JP2010034939A - Moving image coding device and moving image coding method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent the degradation of an image quality of a moving image coding signal after a scene change. <P>SOLUTION: A coding control part controls a quantization part 20 to code a P-picture after the scene change by a code amount (quantization distribution) corresponding to I-picture when a first reference picture after the scene change is the P-picture even though the scene change is generated in GOP final M (in the example of the embodiment, M=3) picture. The total code amount of the GOP to date including the code amount generated by the coding of the P-picture added by "(M-1)×Lcbb" is determined whether exceeding the uppper limit of the total code amount required for the GOP. If it exceeds, the results of coding of the P-picture are disposed, instead, the Cbb signal of a backward prediction (or a forward prediction)/motion vector 0/prediction error 0 is output. The degradation of the moving image quality after the scene change can be prevented as keeping the generated code amount of the GOP not more than a constant value. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a moving image encoding apparatus and a moving image encoding method for encoding a moving image signal.

In a conventional moving image encoding device, when a scene change in which a scene changes is detected in a moving image signal, coarse quantization is performed on an image before the scene change, while on an image after the scene change. There is one that allocates a relatively large amount of information and contributes to an improvement in image quality after a scene change (see, for example, Patent Document 1). This is because it takes into account the general human visual characteristic that the image after the scene change remains in the viewer's consciousness than the image before the scene change.
JP 2007-94762 A

  By the way, in a moving picture coding apparatus, in order to improve the editability of a moving picture coded signal, it may be required to make code amount below a predetermined regulation value per GOP. In such a case, if a scene change occurs in the second half of the GOP, the encoding amount of the image after the scene change usually requires a code amount that is greater than the average code amount. Cannot be assigned, and the image quality of the moving image encoded signal after the scene change deteriorates.

  Accordingly, an object of the present invention is to provide a moving image encoding apparatus and a moving image encoding method capable of preventing image quality deterioration of a moving image encoded signal after a scene change.

In order to solve the above problems, a moving picture coding apparatus according to the present invention includes a coding means for coding a moving picture signal, and a scene change detection means for detecting whether or not a scene change occurs in the moving picture signal. If a scene change is detected by the scene change detecting means, it is determined whether or not the first reference picture after the scene change is detected is a P picture. If it is determined that the P picture is a P picture, Code amount control means for controlling the coding means so as to be coded with a coding distribution equivalent to that of an I picture, and a buffer for outputting a moving picture signal coded by the coding means.
Here, the scene detection means may detect whether or not a scene change occurs in the last M picture of each GOP or the last number M pictures of each GOP for each GOP of the moving image signal.
Further, the code amount control means further includes a code amount from the GOP head to the P picture including the code amount obtained by encoding the P picture with a quantization distribution equivalent to that of the I picture, and a picture for indicating the picture in advance. A value obtained by adding a minimum required code amount after the P picture obtained by multiplying a value obtained by subtracting 1 from the value of M to the code amount of the minimum signal having the minimum code length, and the total code amount required for the GOP When the added value is unlikely to exceed the total code amount upper limit value based on the upper limit value, the buffer is controlled to output the encoding result of the P picture, while it is likely to exceed Alternatively, the buffer may be controlled so as to discard the encoding result of the P picture.
Further, the code amount control means, based on the added value and the total code amount upper limit value, when the added value is likely to exceed the total code amount upper limit value required for GOP. The buffer may be controlled so as to discard the encoding result of the P picture, and the minimum signal may be output via the buffer.
Further, the code amount control means, based on the added value and the total code amount upper limit value, when the added value is likely to exceed the total code amount upper limit value required for GOP. , Controlling the buffer to discard the encoding result of the P picture, encoding the picture immediately before the scene change as a P picture, and outputting the encoding result and the minimum signal through the buffer The encoding means may be controlled to do so.
The code amount control means may further encode the P picture when the added value is unlikely to exceed the total code amount upper limit value based on the added value and the total code amount upper limit value. The buffer is controlled so as to output the result, and the encoding unit encodes the picture immediately before the scene change by quantization distribution equivalent to the B picture, and the B picture immediately before the scene change is quantized distribution equivalent to the B picture. The minimum required code amount after the B picture obtained by multiplying the code amount from the beginning of the GOP including the code amount encoded in step B to the B picture by the value obtained by subtracting 2 from the M value to the code amount of the minimum signal When the added value is unlikely to exceed the total code amount upper limit value based on the value obtained by adding the above and the total code amount upper limit value required for the GOP, The buffer is controlled to output the encoding result, and if it is likely to be exceeded, the buffer is controlled to discard the encoding result of the B picture, and the minimum signal is passed through the buffer. You may make it output.
The code amount control means further determines whether or not a picture to be encoded exists between the picture immediately before the scene change and the P picture before the scene change. May encode the picture to be encoded as a B picture and control the encoding means to output the encoding result.
Further, another moving picture coding apparatus according to the present invention divides each moving picture signal by M (M is a value of M of MN values) for each GOP of the moving picture signal. A scene change detection means for detecting whether or not a scene change occurs in each M picture, and whether or not the first reference picture after the scene change is detected is a P picture, and is determined to be a P picture. In this case, a code amount control unit that controls the encoding unit to encode the P picture with a quantization distribution equivalent to that of the I picture, and a buffer that outputs a moving image signal encoded by the encoding unit Have.
The moving picture encoding method of the present invention includes a step of detecting whether or not a scene change occurs in the moving picture signal, and an initial change after the scene change is detected when an end change is detected in the moving picture signal. Whether or not the reference picture is a P picture, and when it is determined that the reference picture is a P picture, encoding the P picture with an encoding distribution equivalent to that of the I picture, and the encoded moving picture Outputting a signal. Here, even in the moving picture coding method of the present invention, code amount control equivalent to the code amount control means in the moving picture coding apparatus may be performed.

