JP5194082B2 - Image encoding device - Google Patents

Description

  The present invention relates to an image encoding apparatus that encodes data relating to a moving image. The present invention relates to an H.264 / AVC encoding apparatus.

  Along with the development of digital technology, the technology for encoding video has also evolved and developed. However, the amount of data of video information, particularly moving image information, is very large. When a digital video encoded data is transferred by a medium such as a broadcast or a DVD, the transfer amount becomes very large. In particular, in high-definition broadcasting that has recently been put into practical use, the data amount is six times that of conventional SD (Standard Definition) video.

  With the development of digital video technology, in order to process the increasing amount of data, a technology for compressing data is used and developed for digital video data. The development has emerged as a compression technique specialized for video data that takes advantage of the characteristics of video data. In addition, with the improvement of information processing capabilities of computers and other devices, complex computations in compression techniques are possible, and the compression rate of video data has been greatly increased. For example, a compression technique employed in satellite and terrestrial digital high-definition broadcasting is a method called MPEG2, and satellite digital high-definition broadcasting compresses video data to about 1/30 by MPEG2.

  AVC / H.standardized as the next video compression technology of MPEG2. H.264 is a standard that realizes a compression rate about twice that of MPEG2. AVC / H. H.264 implements many compression technologies and combines them to achieve a high compression ratio. For this reason, the amount of calculation is also greatly increased.

  AVC / H. One compression technique implemented in H.264 is entropy coding (variable length coding). Two systems, CAVLC and CABAC, are prepared as entropy encoding systems. CAVLC is an abbreviation for Context-based Adaptive Variable Length Coding (context-adaptive variable-length coding system), and when coding DCT coefficients, a run and level that is a length of consecutive 0s and levels are variable-length coding tables. In this method, encoding is performed from the direction opposite to the scanning direction using.

  CABAC is an abbreviation for Context-based Adaptive Binary Arithmetic Coding (context-adaptive binary arithmetic coding), and is a method generally called arithmetic coding. . In CABAC, in addition to normal arithmetic coding, a context is assigned to each code to be compressed, and the appearance frequency is changed for each context.

  The encoding in CABAC is mainly divided into two processes. The first is a process called binarization that converts multi-value information to be encoded, called “Syntax”, into binary data. The second is a process for binary data converted by binarization. The context is calculated and arithmetic coding is performed.

  In arithmetic coding, 0/1 appearance probability is obtained for each bit of binary data, encoding processing is performed, and appearance probability is simultaneously updated. Therefore, processing is performed for each bit, and the operation speed is normally 1 bit / 1 clock. Moreover, since the appearance probability is adaptively switched for each bit and updated by the value of the bit, it is difficult to accurately know the output code amount before encoding.

  In addition, AVC / H. In H.264, the maximum code amount of a macroblock is limited to 3200 bits in the case of 4-2-0 format and bit_depth 8 bits. As a result of encoding a macroblock, if the code amount exceeds 3200 bits, it is necessary to change the encoding condition of the macroblock and perform encoding again. There is also a method in which mb_type is set to I_PCM and I_PCM data is stored as a bit stream instead of encoded data. Since it is difficult to accurately know the code amount output by CABAC until the encoded data is actually output after being encoded, whether or not the maximum code amount limit of the macroblock is satisfied is determined by the CABAC code. It is done after conversion.

  On the other hand, the amount of data input to the arithmetic coding circuit is monitored, and when this amount of data exceeds a certain amount within a certain unit, this input data is not used as encoded data, but is encoded with different parameters or not. A technique using compressed data as input data is also disclosed (for example, Patent Document 1).

  FIG. 15 is a functional block diagram showing a functional configuration of a conventional image encoding device.

  An image encoding device 50 shown in FIG. 15 is an example of a conventional image encoding device that encodes input image data and outputs it as a bit stream. An outline of the operation of the image encoding device 50 will be described below.

  Note that the image data is data relating to a moving image, such as substantive data such as macroblock data constituting the moving image, control information such as a macroblock type (mb_type), and other attribute information.

  When moving image data to be encoded is input, the data is input to the motion prediction unit 51, the in-plane prediction unit 52, and the I_PCM data buffer 53. The motion prediction unit 51 detects a block similar to the encoding target block from other pictures, obtains a difference image between the encoding target block and the detected block of the other picture, the difference image and the motion vector Is output. The in-plane prediction unit 52 predicts an image of an encoding target block in the screen using an image of an adjacent block, obtains a difference image between the image of the encoding target block and the predicted image, and calculates the difference image. And information indicating the prediction method is output. The I_PCM data buffer 53 is a buffer for storing I_PCM data that is original image data of an input moving image.

  The data output from the motion prediction unit 51 or the in-plane prediction unit 52 is selected by the input selection unit 54 and is input to the orthogonal transformation unit 55 that performs discrete cosine transformation (DCT), which is a kind of orthogonal transformation. . The orthogonal transform unit 55 outputs the DCT coefficient obtained by DCT to the quantization unit 56 and is quantized by the quantization unit 56. The quantized DCT coefficient is input to the binarizer 57 and converted into binary data, which is encoded by the CABAC encoder 59. The encoded data is input to the output selection unit 61 through the output buffer 61.

  Here, the I_PCM determination unit 58 monitors the output data amount from the binarizer 57 in the above-described data flow, and uses the data amount of the binarized data to output data to be output as a bit stream from CABAC. Whether the data is encoded by the encoder 59 or the I_PCM data stored in the I_PCM data buffer 53 can be determined. Alternatively, the output data amount from the CABAC encoder 59 is monitored, and the data output as a bit stream is determined as I_PCM data depending on whether or not the code amount that is the actual encoded data amount exceeds 3200 bits. Or encoded data can be determined. The output selection unit 61 outputs encoded data or I_PCM data according to the determination of the I_PCM determination unit 58.

JP 2004-135251 A

  Here, AVC / H. In H.264, information of the previous macroblock is used for in-plane prediction, motion prediction, and the like. Therefore, if the determination as to whether or not the maximum code amount limit of the macroblock is satisfied is performed after CABAC as in the above-described conventional technique, the next macroblock until the CABAC encoding of the macroblock is completed. In-plane prediction and motion prediction cannot be performed.

  In view of the above-described problems, an object of the present invention is to provide an image encoding apparatus that entropy-encodes binarized data and that processes image data at higher speed. .

  The image encoding device of the present invention includes an image encoding unit that encodes image data, an entropy encoding unit that binarizes and entropy-encodes encoded data obtained by the image encoding unit, and the entropy code Before the encoding unit binarizes the encoded data, it corresponds to a code amount when the encoded data of the predetermined unit is entropy-encoded based on information obtained from the encoded data of the predetermined unit. A prediction means for obtaining a prediction value; a comparison means for comparing the prediction value obtained by the prediction means with a predetermined threshold; and as a result of comparison by the comparison means, if the prediction value is greater than the threshold, the predetermined value A determination unit that determines to output I_PCM data that is data before entropy encoding of encoded data in units of If it is determined that outputs the _PCM data, the I_PCM outputs data, the I_PCM if it is not determined that the output data, and output means for outputting the entropy-coded data by the entropy coding means.

  With this configuration, the image encoding apparatus can determine whether or not the output data is I_PCM data before the encoded data of a predetermined unit is binarized. That is, it is possible to speed up the processing of image data by performing the above determination before performing the processing related to entropy coding.

  Further, in the image encoding device of the present invention, the predetermined unit is a macroblock, and the image encoding means uses the information of adjacent macroblocks adjacent to the encoding target macroblock to encode the encoding target macroblock. Motion prediction means for calculating a predicted motion vector mvp of the current macroblock, and when the current macroblock and the next macroblock to be coded next in coding order are adjacent to each other, the motion prediction means As a result of the comparison by the comparison means for the current macroblock, if the predicted value is larger than the threshold value, the information of the adjacent macroblock of the next macroblock is corrected, and the predicted motion vector mvp of the next macroblock is changed. Calculation may be performed.

  In the image encoding device of the present invention, the predetermined unit is a macroblock, and the image encoding means (1) uses the information of an adjacent macroblock adjacent to the encoding target macroblock to perform the encoding. A prediction motion vector mvp of the target macroblock is calculated, (2) a motion vector mv of the encoding target macroblock is calculated using the prediction motion vector mvp, and (3) the prediction motion vector mvp and the motion vector mv The motion prediction means calculates the difference mvd between the current macroblock and the next macroblock that is next encoded in the encoding order and is adjacent to the next macroblock. The predicted motion vector mvp, the motion vector mv, and the difference mvd are calculated with respect to the macroblock in the procedure (1), (2), and (3) As a result of the comparison by the comparison means for the current macroblock, if the predicted value is larger than the threshold value, the information on the adjacent macroblock of the next macroblock is corrected, and the next macroblock is corrected by the procedure (1). The predicted motion vector mvp of the next macroblock is calculated again, and the difference mvd of the next macroblock is calculated using the calculated motion vector mvp calculated again and the motion vector mv obtained in the procedure (2). May be calculated again.

