Picture memory 205 Prediction decoding unit 206 Quantized value determination unit 207 Addition operation unit 251 Code sequence signal 252 Residual code Signal 253 residual decoded signal 254 prediction image Issue 255 the reconstructed image signal 256 prediction information signal 257 quantized value signal 258 output image signals

Claims (10)

A video encoding device that encodes an input video in units of blocks,
A predicted image generation unit that generates a predicted image corresponding to the encoding target image;
A subtractor that generates a difference image between the encoding target image and the generated predicted image;
A prediction residual encoding unit that performs orthogonal transform processing and quantization processing on the output of the subtractor to generate a residual coefficient;
A prediction residual decoding unit that performs an inverse quantization process and an inverse orthogonal transform process on the residual coefficient to generate a residual decoded image;
An adder that generates a reconstructed image by adding the prediction image generated by the prediction image generation unit and the residual decoded image generated by the prediction residual decoding unit;
A header code string generation unit that generates header information including at least the prediction information used when generating the predicted image;
In the first mode, the residual coefficient generated by the prediction residual encoding unit is variable-length encoded to generate a first coefficient code sequence, and the header information generated by the header information generation unit in the coefficient code sequence In the second mode, the coefficient code string and the header information are output while the reconstructed image generated by the adder is not subjected to variable length coding in the second mode. A coefficient code string generator that outputs the coefficient code string and the header information in a state in which the header information generated by the header information generator is associated with the coefficient code string;
A video encoding device comprising: A video encoding device that encodes an input video in units of blocks,
A predicted image generation unit that generates a predicted image corresponding to the encoding target image;
A subtractor that generates a difference image between the encoding target image and the generated predicted image;
A prediction residual encoding unit that performs orthogonal transform processing and quantization processing on the output of the subtractor to generate a residual coefficient;
A prediction residual decoding unit that performs an inverse quantization process and an inverse orthogonal transform process on the residual coefficient to generate a residual decoded image;
An adder that generates a reconstructed image by adding the prediction image generated by the prediction image generation unit and the residual decoded image generated by the prediction residual decoding unit;
A header code string generation unit that generates header information including at least the prediction information used when generating the predicted image;
In the first mode, the residual coefficient generated by the prediction residual encoding unit is variable-length encoded to generate a coefficient code sequence, and the header information generated by the header information generation unit is associated with the coefficient code sequence. In the second mode, the coefficient code string and the header information are output in the attached state, and in the second mode, the coding target image is directly used as a coefficient code string without variable length coding, and the coefficient A coefficient code string generation unit that outputs the coefficient code string and the header information in a state in which the header information generated by the header information generation unit is associated with the code string;
A video encoding device comprising:   The coefficient code string generation unit associates the coefficient code string and the header information with an identifier indicating whether or not the reconstructed image is directly used as a coefficient code string. The moving picture encoding apparatus according to claim 1, wherein the identifier is output.   The coefficient code string generation unit associates the coefficient code string and the header information with an identifier indicating whether or not the encoding target image is directly used as a coefficient code string. The moving image encoding apparatus according to claim 2, wherein the information and the identifier are output.   The coefficient code string generation unit includes the identifier at a position where the coefficient code string is processed before the coefficient code string is decoded in a moving image decoding apparatus that decodes the output code string, of the output code string that is an output result. The moving picture encoding apparatus according to claim 3 or 4, wherein   The identifier is an identifier used in common for the coefficient code string in the first mode and the coefficient code string in the second mode, and one of the identifier information is encoded as a coefficient code string in the first mode. And the other one indicates that encoding is performed as a coefficient code string in the second mode, and indicates whether or not there is an encoded residual coefficient. The moving image encoding device according to claim 3 or 4.   The coefficient code string generation unit outputs the code string using the second mode when there is a possibility that the code amount in the entire block of the code string to be output exceeds a specific size. The moving image encoding apparatus according to claim 2.   When the code amount in the entire block of the code sequence exceeds a specific size as a result of outputting the code sequence in the first mode, the coefficient code sequence generation unit generates a code sequence generated in the first mode. Instead, the moving picture coding apparatus according to claim 1 or 2, wherein a code string generated using the second mode is output. A moving image encoding method for encoding an input moving image in block units,
Generate a predicted image corresponding to the encoding target image,
Generating a difference image between the encoding target image and the generated predicted image;
An orthogonal transformation process and a quantization process are performed on the output of the subtractor to generate a residual coefficient,
Performing an inverse quantization process and an inverse orthogonal transform process on the residual coefficient to generate a residual decoded image;
Generating a reconstructed image by adding the prediction image generated by the prediction image generation unit and the residual decoded image generated by the prediction residual decoding unit;
Generate header information including at least the prediction information used when generating the predicted image,
In the first mode, the generated residual coefficient is variable-length-encoded to generate a first coefficient code string, and the coefficient code string and the coefficient code string and the generated header information are associated with the coefficient code string. While outputting the header information, in the second mode, the reconstructed image is directly used as a coefficient code string without performing variable length coding on the generated reconstructed image, and the generated header information is included in the coefficient code string. A moving picture coding method for outputting the coefficient code string and the header information in a state of being associated with each other. A moving image encoding method for encoding an input moving image in block units,
Generate a predicted image corresponding to the encoding target image,
Generating a difference image between the encoding target image and the generated predicted image;
An orthogonal transformation process and a quantization process are performed on the output of the subtractor to generate a residual coefficient,
Performing an inverse quantization process and an inverse orthogonal transform process on the residual coefficient to generate a residual decoded image;
Generating a reconstructed image by adding the prediction image generated by the prediction image generation unit and the residual decoded image generated by the prediction residual decoding unit;
Generate header information including at least the prediction information used when generating the predicted image,
In the first mode, the coefficient code string and the header are generated in a state in which the generated residual coefficient is variable length encoded to generate a coefficient code string, and the generated header information is associated with the coefficient code string. While outputting information, in the second mode, without encoding the encoding target image as a variable length code, the encoding target image is used as it is as a coefficient code string, and the generated header information is associated with the coefficient code string. A coefficient code string generation unit that outputs the coefficient code string and the header information,
A video encoding method comprising:

Images