  According to the present invention, at least a M-picture is divided within a moving image signal or for each GOP of the moving image signal, and it is detected whether or not a scene change occurs in each M picture, and a scene change is detected. In this case, it is determined whether or not the first picture after detection of the scene change is a P picture. If it is determined that the first picture is a P picture, the P picture is encoded with an encoding distribution equivalent to that of an I picture. It is possible to prevent image quality deterioration of a moving image encoded signal after a scene change.

Embodiment 1 FIG.
Embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram illustrating a configuration example of a moving image encoding apparatus according to an embodiment of the present invention.

  In FIG. 1, the moving image encoding apparatus includes a scene detection unit 10, a rearrangement unit 12, a subtracter 14, a first mode switching unit 16, a DCT unit 18, a quantization unit 20, a variable length encoding unit 22, and a buffer 24. , An inverse quantification unit 26, an inverse DDCT unit 28, an adder 30, a second mode switching unit 32, a reference image memory 34, a motion compensation prediction unit 36, and an encoding control unit 38. Needless to say, it may have other components. In this apparatus, MPEG2 or H.264 is also used. The H.264 / AVC encoding method will be described, but the present invention is not limited to this.

  Next, the operation will be described.

  When a moving image signal photographed by a video camera or the like is input to this apparatus, the scene detection unit 10 firstly, for each GOP of the moving image signal, the last M of each GOP (M is the value of M of the MN value). In the picture, it is detected whether or not a scene change in which the scene changes suddenly occurs and is output to the rearrangement unit 12. At that time, when a scene change is detected, the scene detection unit 10 outputs the scene change detection signal to the encoding control unit 38. The scene change detection method is performed based on a known method, for example, based on a change in a luminance signal or the like in a moving image signal. In the present embodiment, the GOP structure in which the MN values are N = 15 and M = 3, respectively, will be described. Also, for each GOP of the moving image signal, whether the scene detection unit 10 determines whether it is within the last M picture of each GOP (M is the value of M of the MN value), the encoding control unit 36 or the rearrangement unit. 12 may be determined by inputting a signal indicating the GOP structure, or may be determined by the rearrangement unit 12.

  Next, the rearrangement unit 12 rearranges the moving image signals in the order of encoding, and outputs them to the subtracter 14, the first mode switching unit 16, and the motion compensation prediction unit 36. Here, the rearrangement unit 12, when the DCT unit 18 is in an intra coding mode that outputs an I (intra-screen coding) picture according to the coding mode in the DCT unit 18, In the inter coding mode in which the DCT unit 18 outputs a P (forward prediction coding) picture or a B (bidirectional prediction coding) picture. Is output to the DCT unit 18 via the subtractor 14. If rearrangement is unnecessary, the rearrangement unit 12 may be omitted.