  In the image encoding device of the present invention, the predetermined unit is a macroblock, and the image encoding means (1) uses the information of an adjacent macroblock adjacent to the encoding target macroblock to perform the encoding. A prediction motion vector mvp of the target macroblock is calculated, (2) a motion vector mv of the encoding target macroblock is calculated using the prediction motion vector mvp, and (3) the prediction motion vector mvp and the motion vector mv The motion prediction means calculates the difference mvd between the current macroblock and the next macroblock that is next encoded in the encoding order and is adjacent to the next macroblock. The prediction motion vector mvp and the motion vector mv are calculated for the macroblock in the steps (1) and (2). Before the difference mvd for the next macroblock is calculated by the motion prediction means in the procedure of (3), the current macroblock is compared, and the motion prediction means further includes the current macroblock. As a result of the comparison by the comparison means for the block, when the predicted value is larger than the threshold, the information of the adjacent macroblock of the next macroblock is corrected, and the next macroblock of the next macroblock is corrected by the procedure of (1). A predicted motion vector mvp is calculated again, and the difference mvd of the next macroblock is calculated using the calculated motion vector mvp calculated again and the motion vector mv obtained in the procedure (2). Also good.

  In addition, the image encoding device of the present invention is adjacent to the current macroblock, and a storage unit that stores at least the image data constituting the current macroblock that is the macroblock to be encoded among the input image data. Based on the motion vector calculated by the vector calculation unit, the vector calculation unit for calculating the motion vector of the image data constituting the current macro block, using the motion vector of a predetermined macro block of the macro block, An image encoding unit that encodes image data constituting the current macroblock, and image data before the current encoding unit stores the current macroblock stored in the storage unit, or the image Selects and outputs one of the image data encoded by the encoding unit A selection unit, and when the output selection unit selects and outputs image data stored in the storage unit for the current macroblock, the vector calculation unit is configured to output a motion vector of the current macroblock. The predicted motion vector of the next macroblock may be calculated in the coding order with respect to the current macroblock.

  In addition, the image encoding device of the present invention is adjacent to the current macroblock, and a storage unit that stores at least the image data constituting the current macroblock that is the macroblock to be encoded among the input image data. Based on the motion vector calculated by the vector calculation unit, the vector calculation unit for calculating the motion vector of the image data constituting the current macro block, using the motion vector of a predetermined macro block of the macro block, An image encoding unit that encodes image data constituting the current macroblock, and image data before the current encoding unit stores the current macroblock stored in the storage unit, or the image Selects and outputs one of the image data encoded by the encoding unit A selection unit, and when the output selection unit selects and outputs the image data stored in the storage unit for the current macroblock, the vector calculation unit configures the current macroblock Without using the motion vector of the image data, the motion vector of the image data constituting the next macroblock in the encoding order with respect to the current macroblock is predicted, and the predicted motion vector and the vector calculation unit calculated A difference vector that is a difference value from a motion vector of a current macroblock is calculated, and the image encoding unit outputs the difference vector linked to the encoded image data that constitutes the current macroblock. Good.

  The predetermined unit is a macroblock, and the image encoding unit includes an orthogonal transform unit that orthogonally transforms the image data, and a quantization unit that quantizes the coefficient obtained by the orthogonal transform unit. The predicting means includes coefficient converting means for converting the quantized coefficient obtained by the quantizing means into the number of bits when the quantized coefficient is binarized, and the coefficient for each macroblock. There may be provided addition means for adding the number of bits obtained by the conversion means and using the value obtained by the addition as the predicted value.

  In addition, the image encoding unit further includes a calculation unit that obtains a motion vector of the macroblock, and the prediction unit further binarizes the motion vector obtained by the calculation unit. Motion conversion means for converting into the number of bits in the case of being performed, and the addition means, for each macroblock, the number of bits obtained by the coefficient conversion means and the number of bits obtained by the motion conversion means The values obtained by the addition may be used as the predicted values.

  Further, the calculation means further obtains encoded information other than the motion vector of the macroblock, and the prediction means further determines the encoded information obtained by the calculation means to be 2 Coding information converting means for converting the number of bits into a number of bits when converted into a value, and the adding means is obtained for each macroblock by the number of bits obtained by the coefficient converting means and the motion converting means. The number of bits and the number of bits obtained by the encoded information conversion unit may be added, and a value obtained by the addition may be used as the predicted value.

  The predetermined unit is a macroblock, and the image encoding unit includes an orthogonal transform unit that orthogonally transforms the image data, and a quantization unit that quantizes the coefficient obtained by the orthogonal transform unit. The predicting means converts the quantized coefficient obtained by the quantizing means into a code amount when the quantized coefficient is entropy-encoded, and the coefficient for each macroblock. An addition unit that adds the code amounts obtained by the conversion unit and uses the added value as the predicted value may be included.

  In addition, the image encoding unit further includes a calculation unit for obtaining a motion vector of the macroblock, and the prediction unit further encodes the motion vector obtained by the calculation unit with the motion vector being entropy encoded. Motion conversion means for converting into a code amount in the case of being performed, the adding means, for each macroblock, the code amount obtained by the coefficient conversion means and the code amount obtained by the motion conversion means It is good also considering adding and making the added value the said predicted value.

  In addition, the calculation means further obtains encoded information other than the motion vector of the macroblock, and the prediction means further calculates the encoded information obtained by the calculation means, and the encoding information is entropy. Coding information converting means for converting into a code amount in the case of encoding, and the adding means is obtained for each macroblock by the code amount obtained by the coefficient converting means and the motion converting means. The code amount and the code amount obtained by the encoded information conversion means may be added, and the added value may be used as the predicted value.

  Further, the predetermined unit is a macroblock, and the predicting means further obtains an encoding condition as a condition for entropy encoding the macroblock, and calculates the predicted value according to the encoding condition. You may have a correction | amendment means to correct | amend.

  The image encoding means includes orthogonal transform means for orthogonally transforming the image data, and quantization means for quantizing a coefficient obtained by the orthogonal transform means, wherein the encoding condition is the macro The type of picture in which the block is included, the value of the quantization parameter when the quantization means quantizes, the value of the output rate of the data output from the image encoding device, and the type of residual_block for entropy encoding It may be at least one.

  The image encoding means includes an orthogonal transform means for orthogonally transforming the encoded data, a quantization means for quantizing a coefficient obtained by the orthogonal transform means, and a quantization obtained by the quantization means. Inverse quantization means for inversely quantizing the post coefficients, and inverse orthogonal transform means for obtaining a local decoded image by inverse orthogonal transform of the value obtained by the inverse quantization means, wherein the I_PCM data is the local It may be a decoded image.

  The I_PCM data may be image data input to the image encoding device.

  The predetermined unit is a macroblock, and the threshold value is AVC / H. The maximum code amount of a macroblock defined by H.264 may be used.

  The predetermined unit is a macroblock, and the threshold code amount is AVC / H. The maximum code amount of the macroblock defined by H.264 may be a value obtained by subtracting a predetermined margin value.

  The entropy encoding may be Context-based Adaptive Binary Arithmetic Coding (CABAC).

  The present invention can also be realized as an integrated circuit having characteristic components included in the image coding apparatus of the present invention. The present invention can also be realized as an image encoding method including characteristic processing executed by the image encoding apparatus of the present invention, and a program for causing a computer to execute the image encoding method.

  According to the present invention, it is determined whether or not to output I_PCM data before performing binarization processing for entropy coding. Thereby, it is possible to speed up the processing of the image data. For example, when the predetermined unit is a macroblock, the next in-plane prediction and motion prediction processing can be performed as soon as the orthogonal transformation and quantization of the macroblock are completed. Therefore, pipeline processing for each macroblock is facilitated, and the processing speed can be increased.

It is a functional block diagram which shows the functional structure outline | summary of the image coding apparatus of embodiment of this invention. It is a functional block diagram which shows another example of the functional structure outline | summary of the image coding apparatus shown in FIG. It is a functional block diagram which shows the specific example 1 of a functional structure of the image coding apparatus of embodiment of this invention. It is a functional block diagram which shows the specific example 2 of a functional structure of the image coding apparatus of embodiment of this invention. It is an example of the conversion table for converting the value of the coefficient after quantization into the number of bits. It is a functional block diagram which shows the specific example 3 of a functional structure of the image coding apparatus of embodiment of this invention. It is a functional block diagram which shows the specific example 4 of a functional structure of the image coding apparatus of embodiment of this invention. It is a functional block diagram which shows the specific example 5 of a functional structure of the image coding apparatus of embodiment of this invention. It is a functional block diagram which shows the specific example 6 of a functional structure of the image coding apparatus of embodiment of this invention. It is a functional block diagram which shows the specific example 7 of a functional structure of the image coding apparatus of embodiment of this invention. It is a functional block diagram which shows the specific example 8 of a functional structure of the image coding apparatus of embodiment of this invention. It is a functional block diagram which shows the specific example 9 of a functional structure of the image coding apparatus of embodiment of this invention. It is a figure for demonstrating an example of the efficiency improvement of the process which concerns on the motion prediction of the macroblock to be encoded. It is a figure which shows the time chart of the process which concerns on calculation of mvd. It is a functional block diagram which shows the functional structure of the conventional image coding apparatus.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

  FIG. 1 is a functional block diagram showing an outline of a functional configuration of an image encoding device according to an embodiment of the present invention. First, the outline of the configuration and operation of the image coding apparatus according to the embodiment of the present invention will be described with reference to FIGS. 1 and 2, and then the specific configuration and operation will be described with reference to FIG.

  An image encoding apparatus 1 shown in FIG. 1 is an apparatus that encodes image data and outputs it as a bit stream.

  As illustrated in FIG. 1, the image encoding device 1 includes an image encoding unit 10, an I_PCM data buffer 4, an output selection unit 8, an output unit 9, a code amount prediction unit 20, and an I_PCM determination unit 30. And an entropy encoding unit 40.

  The image encoding unit 10 is a processing unit that encodes input image data, and includes a motion prediction unit 2, an in-plane prediction unit 3, an input selection unit 5, an orthogonal transformation unit 6, and a quantization unit 7. And have.

  The entropy encoding unit 40 is a processing unit that performs CABAC encoding processing on the encoded data obtained by the image encoding unit 10, and includes a binarizer 41 and a CABAC encoder 42. ing.

  It should be noted that illustration of components inherent in the image coding apparatus, such as components that perform inverse quantization and inverse orthogonal transform, is omitted for the sake of simplicity.

  When image data to be encoded is input to the image encoding device 1, the data is input to the motion prediction unit 2, the in-plane prediction unit 3, and the I_PCM data buffer 4. That is, the original image data is stored as I_PCM data in the I_PCM data buffer 4. The data output from the motion prediction unit 2 or the in-plane prediction unit 3 is input to the orthogonal transform unit 6 that performs discrete cosine transform (DCT) by being selected by the input selection unit 5. The orthogonal transform unit 6 outputs the DCT coefficient obtained by DCT to the quantization unit 7 and is quantized by the quantization unit 7.

  The above operation flow is the same as that of the conventional image encoding apparatus. However, in the image encoding device 1 according to the embodiment of the present invention, the code amount prediction unit 20 uses the quantized DCT coefficients output from the quantization unit 7 (hereinafter referred to as “quantized coefficients”) or the like. From the value for each macroblock of the encoded data, a predicted value of the code amount after CABAC encoding for each macroblock can be obtained.

  The data input to the code amount prediction unit 20 is not limited to the output of the quantization unit, and any data that can predict the code amount can be used.

  FIG. 2 is a functional block diagram illustrating another example of a functional configuration outline of the image encoding device illustrated in FIG. 1.

  For example, the data selected by the input selection unit 5 can be used as the input data of the code amount prediction unit 20 as in the image encoding device shown in FIG. Since the output of the input selection unit 5 is converted into substantially unique data by the orthogonal transformation unit and the quantization unit, the same result is obtained even if the output of the input selection unit 5 is input to the code amount prediction unit 20. It is possible. However, it is necessary to change the inside of the code amount prediction unit 20 when the output of the input selection unit 5 is input to the code amount prediction unit 20 and when the output of the quantization unit 7 is input.

  The data output from the input selection unit 5 is also an example of encoded data in the image encoding device of the present invention.

  Furthermore, the I_PCM determination unit 30 has a set threshold for the code amount, and can compare the predicted value with the threshold.

  The I_PCM determination unit 30 instructs the output selection unit 8 based on the comparison result. Specifically, when the predicted value is larger than the threshold value, an instruction to select I_PCM data is issued. Further, when the predicted value is equal to or less than the threshold value, an instruction to select encoded data output from the image encoding unit 10 is issued.

  Also, the I_PCM determination unit 30 notifies the motion prediction unit 2 of the result of the I_PCM determination. The operation of the motion prediction unit that has received this notification will be described in specific examples 10 and 11 described later.

  When receiving an instruction to select I_PCM data, the output selection unit 8 acquires I_PCM data from the I_PCM data buffer 4 and outputs it to the output unit 9. Also, the control information “mb_type = I_PCM” is output to the entropy encoding unit 40.

  The entropy encoding unit 40 encodes only the input control information by the binarizer 41 and the CABAC encoder 42 and outputs the encoded information to the output unit 9.

  The output unit 9 outputs the I_PCM data output from the output selection unit 8 and the encoded control information output from the entropy encoding unit 40 to the outside as a bit stream.

  When receiving an instruction to select encoded data, the output selection unit 8 acquires quantized coefficient data from the quantization unit 7 and outputs the acquired coefficient data to the entropy encoding unit 40.

  The entropy encoding unit 40 encodes the input quantized coefficient data by the binarizer 41 and the CABAC encoder 42 and outputs the encoded coefficient data to the output unit 9. The output unit 9 outputs the encoded data output from the entropy encoding unit 40 to the outside as a bit stream. That is, the output selection unit 8 and the output unit 9 exhibit the function of outputting data based on the I_PCM determination that the output unit in the image encoding device of the present invention has.

  As described above, the image encoding device 1 according to the present invention determines whether the output bit stream is the I_PCM data or the encoded data subjected to the entropy encoding process before the entropy encoding process. Can be determined.

  That is, whether or not to output I_PCM data as a bit stream (hereinafter referred to as “I_PCM determination”) is performed for each macroblock, but the timing of this determination can be made earlier. Finally, when I_PCM data that is original image data is output as a bit stream, it is not necessary to perform processing for entropy coding. In other words, the processing speed is improved by determining the necessary operation at an earlier timing.

  Hereinafter, specific examples of operations and configurations of the image encoding device 1 will be described focusing on the specific configurations and operations of the code amount prediction unit 20 and the I_PCM determination unit 30.

(Specific example 1)
FIG. 3 is a functional block diagram illustrating a specific example 1 of the functional configuration of the image encoding device 1.

  As illustrated in FIG. 3, the I_PCM determination unit 30 includes a code amount comparison unit 31 and a threshold storage unit 32. Other components are the same as those of the image encoding apparatus shown in FIG.

  The threshold value storage unit 32 is a storage unit that stores a code amount threshold value that serves as a reference value for I_PCM determination.

  The code amount comparison unit 31 is a processing unit that compares the prediction value of the code amount for each macroblock output from the code amount prediction unit 20 with the threshold value stored in the threshold value storage unit 32. As a result of the comparison, if the predicted value is larger than the threshold value, the output selection unit 8 is instructed to select I_PCM data. Further, when the predicted value is equal to or less than the threshold value, the output selection unit 8 is instructed to select the quantized coefficient output from the image encoding unit 10 and the encoding information. The subsequent data flow is as described with reference to FIG.

  Here, the threshold value stored in the threshold value storage unit 32 is, for example, AVC / H. H.264 is 3200 bits, which is the maximum code amount of a macroblock defined in H.264. That is, in the image encoding device 1, I_PCM data is output as a bit stream for a macroblock whose code amount is predicted to exceed 3200 bits.

  The threshold value stored in the threshold value storage unit 32 is AVC / H. The value is not the same as the maximum code amount of the macroblock defined in H.264, but may be a value obtained by subtracting a certain margin value from the maximum code amount in order to make a safer determination. In addition, instead of storing a value obtained by subtracting a certain margin value as a threshold value in the threshold value storage unit 32, the code amount comparison unit 31 subtracts a predetermined margin value from the threshold value that is the maximum code amount to obtain a predicted value. You may compare with.

  In addition, the I_PCM determination may not be performed in units of macroblocks, and may be in larger units or in smaller units. For example, you may change according to the standard of entropy encoding.

(Specific example 2)
FIG. 4 is a functional block diagram illustrating a specific example 2 of the functional configuration of the image encoding device 1.