  In the case of the inter coding mode, the subtractor 14 differs from the moving image signal from the rearrangement unit 12 and the prediction image from the motion compensation prediction unit 36 (residual) according to the coding mode in the DCT unit 18. The difference signal obtained by taking is output to the first mode switching unit 16.

  The first mode switching unit 16 uses the moving image signal from the rearranging unit 12 as it is in the intra coding mode, or in the sub coding unit 14 in the inter coding mode, according to the coding mode in the DCT unit 18. The predicted image and the difference signal obtained by the difference are switched and output to the DCT unit 18.

  The DCT unit 18 DCT-transforms the moving image signal from the first mode switching unit 16 according to the encoding mode, and outputs it to the quantization unit 20.

  The quantization unit 20 quantizes the DCT coefficient from the DCT unit 18 by the quantization step instructed from the encoding control unit 38 and outputs the quantized coefficient to the variable length encoding unit 22 and the inverse quantification unit 26. To do.

  The variable length coding unit 22 performs variable length coding on the quantization coefficient, outputs the quantized coefficient to the buffer 24, and outputs the coded coefficient as a coded bit string to the outside.

  On the other hand, the inverse quantification unit 26 inversely quantizes the quantization coefficient from the quantization unit 20, opposite to the quantization in the quantization unit 20, restores the DCT coefficient, and outputs it to the inverse DDCT unit 28.

  In the inverse DCT (IDCT) unit 28, the restored DCT coefficient from the inverse quantification unit 26 is subjected to inverse DCT (IDCT) conversion to be a decoded signal. Depending on the encoding mode of the DCT unit 18, an adder 30 and Output to the second mode switching unit 32. That is, in the intra coding mode, the signal is output to the second mode switching unit 32 without going through the adder 30, and in the inter coding mode, the adder 30 is output.

  The adder 30 adds the decoded signal from the inverse DCT (IDCT) unit 28 and the prediction signal from the prediction signal motion compensation prediction unit 36 that is the same as the prediction signal input to the subtracter 14 only in the inter coding mode. Then, it is output to the second mode switching unit 32 as a restored image signal.

  The second mode switching unit 32 outputs the restored image signal to the reference image memory 34 according to the encoding mode. The reference image memory 34 stores the restored image signal from the second mode switching unit 32 as a reference image.

  In the inter coding mode, the motion compensation prediction unit 36 inputs the input moving image signal from the rearrangement unit 12 and searches the reference image memory 34 for a reference image having the smallest difference in the input moving image signal. Then, motion compensated prediction for outputting a predicted image and a motion vector is performed. Note that the predicted image subjected to motion compensation prediction is output to the subtractor 14 and the adder 30, and the motion vector is output to the variable length coding unit 22.

  Based on the scene change detection signal from the scene change detection unit 10, the encoding control unit 38 outputs a quantization control signal to the quantization unit 20 as will be described later, while the buffer 24 stores the encoding result buffer. Outputs control signals and the like.

  FIG. 2 is a diagram illustrating a coding order of pictures in the GOP when a scene change does not occur.

  FIG. 2 shows an example of a GOP structure with MN values of N = 15 and M = 3, for example, but the present invention is not limited to this GOP structure. Note that the N value of the MN value indicates the number of pictures constituting the GOP, and the M value indicates the P picture repetition period, that is, (M−1) indicates the number of B pictures between the P pictures.

  First, as shown to Fig.2 (a), the input moving image signal (INPUT) has the images 0-14, for example.

  Next, the rearrangement unit 12 rearranges the input moving image signal shown in FIG. 2A in the encoding order so that the intra picture comes first as shown in FIG. The encoding unit 20 and the variable length encoding unit 22 encode the input moving image signal in the rearranged order. In FIG. 2B, I picture 2 (I2), B picture 0 (B0), B picture 1 (B1), P picture 5 (P5), B picture image 3 (B3), B picture 4 (B4), Sorted in the order ...

  Then, as shown in FIG. 2C, the buffer 24 holds the encoded images from the variable length encoding unit 24 in the same arrangement as in FIG. 2B, and outputs them in an arrangement with a delay of one timing. To do.

  On the other hand, in a moving picture decoding apparatus (not shown), an encoded picture sequence as shown in FIG. 2 (c) output from the moving picture encoding apparatus is input and decoded as shown in FIG. 2 (d). As shown in FIG. 2 (e), the decoded signals are rearranged in the reproduction order and output.

  The above is the encoding order of each picture when no scene change occurs in the present apparatus.