  As shown in FIG. 4, the code amount prediction unit 20 includes a coefficient bit number conversion unit 21 and an addition unit 25. In the image encoding device 1, the motion prediction unit 2, the in-plane prediction unit 3, and the input selection unit 5 constitute an encoding method calculation unit 10 a. Further, a merge unit 11 is provided.

  The encoding method calculation unit 10a separately outputs encoding information obtained from input image data and pixel difference data. The pixel difference data is input to the merge unit 11 via the orthogonal transform unit 6 and the quantization unit 7. The merge unit 11 merges the encoding information output from the encoding method calculation unit 10a and the quantized coefficient data obtained from the quantization unit 7.

  Encoding information is information such as a motion vector, mb_type, and intra prediction direction of a macroblock to be encoded. This is information relating to encoding of macroblocks.

  Other components are the same as those of the image encoding device 1 illustrated in FIG. 3, and the I_PCM determination unit 30 includes a code amount comparison unit 31 and a threshold storage unit 32.

  The coefficient bit number conversion unit 21 is an example of a coefficient conversion unit in the image encoding device of the present invention, and the quantized coefficient value output from the quantization unit 7 is converted into a binarizer. This is a processing unit for converting into the number of bits of binary data when binarized at 41.

  Specifically, the coefficient bit number conversion unit 21 has the conversion table shown in FIG. 5, and converts the value of the quantized coefficient into the number of bits using the conversion table.

  FIG. 5 is an example of a conversion table for converting the value of the quantized coefficient into the number of bits. As shown in FIG. 5, if the value of the quantized coefficient is known, the number of bits of binary data is determined.

  The number of bits output from the coefficient bit number conversion unit 21 is added for each macroblock in the addition unit 25 and output to the I_PCM determination unit 30. The number of bits added and output for each macroblock in the adding unit 25 is hereinafter referred to as “coefficient bit number”. That is, the number of coefficient bits is output to the I_PCM determination unit 30 as a predicted value.

  In the I_PCM determination unit 30, the threshold storage unit 32 stores a threshold for performing I_PCM determination using the number of coefficient bits output from the code amount prediction unit 20. That is, the prediction value output from the code amount prediction unit 20 is not the prediction value of the code amount after CABAC encoding for each macroblock, but the number of bits of binary data before the arithmetic encoding for each macroblock. is there. Therefore, it has a threshold value corresponding to the number of bits.

  The code amount comparison unit 31 compares the prediction value, which is the number of coefficient bits received from the code amount prediction unit 20, with the threshold value stored in the threshold value storage unit 32. If the prediction value is larger than the threshold value, the output selection unit 8 is instructed to select I_PCM data. When the predicted value is less than or equal to the threshold value, the output selection unit 8 is instructed to select the encoded data output from the image encoding unit 10. The subsequent data flow is as described with reference to FIG.

  As described above, the image encoding device 1 illustrated in FIG. 4 can obtain the number of bits as a predicted value when all the quantized coefficients included in one encoding target macroblock are converted into binary data. . Furthermore, I_PCM determination for each macroblock can be performed by comparing the predicted value with a threshold value.

  Note that the coefficient bit number conversion unit 21 converts the value of the quantized coefficient into the number of bits using the conversion table shown in FIG. However, the conversion may be performed by means other than the conversion table, for example, using a function that can obtain the same conversion result. In the following description, the same applies to conversion using a conversion table.

(Specific example 3)
FIG. 6 is a functional block diagram illustrating a specific example 3 of the functional configuration of the image encoding device 1.

  The functional configuration of the image encoding device 1 illustrated in FIG. 6 includes a motion vector bit number converting unit 22 and an adding unit 27 in addition to the components included in the image encoding device 1 illustrated in FIG. It is the structure which has.

  The motion vector bit number conversion unit 22 is an example of a motion conversion unit in the image encoding device of the present invention. The motion vector bit number conversion unit 22 acquires a motion vector from the encoding information output from the encoding method calculation unit 10a. This is a processing unit that converts the number of bits of binary data when binarized by the binarizer 41. For this conversion, a conversion table similar to the conversion table shown in FIG. 5 is used. That is, the motion vector bit number conversion unit 22 has a conversion table for obtaining the number of bits of one binary data from one motion vector, and the motion vector is converted into a binarizer 41 using this conversion table. Is converted into the number of bits of binary data when binarized.

  The number of bits output from the motion vector bit number conversion unit 22 is added for each macroblock in the addition unit 27. The number of bits added and output for each macroblock in the adding unit 27 shown in FIG. 6 is hereinafter referred to as “motion vector bit number”. The number of bits output from the coefficient bit number conversion unit 21 is also added for each macroblock in the addition unit 25. Further, the sum of the number of coefficient bits and the number of motion vector bits output from the addition unit 25 and the addition unit 27 is output to the I_PCM determination unit 30. That is, the sum of the number of coefficient bits and the number of motion vector bits is output to the I_PCM determination unit 30 as a predicted value.

  In the I_PCM determination unit 30, the threshold storage unit 32 stores a threshold for performing I_PCM determination using the sum of the number of coefficient bits and the number of motion vector bits. That is, it has a threshold value corresponding to the sum of the number of coefficient bits and the number of motion vector bits output from the code amount prediction unit 20 as a prediction value.

  The code amount comparison unit 31 compares the prediction value that is the sum of the coefficient bit number and the motion vector bit number received from the code amount prediction unit 20 with the threshold value stored in the threshold value storage unit 32, and the prediction value is If it is larger than the threshold value, the output selection unit 8 is instructed to select I_PCM data. When the predicted value is less than or equal to the threshold value, the output selection unit 8 is instructed to select the encoded data output from the image encoding unit 10. The subsequent data flow is as described with reference to FIG.

  As described above, the image encoding device 1 shown in FIG. 6 has the number of bits when all the quantized coefficients included in one encoding target macroblock are converted to binary data, and all the macroblocks. The number of bits when the motion vector is converted to binary data is added, and the added value is used as a predicted value. Thereby, as compared with the image encoding device 1 shown in FIG. 4, a value closer to the number of bits when the encoding target macroblock is actually binarized can be set as the predicted value. Furthermore, I_PCM determination for each macroblock can be performed by comparing the predicted value with a threshold value.

(Specific example 4)
FIG. 7 is a functional block diagram showing a specific example 4 of the functional configuration of the image encoding device 1.

  The functional configuration of the image encoding device 1 illustrated in FIG. 7 is the same as the configuration of the image encoding device 1 illustrated in FIG. 6 except that the encoded information bit number conversion unit 23 is replaced with a code amount in place of the motion vector bit number conversion unit 22. This is a configuration included in the prediction unit 20.

  The encoding information bit number conversion unit 23 acquires the encoding information output from the encoding method calculation unit 10a, and is binary when the encoding information including the motion vector is binarized by the binarizer 41. It is a processing unit that converts data into the number of bits. That is, the processing unit outputs the number of bits when the motion vector is binarized and the number of bits when the encoded information other than the motion vector is binarized. Therefore, this processing unit also serves as the motion vector bit number conversion unit 22 shown in FIG.

  Note that the encoded information bit number conversion unit 23 has a function of obtaining the number of bits after binarization from the target data, which each of the motion conversion unit and the encoded information conversion unit in the image encoding device of the present invention has. Realized.

  For the conversion of the encoded information, a conversion table similar to the conversion table shown in FIG. 5 is used. In other words, the encoded information bit number conversion unit 23 is one or more conversion tables in which the number of bits of one binary data is obtained from each piece of information such as one motion vector and one mb_type included in the encoded information. Using this conversion table, the encoded information is converted into the number of bits of binary data when binarized by the binarizer 41.

  The number of bits output from the encoded information bit number converting unit 23 is added by the adding unit 27 for each macroblock. The number of bits added and output for each macroblock in the adding unit 27 shown in FIG. 7 is hereinafter referred to as “the number of encoded information bits”. The number of bits output from the coefficient bit number conversion unit 21 is also added for each macroblock in the addition unit 25.

  Furthermore, the sum of the number of coefficient bits and the number of encoded information bits output from the addition unit 25 and the addition unit 27 is output to the I_PCM determination unit 30. That is, the sum of the number of coefficient bits and the number of encoded information bits is output to the I_PCM determination unit 30 as a predicted value.

  That is, in the image encoding device 1 shown in FIG. 6, the sum of the number of bits of binary data after binarization of each quantized coefficient and motion vector for each macroblock is used as a predicted value. The encoding apparatus 1 further converts encoding information such as mb_type other than the motion vector into the number of bits of binary data after binarization, and adds it to the prediction value for each macroblock.

  In the I_PCM determination unit 30, the threshold storage unit 32 stores a threshold for performing I_PCM determination using the sum of the number of coefficient bits and the number of encoded information bits. That is, it has a threshold value corresponding to the sum of the number of coefficient bits output from the code amount prediction unit 20 and the number of encoded information bits as a predicted value.

  The code amount comparison unit 31 compares the prediction value, which is the sum of the number of coefficient bits received from the code amount prediction unit 20 and the number of encoded information bits, with the threshold value stored in the threshold value storage unit 32, and calculates the prediction value. Is larger than the threshold value, the output selection unit 8 is instructed to select I_PCM data. When the predicted value is equal to or less than the threshold value, the output selection unit 8 is instructed to select the encoded data output from the image encoding unit 10. The subsequent data flow is as described with reference to FIG.

  As described above, the image encoding device 1 shown in FIG. 7 has the number of bits when all the quantized coefficients included in one encoding target macroblock are converted to binary data, and all the macroblocks. The number of bits when the motion vector is converted into binary data and the number of bits when all encoded information other than the motion vector of the macroblock is converted into binary data are added, and the added value is used as a predicted value. . Thereby, as compared with the image encoding device 1 shown in FIG. 6, a value closer to the number of bits when the encoding target macroblock is actually binarized can be set as the predicted value. Furthermore, I_PCM determination for each macroblock can be performed by comparing the predicted value with a threshold value.

(Specific example 5)
FIG. 8 is a functional block diagram showing a specific example 5 of the functional configuration of the image encoding device 1.

  The functional configuration of the image encoding device 1 illustrated in FIG. 8 is the same as the configuration of the image encoding device 1 illustrated in FIG. 4 except that the coefficient code amount conversion unit 24 is replaced with the code amount prediction unit 20 instead of the coefficient bit number conversion unit 21. It is the structure which has.

  The coefficient code amount conversion unit 24 is another example of coefficient conversion means in the image encoding device of the present invention, and the value of the quantized coefficient output from the quantizing unit 7 is binary. It is a processing unit that converts the code amount into a code amount when encoded by the CABAC encoder 42 after being converted into a prediction value.

  That is, unlike the code amount prediction unit 20 in each image encoding device 1 shown in FIGS. 4, 6, and 7, the code amount prediction unit 20 converts the quantized coefficient for each macroblock into binary data. Instead of outputting the number of bits as a predicted value, the code amount when the binary data for each macroblock is CABAC encoded is output as the predicted value.

  The coefficient code amount conversion unit 24 has a conversion table or a function for converting the value of one quantized coefficient into one encoded code amount. These conversion tables and functions are obtained from experimental values or theoretical values. The motion vector code amount conversion unit 26 and the encoded information code amount conversion unit 28 which will be described later with reference to FIGS. 9 and 10 also use the same conversion table or function to convert the information input to the code amount.

  The code amount output from the coefficient code amount conversion unit 24 is added for each macroblock in the addition unit 25. The code amount added and output for each macroblock in the adding unit 25 shown in FIG. 8 is hereinafter referred to as “coefficient code amount”. That is, the coefficient code amount is output to the I_PCM determination unit 30 as a predicted value.

  In the I_PCM determination unit 30, the threshold storage unit 32 stores a threshold for performing I_PCM determination using the coefficient code amount output from the code amount prediction unit 20. The prediction value output from the code amount prediction unit 20 is a prediction value of the code amount after CABAC encoding for each macroblock of the quantized coefficient, and the threshold value storage unit 32 has a threshold value corresponding to the prediction value. It is remembered. For example, the AVC / H. A value of 3200 bits or less, which is the maximum code amount of a macroblock defined by H.264, is stored as a threshold value.

  The code amount comparison unit 31 compares the prediction value, which is the coefficient code amount received from the code amount prediction unit 20, with the threshold value stored in the threshold value storage unit 32. If the prediction value is greater than the threshold value, the output selection unit 8 is instructed to select I_PCM data. When the predicted value is equal to or less than the threshold value, the output selection unit 8 is instructed to select the encoded data output from the image encoding unit 10. The subsequent data flow is as described with reference to FIG.

  As described above, the image encoding device 1 illustrated in FIG. 8 can obtain the code amount when all the quantized coefficients included in one encoding target macroblock are CABAC encoded as a predicted value. Furthermore, I_PCM determination for each macroblock can be performed by comparing the predicted value with a threshold value.

(Specific example 6)
FIG. 9 is a functional block diagram illustrating a specific example 6 of the functional configuration of the image encoding device 1.

  The functional configuration of the image encoding device 1 illustrated in FIG. 9 includes a motion vector code amount conversion unit 26 and an addition unit 27 in addition to the components included in the image encoding device 1 illustrated in FIG. It is the structure which has.

  The motion vector code amount conversion unit 26 is another example of motion conversion means in the image encoding device of the present invention, acquires a motion vector from the encoding information output from the encoding method calculation unit 10a, This is a processing unit for converting a vector into a code amount when the vector is binarized and further encoded by the CABAC encoder 42. For this conversion, a conversion table or a function is used as described above.

  The code amount output from the motion vector code amount conversion unit 26 is added by the adding unit 27 for each macroblock. The code amount added and output for each macroblock in the adder 27 shown in FIG. 9 is hereinafter referred to as “motion vector code amount”. Further, the code amount output from the coefficient code amount conversion unit 24 is also added for each macroblock in the addition unit 25. Further, the sum of the coefficient code amount and the motion vector code amount output from the addition unit 25 and the addition unit 27 is output to the I_PCM determination unit 30. That is, the sum of the coefficient code amount and the motion vector code amount is output to the I_PCM determination unit 30 as a predicted value.

  In the I_PCM determination unit 30, the threshold storage unit 32 stores a threshold for performing I_PCM determination using the sum of the coefficient code amount and the motion vector code amount. That is, it has a threshold value corresponding to the sum of the coefficient code amount and the motion vector code amount output from the code amount prediction unit 20 as a prediction value. For example, the AVC / H. A value of 3200 bits or less, which is the maximum code amount of a macroblock defined by H.264, is stored as a threshold value.

  The code amount comparison unit 31 compares the prediction value that is the sum of the coefficient code amount and the motion vector code amount received from the code amount prediction unit 20 with the threshold value stored in the threshold value storage unit 32, and the prediction value is If it is larger than the threshold value, the output selection unit 8 is instructed to select I_PCM data. When the predicted value is equal to or less than the threshold value, the output selection unit 8 is instructed to select the encoded data output from the image encoding unit 10. The subsequent data flow is as described with reference to FIG.

  As described above, the image encoding device 1 illustrated in FIG. 9 includes the code amount when all the quantized coefficients included in one encoding target macroblock are CABAC encoded, and all the motions of the macroblock. The code amount when the vector is CABAC encoded is added, and the added value is used as a predicted value. Accordingly, compared to the image encoding device 1 shown in FIG. 8, a value closer to the code amount when the encoding target macroblock is actually CABAC encoded can be set as the predicted value. Furthermore, I_PCM determination for each macroblock can be performed by comparing the predicted value with a threshold value.

(Specific example 7)
FIG. 10 is a functional block diagram illustrating a specific example 7 of the functional configuration of the image encoding device 1.

  The functional configuration of the image encoding device 1 illustrated in FIG. 10 is the same as the configuration of the image encoding device 1 illustrated in FIG. 9 except that the encoded information code amount conversion unit 28 is replaced with a code amount in place of the motion vector code amount conversion unit 26. This is a configuration included in the prediction unit 20.

  The encoding information code amount conversion unit 28 acquires the encoding information output from the encoding method calculation unit 10a, and the encoding information including the motion vector is binarized and further encoded by the CABAC encoder 42. It is a processing part which converts into the amount of codes in the case. That is, the processing unit outputs a code amount when the motion vector is CABAC-encoded and a code amount when encoding information other than the motion vector is CABAC-encoded. Therefore, this processing unit also serves as the motion vector code amount conversion unit 26 shown in FIG.

  Note that the encoded information code amount conversion unit 28 has a function of obtaining a code amount after entropy encoding from target data, which each of the motion conversion unit and the encoded information conversion unit in the image encoding apparatus of the present invention has. Realized.

  As described above, a conversion table, a function, or the like is used for conversion of the encoded information.

  The code amount output from the encoded information code amount conversion unit 28 is added by the adding unit 27 for each macroblock. The code amount added and output for each macroblock in the adding unit 27 shown in FIG. 10 is hereinafter referred to as “encoded information code amount”. Further, the code amount output from the coefficient code amount conversion unit 24 is also added for each macroblock in the addition unit 25. Further, the sum of the encoded information code amount and the motion vector code amount output from the addition unit 25 and the addition unit 27 is output to the I_PCM determination unit 30. That is, the sum of the coefficient code amount and the encoded information code amount is output to the I_PCM determination unit 30 as a predicted value.