  FIG. 3 is a diagram illustrating the coding order of pictures in the GOP when a scene change occurs. In the first embodiment, it is assumed that a scene change occurs in the last M of the GOP (M = 3 in the example of the present embodiment).

  That is, as shown in FIG. 3A, for example, the processing of this embodiment when a scene change is detected between the B picture 13 (B13) and the P picture 14 (P13) is shown. .

  That is, the scene change detection unit 10 monitors a change in luminance between images, and when the luminance exceeds a predetermined threshold, a flag Fcng indicating that a scene change has been detected is set to 1, The data is output to the encoding control unit 38. Here, as shown in FIG. 3A, since it is detected that a scene change has occurred between the B picture 13 (B13) and the picture 14 (P13), the scene change detection unit 10 performs the process shown in FIG. As shown in b), the post scene change detection flag Fcng is set to 1 after the P picture 14 (P13).

  The encoding control unit 38 performs the following procedure for each picture type based on the state of the scene change detection flag Fcng from the scene change detection unit 10, that is, whether Fcng is 0 or 1. Code amount control is performed. As a result, the DCT unit 18, the quantization unit 20, and the variable length coding unit 22 perform the coding in the arrangement of images as shown in FIG.

(1) I Picture Code For the I picture, the coding control unit 38 performs coding as usual regardless of the value of the scene change detection flag Fcng, and sets the scene change detection flag Fcng to 0 after the completion of coding. To do. That is, the scene change detection flag Fcng is always reset by encoding the I picture.

(2) Coding of P picture For P pictures, the coding control unit 38, when the scene change detection flag Fcng value is 0, as usual, that is, by a code amount (quantization distribution) equivalent to a normal P picture. A quantization control signal is sent to the quantization unit 20 so as to perform encoding.

  On the other hand, when the scene change detection flag Fcng value is 1, that is, when the first reference picture after the scene change is a P picture, the encoding control unit 38 converts the corresponding P picture after the scene change into an I picture. A quantization control signal is sent to the quantization unit 20 so as to perform encoding with a considerable amount of code (quantization distribution). It should be noted that the code amount (quantization distribution) is the largest for I pictures, then P pictures and B pictures in this order, and the code of each picture constituting the GOP so as not to exceed the upper limit of the total GOP code amount. The amount is decided. As a result, the P picture after the scene change can be encoded by quantization distribution equivalent to the I picture, that is, encoding distribution equivalent to the I picture, so that a sufficient code amount can be allocated to the P picture. Further, it is possible to prevent the deterioration of the image quality of the moving image encoded signal after the scene change.

  Next, the encoding control unit 38 becomes the minimum codeword for the B picture as a minimum signal with a fixed code length that is not dependent on the input moving image signal, and becomes the minimum codeword for the B picture (future) or before (before) The code length of the C picture of backward prediction (or forward prediction) / motion vector 0 / prediction error 0 indicating that the past image is used as it is is assumed to be Lcbb, and the code amount generated by the encoding of this P picture is It is determined whether or not the value obtained by adding “(M−1) × Lcbb” to the total GOP code amount to date including the present exceeds the upper limit of the total code amount required for the GOP. This is because the P picture 14 (P14), which is the P picture immediately after the occurrence of the scene change, is followed by the encoding of the B picture 12 (B12) and the B picture 13 (B13). Whether or not picture 14 (P14) is encoded with a code amount equivalent to an I picture is determined whether or not the GOP total code amount upper limit value is not exceeded.

  In this embodiment, since M = 3, the encoding control unit 38 determines whether or not the value obtained by adding “2 × Lcbb” exceeds the upper limit of the total code amount required for the GOP. If the P picture encoding result is discarded, the P picture skip flag FPskip indicating that the P picture encoding result is discarded is set to 1, and the scene change detection flag Fcng is set to 0. . On the other hand, if not exceeding, the encoding result of the P picture is adopted as it is, the P picture skip flag FPskip is set to 0, and the scene change detection flag Fcng is set to 0.

  Since FIG. 3 shows a case where the value obtained by adding “2 × Lcbb” exceeds the GOP total code amount upper limit value, the encoding control unit 38 encodes the P picture 14 (P14), In contrast, a buffer control signal is transmitted so as to discard the encoding result of P picture 14 (P14), while the encoding result of P picture 14 (P14) is discarded as shown in FIG. While the P picture skip flag FPskip is set to 1, the scene change detection flag Fcng is returned to 0 as shown in FIG.