  That is, in the image encoding device 1 shown in FIG. 9, the sum of the code amount after CABAC encoding for each macroblock of the quantized coefficient and the motion vector is used as the predicted value, but the image encoding device 1 shown in FIG. Further, encoding information such as mb_type other than the motion vector is also converted into a code amount after CABAC encoding, and added to the prediction value for each macroblock.

  In the I_PCM determination unit 30, the threshold storage unit 32 stores a threshold for performing I_PCM determination using the sum of the coefficient code amount and the encoded information code amount. That is, it has a threshold value corresponding to the sum of the encoded information code amount and the motion vector code amount output from the code amount prediction unit 20 as a prediction value.

  The code amount comparison unit 31 compares the prediction value, which is the sum of the coefficient code amount received from the code amount prediction unit 20 and the encoded information code amount, with the threshold value stored in the threshold value storage unit 32, and provides a prediction value. Is larger than the threshold value, the output selection unit 8 is instructed to select I_PCM data. When the predicted value is equal to or less than the threshold value, the output selection unit 8 is instructed to select the encoded data output from the image encoding unit 10. The subsequent data flow is as described with reference to FIG.

  As described above, the image encoding device 1 shown in FIG. 10 has the code amount when all the quantized coefficients included in one encoding target macroblock are CABAC encoded, and all the motions of the macroblock. The code amount when the vector is subjected to 2CABAC encoding and the code amount when all encoding information other than the motion vector of the macroblock is CABAC encoded are added, and the added value is used as a predicted value. As a result, as compared with the image encoding device 1 shown in FIG. 9, a value closer to the code amount when the macroblock to be encoded is actually CABAC encoded can be set as the predicted value. Furthermore, I_PCM determination for each macroblock can be performed by comparing the predicted value with a threshold value.

(Specific example 8)
FIG. 11 is a functional block diagram illustrating a specific example 8 of the functional configuration of the image encoding device 1.

  The functional configuration of the image encoding device 1 illustrated in FIG. 11 is a configuration in which the value correction unit 29 is included in the code amount prediction unit 20 in addition to the configuration of the image encoding device 1 illustrated in FIG. 10.

  The value correction unit 29 acquires a coding condition that is a condition for entropy coding of the macroblock to be coded, and corrects the prediction value obtained in the code amount prediction unit 20 according to the coding condition. It is a processing unit.

  The macroblock coding conditions include, for example, the type of picture (I / P / B) including the macroblock, the mb_type, the type of residual block (residual_block), the quantization parameter, and the target output rate. It is. Since the code amount after CABAC encoding varies depending on these encoding conditions, the value correction unit 29 corrects the predicted value according to the variation.

  Specifically, a conversion table or function for converting input information into a code amount, which each of the coefficient code amount conversion unit 24 and the encoded information code amount conversion unit 28 has, is set according to these encoding conditions. rewrite. Thereby, the predicted value is corrected.

  These encoding conditions can be obtained from, for example, picture information and slice information, and the picture information and slice information can be obtained from a slice header, SPS, and PPS. The picture information and slice information are acquired by the value correction unit 29 by a predetermined means. Further, the target output rate is input to the image encoding device 1 by the user of the image encoding device 1 by a predetermined means, and the value of the output rate is acquired by the value correction unit 29.

  Therefore, the code amount output from each of the coefficient code amount conversion unit 24 and the encoded information code amount conversion unit 28 is corrected according to the encoding condition of the macroblock to be encoded. The two code amounts are added by the adder 25 and the adder 27 and output as a coefficient code amount and an encoded information code amount. Furthermore, the sum of the coefficient code amount and the encoded information code amount is output to the I_PCM determination unit 30 as a predicted value.

  In the I_PCM determination unit 30, the threshold storage unit 32 stores a threshold for performing I_PCM determination using the sum of the coefficient code amount output from the code amount prediction unit 20 and the encoded information code amount. . For example, the AVC / H. 3200 bits, which is the maximum code amount of a macroblock defined by H.264, is stored as a threshold value.

  The code amount comparison unit 31 compares the number of coefficient bits received from the code amount prediction unit 20 with the threshold value stored in the threshold value storage unit 32. If the prediction value is larger than the threshold value, the code amount comparison unit 31 sends the I_PCM to the output selection unit 8. Instructs to select data. When the predicted value is equal to or less than the threshold value, the output selection unit 8 is instructed to select the encoded data output from the image encoding unit 10. The subsequent data flow is as described with reference to FIG.

  As described above, the image encoding device 1 illustrated in FIG. 11 performs the CABAC encoding on all the quantized coefficients included in one encoding target macroblock, similarly to the image encoding device 1 illustrated in FIG. 10. And the amount of code when all the motion vectors of the macroblock are 2CABAC-encoded and the amount of code when all the encoded information other than the motion vector of the macroblock is CABAC-encoded are added. The added value is taken as the predicted value.

  The image encoding device 1 shown in FIG. 11 can obtain the prediction value after performing correction according to the coding condition of the macroblock when the code amount prediction unit 20 obtains the prediction value. That is, the code amount prediction unit 20 can output a more accurate code amount prediction value for each macroblock. Furthermore, I_PCM determination for each macroblock can be performed by comparing the predicted value with a threshold value.

(Specific example 9)
FIG. 12 is a functional block diagram illustrating a specific example 9 of the functional configuration of the image encoding device 1.

  The image encoding device 1 shown in FIG. 12 includes a local decoded image buffer 12 instead of the I_PCM data buffer 4 included in each of the image encoding devices 1 described above. In addition, an inverse quantization unit 17 and an inverse orthogonal transform unit 16 are added to the image encoding unit 10. Although the inverse quantization unit 17 and the inverse orthogonal transform unit 16 are not illustrated in FIGS. 1 to 4 and FIGS. 6 to 11, they are components that the image encoding device 1 originally has as described above.

  In the image encoding device 1 shown in FIG. 12, the quantized coefficient data output from the quantization unit 7 is also input to the inverse quantization unit 17 and is inversely quantized. Further, the inversely quantized data is input to the inverse orthogonal transform unit 16 and subjected to inverse orthogonal transform to be a local decoded image. This local decoded image is stored in the local decoded image buffer 12. The local decoded image is acquired by the in-plane prediction unit 3 and is referred to in the in-screen prediction of the next macroblock.

  12, the functional configurations of the code amount prediction unit 20 and the I_PCM determination unit 30 are the same as those in the image encoding devices illustrated in FIGS. 1 to 4 and FIGS. 6 to 11. This is the same functional configuration as any one of the code amount prediction unit 20 and the I_PCM determination unit 30. That is, the code amount prediction unit 20 outputs the prediction value of binary data after binarization for each macroblock or the code amount after CABAC encoding, and stores the prediction value and the threshold storage unit 32. By comparing with the threshold value, I_PCM determination for each macroblock is performed.

  However, unlike the image encoding devices 1 described above, the image encoding device 1 in FIG. 12 outputs the above-described local decoded image as I_PCM data instead of the original image data of the moving image when the predicted value is larger than the threshold value. To do.

  When original image data is used as I_PCM data, since the reference image used for intra prediction of the next macroblock is changed from the local decoded image to the original image data, intra prediction of the next macroblock is started before the determination. If so, the in-screen prediction process for the next macroblock must be performed again.

  However, since the image encoding device 1 configured as shown in FIG. 12 uses a locally decoded image for I_PCM data, the pixels referred to in the intra prediction of the next macroblock need not be changed. The re-execution process of the intra prediction of the macro block is not performed, and the processing speed is improved.

  As described above, the image encoding device 1 according to the embodiment of the present invention can perform I_PCM determination for each macroblock before the data to be encoded is binarized. Further, the determination can be performed with higher accuracy by using a plurality of types of information or by using the macroblock coding conditions.

  Thus, since the I_PCM determination is performed at an earlier timing than the conventional image encoding device 1, the processing speed is improved. In addition, wasteful processing such as binarization of data that is not output as a bit stream is not performed, and the processing efficiency can be improved.

  In addition, the image encoding device 1 according to the embodiment of the present invention can more efficiently perform the process related to the motion prediction of the macroblock to be encoded, using the result of the determination performed at an earlier timing than the conventional one. it can. This also improves the processing speed of the image encoding device 1 as a whole.

  Hereinafter, as specific example 10 and specific example 11, the efficiency of motion prediction of a macroblock to be encoded will be described.

(Specific Example 10)
FIG. 13 is a diagram for explaining an example of improving the efficiency of processing related to motion prediction of a macroblock to be encoded.

  FIG. 13A is a diagram for explaining a macroblock that is referred to in the calculation of a motion vector predictor. In FIG. 13A, one grid represents one macroblock.

  For the encoding target macroblock shown in FIG. 13A, the motion vector of the left adjacent macroblock is mvLXA, the motion vector of the upper adjacent macroblock is mvLXB, and the motion vector of the upper right adjacent macroblock is mvLXC.

  The predicted motion vector mvpLX of the macroblock to be encoded is obtained using the median value of the three motion vectors mvLXA, mvLXB, and mvLXC.

  FIG. 13B is a diagram in which symbols are assigned to the plurality of macroblocks illustrated in FIG. Hereinafter, [iv] refers to a macroblock with a Roman numeral iv, and [v] refers to a macroblock with a Roman number v. Further, the following processing is performed, for example, in the image encoding device 1 shown in FIG. Also, in any of the image encoding devices 1 shown in other drawings, the I_PCM determination unit 30 has a path for giving an instruction to the motion prediction unit 2.

  First, let us consider a case where [iv] is encoded by motion prediction.

  By motion estimation, mvLX of [iv] is obtained. Since the mvLX in [iv] is the mvLXA in [v], the [v] uses the mvLX in [iv] that is the left adjacent macroblock as the mvLXA, and starts the mvpLX calculation and the motion prediction process.

  Next, when the I_PCM determination unit 30 determines that I_PCM is output as a bit stream for [iv], the mvLX of [iv] is changed. That is, since this means that the [v] mvLXA has been changed, the [v] mvpLX calculation process must be executed again.

  FIG. 13C is a diagram illustrating a time chart of processing related to motion prediction. In the figure, each of T0, T1,... Is a symbol indicating a predetermined continuous period. The outline of processing performed in each period will be described below. The predicted motion vector mvpLX is hereinafter simply referred to as “mvp”. In the following processing, the motion prediction unit 2 performs processing related to motion prediction, the DCT coefficient is calculated by the orthogonal transform unit 6, the post-quantization coefficient is calculated by the quantization unit 7, and the I_PCM determination is performed by the I_PCM determination. Part 30 performs.

  T0: [iv] processing is started, and [iv] mvp calculation and motion prediction are performed.

  Calculation of the DCT coefficient and the quantized coefficient of T1: [iv] is performed. [V] mvp calculation and motion prediction are performed.

  T2: [iv] I_PCM determination is performed. Also, the calculation of the DCT coefficient and the quantized coefficient of [v] is performed.

  Here, as a result of the I_PCM determination of [iv], when it is determined that I_PCM data is output as a bit stream for [iv], the processing of [v] is canceled and mvp is calculated again for [v]. An instruction to do so is given from the I_PCM determination unit 30 to the motion prediction unit 2.

  T3: [v] mvp calculation and motion prediction are performed. In addition, a process for entropy encoding of [iv] is performed.

  Here, in the case of the conventional image encoding device, in the above example, the I_PCM determination is performed after the processing for entropy encoding of [iv] is started. For this reason, the motion prediction for [v] again based on the result of the I_PCM determination is started at T4 at the earliest.

  However, in the image encoding device 1 according to the embodiment of the present invention, the I_PCM determination is performed before the processing for entropy encoding of [iv] is started, specifically, before the binarization processing is started. Done. Therefore, the process related to the motion estimation of [v] based on the result of the I_PCM determination can be started at T3. In addition, the calculation of mvp included in the process related to motion prediction does not use the motion vector of [iv] to be changed, and [ii] and [iii] are macroblocks of [v] upper adjacent and upper right adjacent. ] Motion vectors are used.

  That is, when the current macroblock and the next macroblock to be encoded next in the encoding order are adjacent to each other and the predicted value for the current macroblock is larger than the threshold, the motion prediction unit 2 The information of the adjacent macroblock of the macroblock is corrected, and mvp of the next macroblock is calculated. The current macroblock is, for example, the above [iv]. In this case, the next macroblock is the above [v].

  As described above, when it is determined that the I_PCM data is output as a bit stream for a certain macroblock by performing the I_PCM determination at an earlier timing than before, the process related to motion prediction is started again at an earlier timing than before. be able to. That is, the process related to motion prediction can be efficiently advanced.

  As a result, the processing speed of the image encoding device 1 as a whole can be improved.

(Specific example 11)
In the specific example 10 described above, the flow of processing for starting re-processing of motion prediction at a timing earlier than the prior art using the result of the I_PCM determination performed at an earlier timing than the prior art has been described.

  Here, as shown in the time chart of FIG. 13C, the process related to motion prediction specifically includes the calculation of an actual motion vector (mvp, which is a predicted motion vector), and the mvp and the picture to be referenced. mv) and a process of calculating a difference (mvd) between mvp and mv. Specifically, mvd is calculated using AVC / H. This is performed based on the following expression standardized to H.264.

mvLX [0] = mvpLX [0] + mvd_lX [mbPartIdx] [subMbPartIdx] [0] (Formula 1)
mvLX [1] = mvpLX [1] + mvd_lX [mbPartIdx] [subMbPartIdx] [1] (Formula 2)
It is. By transforming the equation, mvd_lX indicating the difference is obtained by the following equation.

mvd_lX [mbPartIdx] [subMbPartIdx] [0] = mvLX [0] −mvpLX [0] (Formula 3)
mvd_lX [mbPartIdx] [subMbPartIdx] [1] = mvLX [1] −mvpLX [1] (Formula 4)

  In (Expression 1) to (Expression 4), “0” in parentheses means the x-axis component of the vector, and “1” means the y-axis component of the vector. Further, “mbPartIdx” and “subMbPartIdx” are information for specifying the positions in the macroblock, respectively.

  That is, from the above (Equation 3) and (Equation 4), the predicted motion vector (mvp) and the difference (mvd) between the mvp and the actual motion vector (mv) form the following equation.

  mvd = mv−mvp (Formula 5)

  mvd is used as information indicating a motion vector. The smaller the code amount, the smaller the code amount when encoded, that is, as mv and mvp are closer as shown in (Equation 5), The code amount is reduced. Therefore, in motion prediction, a pixel block having a small difference between mvp and mv is preferentially searched for a macroblock to be encoded in a picture to be referenced.

  That is, as described above, when the mvp is calculated again according to the result of the I_PCM determination, the mv is normally changed according to the mvp changed by the redo.

  However, the image coding apparatus 1 according to the embodiment of the present invention calculates mvd using the changed mvp and the already obtained mv without changing the mv. Thereby, the motion prediction process can be completed quickly.

  FIG. 14 is a diagram illustrating a time chart of processing related to calculation of mvd. In the figure, each of T0, T1,... Is a symbol indicating a predetermined continuous period. Further, “iv” and “v” in the figure correspond to symbols given to a plurality of macroblocks shown in FIG. The outline of processing performed in each period will be described below. In the following processing, processing related to motion prediction is performed by the motion prediction unit 2, calculation of DCT coefficients is performed by the orthogonal transformation unit 6, calculation of quantized coefficients is performed by the quantization unit 7, and I_PCM determination is performed by I_PCM determination. Part 30 performs.

  T0: The process of [iv] is started, and mvp calculation and motion prediction of [iv] are performed.

  T1: The DCT coefficient of [iv], the quantized coefficient, and mvd are calculated. [V] mvp calculation and motion prediction are performed.

  T2: [iv] I_PCM determination is performed. The calculation of the DCT coefficient and the quantized coefficient of [v] is performed. Here, as a result of the I_PCM determination of [iv], if it is determined that I_PCM data is output as a bit stream for [iv], an instruction to perform recalculation of [v] in the next period T3 is an I_PCM determination unit 30 to the motion prediction unit 2.

  T3: [v] I_PCM determination and mvd recalculation are performed. Specifically, the mvd is recalculated by calculating mvp, which is a predicted motion vector, and calculating mvd by obtaining a difference between this mvp and mv already obtained in T2. In addition, processing for entropy encoding of [iv] is performed.

  Processing for entropy coding of T4: [v] is performed.

  As described above, when I_PCM determination is performed at an earlier timing than before, when it is determined that I_PCM data is output as a bit stream for a certain macroblock, it is possible to start the motion prediction re-execution at an earlier timing than before. .

  Also, when the current macroblock and the next macroblock are adjacent to each other, the motion prediction unit 2 calculates mvp, mv, and mvd for the next macroblock, and the prediction value for the current macroblock is a threshold value. If it is larger, mvp is recalculated by correcting the information of the adjacent macroblock of the next macroblock, and mvd of the next macroblock is recalculated using the mvp and mv already determined.