  Thereby, when the P picture skip flag FPskip is 1, the buffer 24 uses the buffer control signal from the encoding control unit 38 to change the encoding result of the P picture 14 (P14) as shown in FIG. Discard.

  In addition, when the P picture skip flag FPskip is 1, that is, when the encoding result of the P picture 14 (P14) is discarded as shown in FIG. Since there are no reference images for B picture 12 (B12) and B picture 13 (B13), B picture 13 (B13) immediately before the scene change is converted to P picture 13 (P13) as shown in FIG. Then, a quantization control signal is sent to the quantization unit 20 so as to be quantized with a code amount equivalent to a P picture, and the quantization unit 20 quantizes the DCT coefficient based on the quantization control signal.

(3) Encoding B picture For B pictures, the encoding control unit 38 instructs the quantization unit 20 to quantize with a code amount equivalent to a normal B picture when the P picture skip flag FPskip is 0. The quantization control signal is sent, and the quantization unit 20 quantizes the DCT coefficient based on the quantization control signal.

  Next, the encoding control unit 38 adds (M−2) × Lcbb, here 1 × Lcbb, to the total code amount of the GOP up to now including the code amount generated by encoding the B picture. It is determined whether or not the value exceeds the total code amount upper limit value required for the GOP. When (M−2) × Lcbb exceeds the GOP total code amount upper limit value, the encoding result of the B picture is discarded and the B picture skip flag FBskip indicating the result is set to 1. On the other hand, when (M−2) × Lcbb does not exceed the GOP total code amount upper limit value, the B picture encoding result is used as it is, and the B picture skip flag FBskip indicating the result is set to 0. In the example shown in FIG. 3, the latter (M−2) × Lcbb does not exceed the GOP total code amount upper limit value.

  When the P picture skip flag FPskip is 1, if there is a picture to be encoded between the picture immediately before the scene change and the previous P picture that has not been discarded, the picture is encoded, and then the code is encoded. After adding the conversion column Cbb, the P picture skip flag FPskip is returned to 0.

  For example, in the case of FIG. 3, the B picture 12 (B12) exists between the B picture 13 (B13) immediately before the scene change and the previous P picture 11 (P11) that has not been discarded. As shown in (f), the B picture 12 (B12) is encoded, after which the encoded sequence Cbb is added, and then the P picture skip flag FPskip is returned to 0.

  Although not shown, when the P picture skip flag FPskip is 0 and the B picture skip flag FBskip is 1, the encoding result of the B picture 12 (B12) is also discarded after the P picture (P13). After the P picture (P13), instead of the encoding result of the B picture 12 (B12), the backward prediction (or forward prediction) / motion vector 0 / prediction error 0 Cbb signal is output twice in succession. Thus, the P picture skip flag FPskip is returned to zero. In this case, for example, a case where a scene change is detected between a B picture (B12) and a P picture (P13) is applicable.

  When both the P picture skip flag FPskip and the B picture skip flag FBskip are 0, the B picture is encoded according to a normal procedure. Then, it is determined whether or not the total code amount up to the present including the code amount generated by encoding the last B picture of the GOP exceeds the upper limit of the total code amount of the entire GOP. The encoding result of the last B picture is discarded, and instead a backward prediction (or forward prediction) / motion vector 0 / prediction error 0 Cbb signal is added. The picture encoding result is output as it is.

  Therefore, according to the moving picture coding apparatus of the first embodiment, even if a scene change occurs in the last M of the GOP (M = 3 in the example of the present embodiment), a scene change occurs. When the subsequent first reference picture is a P picture, the encoding control unit 38 causes the quantization unit 20 to encode the P picture after the scene change with a code amount (quantization distribution) equivalent to the I picture. Since the control is performed, a sufficient amount of code can be assigned to the P picture after the scene change, and deterioration of the image quality of the moving image encoded signal after the scene change can be prevented.

  Further, in the first embodiment, when a scene change occurs in the last M picture of the GOP (M = 3 in the example of the present embodiment), the code amount generated by encoding the P picture is changed. It is determined whether or not the value obtained by adding “(M−1) × Lcbb” to the total code amount of the GOP including the present exceeds the upper limit of the total code amount required for the GOP. In this case, the encoding result of this P picture is discarded, and instead, backward prediction (or forward prediction) / motion vector 0 / prediction error 0 Cbb signal, that is, a backward (future) or previous (past) image Since the Cbb signal indicating that the signal is used as it is is output, it is possible to prevent the deterioration of the image quality of the moving image after the scene change and to reduce the generated code amount of GOP to a certain value or less.

  Further, according to the first embodiment, a large amount of resources such as memory is not required without performing complicated processing that changes the GOP structure.

  In the above embodiment, the case where the MN values are N = 15 and M = 3 has been described. However, in the present invention, the MN value is not limited to this, and other MN values may of course be used. The same applies to the second embodiment.

  In the first embodiment, as a minimum signal to be inserted as a B picture, backward prediction (or forward prediction) / motion vector 0 / prediction error 0 Cbb signal, that is, backward (future) or forward (past) However, the present invention is not limited to this, and averaging (1) is performed using both the subsequent (future) and previous (previous) images. Of course, it may be anything as long as it is a signal of a coded bit string having a short code length that does not depend on the input video signal. The same applies to the second embodiment.

  Further, in the first embodiment, the code amount from the GOP head to the P picture including the code amount obtained by encoding the P picture with the same encoding distribution as the I picture, and the code length for indicating the picture in advance The total code amount upper limit required for the GOP is a value obtained by adding the minimum necessary code amount after the P picture obtained by multiplying the value of the minimum signal by 1 or 2 from the value of M to the minimum signal amount. In the present invention, the present invention is not limited to this. However, the present invention is not limited to this, and a difference between the added value and the total code amount upper limit value is obtained, and whether or not the difference exceeds a predetermined threshold value. Of course, it may be determined that the added value is likely to exceed the total code amount upper limit value or not. The same applies to the second embodiment.

  Further, in the first embodiment, when the added value exceeds the total code amount upper limit value required for the GOP based on the addition value and the total code amount upper limit value, Although it has been described that the buffer 24 is controlled so as to discard the encoding result and the minimum signal Cbb is output via the buffer 24, the present invention is not limited to this, and the encoding of the P picture is not limited thereto. Only by controlling the buffer 24 to discard the result, the minimum signal Cbb does not have to be output via the buffer 24 unless the GOP structure is destroyed later. The same applies to the second embodiment.

  Further, in the first embodiment, for each GOP of the moving image signal, it has been described that whether or not a scene change occurs in the last M picture of each GOP (M is the value of M of the MN value) is detected. However, in the present invention, the present invention is not limited to this, and may be within the number of M pictures at the end of the GOP, or even at the beginning of the GOP, near the center, or anywhere within the GOP, not the last M picture of the GOP. .

Embodiment 2.
In the first embodiment, for example, when a scene change occurs in the last M of the GOP (M = 3 in the example of the present embodiment), the total code amount is the upper limit of the GOP in GOP units. Although it was explained not to exceed, it is normal not to know when a scene change occurs.

  Therefore, in the second embodiment, the GOP is divided into M values so that even if a scene change occurs in M value units, the total code amount does not exceed the upper limit value of GOP in M value units.

  For example, in the same manner as in the first embodiment, when the MN values of the GOP are N = 15 and M = 3, respectively, when the 1 GOP is divided into every 3 pictures of the M value indicating the appearance ratio of the P picture, these 3 pictures Of course, the total amount of codes from the beginning of the GOP is determined as one unit for each (frame), and the same processing as in the first embodiment may be performed using one unit for every three frames.

  Therefore, according to the second embodiment, as in the first embodiment, it is possible to prevent image quality deterioration of a moving image after a scene change and to reduce the generated code amount of GOP to a certain value or less.

  In the second embodiment, the GOP is divided into M values, and even if a scene change occurs in units of M pictures, the total code amount does not exceed the upper limit value of GOPs in units of M pictures. Even if a scene change occurs in any part of the picture, the same processing as in the first embodiment is performed in the M picture in which the scene change has occurred, and the subsequent picture coding is not affected. Even if a scene change occurs in the first or first half of the GOP, it is possible to reduce pictures that are adversely affected by image quality deterioration due to the scene change.

It is a block diagram which shows the structural example of the moving image encoder which is embodiment of this invention. It is a figure which shows the encoding order etc. of the picture in GOP when a scene change does not occur. It is a figure which shows the encoding order etc. of the picture in GOP when a scene change generate | occur | produces. It is a figure which shows an example of the table which recorded the imaging | photography information extracted.