  That is, when calculating mvd in motion prediction, mvd is obtained from the already obtained mvp and the changed mvp without obtaining a new mv. As a result, the motion prediction for redo can be completed quickly. As a result, the processing speed of the image encoding device 1 as a whole can be improved.

  Note that the first calculation of mvd may be performed after the I_PCM determination, not before the I_PCM determination of the immediately preceding macroblock. For example, if the time chart shown in FIG. 14 is used, [v] mvd calculation may be performed at T2 after [iv] I_PCM determination.

  In this case, since it is determined in the time chart shown in FIG. 14 that I_PCM data is output as a bit stream for [iv], the calculation of mvd for [v] is performed at T2 and again at T3. . However, as described above, when the first calculation of mvd is performed after the [iv] I_PCM determination, the mvp at that time is used regardless of whether or not the mvp is changed by the [iv] I_PCM determination. What is necessary is just to calculate mvd.

  That is, when the current macroblock and the next macroblock are adjacent to each other, the motion prediction unit 2 calculates mvp and mv for the next macroblock. Also, before the mvd is calculated, the prediction value for the current macroblock is compared with the threshold value, and if the result of the comparison is that the prediction value is greater than the threshold value, the motion prediction unit 2 The information on the macroblock is corrected, mvp is calculated again, and the mvd of the next macroblock is calculated using the mvp and the already determined mv.

  Thereby, regardless of the result of the I_PCM determination of [iv], it is only necessary to calculate mvd once.

  As described above, the image encoding device 1 according to the embodiment of the present invention determines the I_PCM as described in the description of the specific example 10 also in the process related to the motion determination, specifically, in the recalculation of mvp. Can be executed quickly, from mvp recalculation to mv re-search.

  Further, as described in the description of the specific example 11, in the motion prediction in which mvd is recalculated, even when mvp is changed, mvd is calculated using the existing mv. Thereby, motion estimation can be completed early.

  In addition, since the trigger for the recalculation of mvp and the recalculation of mvd is the result output of the above-described I_PCM determination, the recalculation of mvp and the recalculation of mvd can be started at a timing earlier than before.

  As described above, the image encoding device 1 according to the embodiment of the present invention is an image encoding device that entropy-encodes binarized data, and can process image data at higher speed.

  The present invention can be used for an image coding apparatus that entropy codes binarized data. Specifically, AVC / H. H.264 is optimal as an image encoding device that outputs high-definition video or the like that is encoded by a data compression technique defined in H.264.

DESCRIPTION OF SYMBOLS 1 Image coding apparatus 2 Motion estimation part 3 In-plane prediction part 4 PCM data buffer 5 Input selection part 6 Orthogonal transformation part 7 Quantization part 8 Output selection part 9 Output part 10 Image encoding part 10a Coding system calculating part 11 Merge Unit 12 local decoded image buffer 16 inverse orthogonal transform unit 17 inverse quantization unit 20 code amount prediction unit 21 coefficient bit number conversion unit 22 motion vector bit number conversion unit 23 encoded information bit number conversion unit 24 coefficient code amount conversion unit 25 addition Unit 26 vector code amount conversion unit 27 addition unit 28 encoded information code amount conversion unit 29 value correction unit 30 PCM determination unit 31 code amount comparison unit 32 threshold storage unit 40 entropy encoding unit 41 binarizer 42 CABAC encoder

Claims (4)

Image encoding means for encoding image data using at least motion prediction ;
Entropy encoding means for binarizing and entropy encoding the encoded data obtained by the image encoding means;
Before the entropy encoding means binarizes the encoded data, the block is entropy encoded based on encoding information in a block composed of a plurality of pixels included in the encoded data . A prediction means for obtaining a prediction value indicating a code amount;
Comparing means for comparing the predicted value obtained by the predicting means with a predetermined threshold;
A determination unit that determines that I_PCM data, which is data before entropy encoding the encoded data of the predetermined unit, is output when the predicted value is larger than the threshold as a result of the comparison by the comparison unit;
When it is determined by the determination means that the I_PCM data is to be output, the I_PCM data is output, and when it is not determined that the I_PCM data is to be output, the entropy-encoded data is output by the entropy encoding means. An output means,
The image encoding means includes
As a result of the determination, when a block that is dependent on the motion prediction process of the encoding target block among the adjacent blocks adjacent to the encoding target block is output as I_PCM data, the motion vector in the encoding target block Recalculate the difference motion vector mvd in the encoding target block based on the motion vector whose value is fixed in the adjacent block and the motion vector already calculated in the encoding target block without recalculating mv An image encoding device. Image encoding means for encoding image data using at least motion prediction ;
Entropy encoding means for binarizing and entropy encoding the encoded data obtained by the image encoding means;
Before the entropy encoding means binarizes the encoded data, the block is entropy encoded based on encoding information in a block composed of a plurality of pixels included in the encoded data . A prediction means for obtaining a prediction value indicating a code amount;
Comparing means for comparing the predicted value obtained by the predicting means with a predetermined threshold;
A determination unit that determines that I_PCM data, which is data before entropy encoding the encoded data of the predetermined unit, is output when the predicted value is larger than the threshold as a result of the comparison by the comparison unit;
When it is determined by the determination means that the I_PCM data is to be output, the I_PCM data is output, and when it is not determined that the I_PCM data is to be output, the entropy-encoded data is output by the entropy encoding means. An output means,
The image encoding means includes
As a result of the determination, out of adjacent blocks adjacent to the encoding target block, a block that is dependent on the motion prediction process of the encoding target block is output as I_PCM data.
(I) The information of the neighboring block is corrected, and the predicted motion vector mvp of the encoding target block is recalculated without using the motion vector of the block having a dependency relationship with the motion prediction processing of the encoding target block. ,
(Ii) (a) the recalculated predicted motion vector mvp, and (b) a motion vector already calculated in the encoding target block obtained without recalculating the motion vector mv in the encoding target block; An image encoding device that recalculates the difference motion vector mvd in the encoding target block based on the above . Image encoding means for encoding image data using at least motion prediction ;
Entropy encoding means for binarizing and entropy encoding the encoded data obtained by the image encoding means;
Before the entropy encoding means binarizes the encoded data, the block is entropy encoded based on encoding information in a block composed of a plurality of pixels included in the encoded data . A prediction means for obtaining a prediction value indicating a code amount;
Comparing means for comparing the predicted value obtained by the predicting means with a predetermined threshold;
A determination unit that determines that I_PCM data, which is data before entropy encoding the encoded data of the predetermined unit, is output when the predicted value is larger than the threshold as a result of the comparison by the comparison unit;
If it is determined by the determination means that the I_PCM data is to be output, the I_PCM data is output, and if it is not determined that the I_PCM data is to be output, the entropy encoding data is output by the entropy encoding means. An output means,
The image encoding means includes
Before the comparison means calculates the differential motion vector mvd in the encoding target block by the image encoding means, the motion prediction of the encoding target block among the adjacent blocks adjacent to the encoding target block Perform comparisons on blocks that have processing dependencies,
As a result of the determination, when a block having a dependency relationship with the motion prediction process of the encoding target block is output as I_PCM data,
(I) The information of the adjacent block is corrected, and the predicted motion vector mvp of the encoding target block is recalculated without using the motion vector of the block having a dependency relationship with the motion prediction processing of the encoding target block. ,
(Ii) (a) the recalculated predicted motion vector mvp, and (b) a motion vector already calculated in the encoding target block obtained without recalculating the motion vector mv in the encoding target block; An image encoding device that recalculates the difference motion vector mvd in the encoding target block based on the above . Among the input image data, a storage unit that stores at least image data constituting the encoding target block ;
A vector calculating unit for calculating a predetermined using motion vectors of neighboring blocks, the motion vector of the image data constituting the encoding target block among neighboring blocks adjacent to the encoding target block,
Based on the motion vector calculated in the vector calculation unit, an image encoding unit that encodes image data constituting the encoding target block;
Select and output either the image data before the image encoding unit encodes the encoding target block stored in the storage unit or the image data encoded by the image encoding unit An output selection unit for
When the output selection unit selects and outputs the image data stored in the storage unit for the encoding target block,
The vector calculation unit
(I) without using the motion vector of the image data constituting the coding target block, predicting the motion vector of the image data constituting the next block in the coding order with respect to the coding target block;
(Ii) based on (a) the predicted motion vector and (b) the motion vector already calculated in the encoding target block, obtained without recalculating the motion vector mv in the encoding target block Calculating a difference vector that is a difference value between the predicted motion vector and the motion vector of the encoding target block;
The image encoding unit is an image encoding device that associates and outputs the difference vector to encoded image data constituting the encoding target block.