Explanation of symbols

  10 scene detection unit, 12 rearrangement unit, 14 subtractor, 16 first mode switching unit, 18 DCT unit, 20 quantization unit, 22 variable length coding unit, 24 buffer, 26 inverse quantification unit, 28 inverse DDCT Units 28, 30 adder, 32 second mode switching unit, 34 reference image memory, 36 motion compensation prediction unit, 38 encoding control unit.

Claims (8)

Encoding means for encoding a moving image signal;
Scene change detection means for detecting whether or not a scene change occurs in the moving image signal;
When a scene change is detected by the scene change detection means, it is determined whether or not the first reference picture after the scene change is detected is a P picture. A code amount control means for controlling the coding means to perform coding with a coding distribution equivalent to that of a picture;
A buffer for outputting a moving image signal encoded by the encoding means;
A moving picture encoding apparatus having: The moving image encoding device according to claim 1,
The code amount control means includes:
further,
The code amount from the GOP head to the P picture including the code amount obtained by encoding the P picture with the same quantization distribution as that of the I picture, and the code of the minimum signal with the minimum code length for indicating the picture in advance Based on a value obtained by adding a minimum required code amount after the P picture obtained by multiplying a value obtained by subtracting 1 from the value of M to the amount, and the total code amount upper limit value required for the GOP, the added value Is not likely to exceed the total code amount upper limit value, the buffer is controlled so as to output the encoding result of the P picture, while if it is likely to exceed, the encoding result of the P picture is A moving picture coding apparatus for controlling the buffer to be discarded. The moving image encoding device according to claim 2,
The code amount control means includes:
Based on the added value and the total code amount upper limit value, if the added value is likely to exceed the total code amount upper limit value required for the GOP, the encoding result of the P picture is A moving picture coding apparatus that controls the buffer to be discarded and further outputs the minimum signal through the buffer. The moving image encoding device according to claim 2,
The code amount control means includes:
Based on the added value and the total code amount upper limit value, if the added value is likely to exceed the total code amount upper limit value required for the GOP, the encoding result of the P picture is The buffer is controlled to be discarded, and the picture immediately before the scene change is encoded as a P picture, and the encoding means is controlled to output the encoding result and the minimum signal through the buffer. , Moving picture coding apparatus. The moving image encoding device according to claim 2,
The code amount control means includes:
further,
Based on the added value and the total code amount upper limit value, when the added value is unlikely to exceed the total code amount upper limit value, the buffer is controlled to output the encoding result of the P picture. The GOP head including the code amount obtained by encoding the picture immediately before the scene change by the encoding means with the quantization distribution equivalent to the B picture and encoding the B picture immediately before the scene change with the quantization distribution equivalent to the B picture. A value obtained by adding the code amount from the B picture to the B picture, the minimum required code amount after the B picture obtained by multiplying the code amount of the minimum signal by the value obtained by subtracting 2 from the M value, and GOP Based on the requested total code amount upper limit value, if the added value is unlikely to exceed the total code amount upper limit value, the buffer outputs the encoding result of the B picture. While control, if exceeded likely controls the buffer to discard the encoded result of the B-picture, and the minimum signal to output through the buffer, the moving picture coding apparatus. The moving image encoding device according to claim 4,
The code amount control means includes:
further,
It is determined whether or not there is a picture to be encoded between the picture immediately before the scene change and the P picture before the scene change, and if so, the picture to be encoded is determined as a B picture. A moving picture coding apparatus for controlling the coding means to perform coding and to output the coding result. Encoding means for encoding a moving image signal;
Scene change detecting means for dividing each MOP of a moving image signal for each M (M is the value of M of MN values) and detecting whether or not a scene change occurs in each M picture;
It is determined whether or not the first reference picture after detection of a scene change is a P picture. If it is determined that the P picture is a P picture, the encoding is performed so that the P picture is encoded with a quantization distribution equivalent to that of an I picture. Code amount control means for controlling the means;
A buffer for outputting a moving image signal encoded by the encoding means;
A moving picture encoding apparatus having: Detecting whether a scene change occurs in the moving image signal;
If an end change is detected in the video signal, it is determined whether or not the first reference picture after the scene change is detected is a P picture. If it is determined that the P picture is a P picture, the P picture is referred to as an I picture. Encoding with an equivalent encoding distribution;
Outputting the encoded video signal;
A video encoding method comprising:

Images