JP2002010271A - Device for coding video - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device for coding video such that a special buffer for determining whether or not the entire picture is to be skipped becomes unnecessary and the sequence of video is not reversed even if the entire picture is skipped. SOLUTION: A comparative determination means 125 compares a predicted amount of accumulation in the receiving buffer of this decoding device with a threshold set on each of picture types, on every single picture data before the picture data is decoded. The comparative determination means 125 determines that a coding processing of image data is to be skipped, making DCT means to stop the coding processing, when the predicted amount of accumulation in the receiving buffer is under the threshold. Further, the comparative determination means 125 produces the outputs of all skip pictures such that all remaining macro-blocks except the top and bottom of a slice layer in a skip picture storing memory 126 consist of the skipped macro-blocks.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video encoding apparatus, and more particularly, to a video encoding apparatus that skips encoding of image data and adjusts a code amount.

[0002]

2. Description of the Related Art In video encoding based on the MPEG standard, the amount of code is controlled while predicting the amount of stored pictures in a receiving buffer of a decoding device. This is VBV
(Video Buffering Verification
r) Code amount control by model.

FIG. 20A shows a normal transition of the estimated accumulation amount of the reception buffer. As shown in the drawing, pictures are input to the reception buffer at a constant rate. Then, DTS (Decoding Time Stamp)
At the time indicated by p), one picture is output from the reception buffer for decoding. When the decoding device has a display unit of the NTSC system, the DTS is 1/30 second if one frame is allocated to one picture, and 1/60 if one field is allocated to one picture. Set every second. Also, when the decoding device has a display unit of the PAL system, the DTS is one every 1/25 second when one frame is allocated to one picture, and when one field is allocated to one picture. / 50 seconds.

[0004] Normally, as shown in FIG.
Assuming that the bit generation amount of the picture decoded in TS1 is D1, in DTS1, the predicted storage amount of the reception buffer decreases from V1 to V1 ^* (= V1−D1). V
In the bV model, the code amount is controlled so that the predicted accumulation amount of the reception buffer does not cause overflow or underflow.

At the time of a scene change or the like, FIG.
As shown in (b), the reception buffer may underflow due to the continuation of pictures having a large bit generation amount. That is, assuming that the bit generation amount of the picture decoded by DTS3 is D3,
From V3−D3 <0, the predicted accumulation amount of the reception buffer is less than 0. This is because the picture decoded by DTS3 has not been input to the reception buffer yet.
In such a case, underflow is avoided by increasing the quantization scale to reduce the amount of generated bits.

[0006] As shown in FIG. 20 (c), the reception buffer may overflow due to consecutive pictures having a small bit generation amount. That is, before the predicted accumulation amount of the reception buffer is decoded by DTS3, V
3 (= V2 ^* + R, R is the amount of bits input to the receiving buffer of the decoding device for each DTS interval (decoding time interval of one picture), in other words, the bits transmitted from the encoding device for each DTS interval V3), and V3 exceeds the reception buffer capacity. In such a case, overflow is avoided by reducing the quantization scale and increasing the bit generation amount.

However, as described above, when the quantization scale is increased to prevent the underflow and the amount of generated bits is extremely reduced, the image quality is deteriorated. Therefore, the following methods have been used together with the adjustment of the quantization scale. The first method is
This is a method using a so-called skipped macro block. In MPEG, encoding processing is performed in units of macroblocks, which are 16 × 16 pixel blocks. A skipped macroblock is a macroblock consisting of a special code that defines that the same image as the reference image should be displayed at that position, and has a very small data amount. Therefore,
If an underflow is likely to occur, the skipped macroblock may be transmitted without encoding the original image.

However, according to this method, an original image portion is displayed at a position of a macroblock which is not a skipped macroblock, and a reference image portion is displayed at a position of a skipped macroblock. Become
An image that does not match as a whole is obtained. FIG. 21A shows an example of a screen normally displayed.

FIG. 21B shows an example of a screen displayed when a skipped macroblock is used. In the second frame, the upper half macroblock uses a macroblock that is not a skipped macroblock, so the original image portion of the second frame is displayed. Since the lower half macroblock uses a skipped macroblock, The image portion of the first frame is displayed, and the image does not match as a whole.

[0010] The second method is a method of encoding pseudo image data. Pseudo image data is image data that is the median of the range of pixel values. For example, if the pixel value is represented by 8 bits, the median is 12
It becomes 8. In MPEG, the difference value between the pixel value and the median value of the image data is encoded, so the amount of encoded data of the image data whose image value is the median value is minimum. Therefore, when underflow is likely to occur for each macroblock, image data having a median pixel value may be encoded without performing the original encoding.

However, according to this method, a gray image is displayed at the position where the pixel value is set to the median value. FIG.
(C) shows an example of a screen displayed when a pseudo image is used. In the second frame, the macroblock in the upper half is subjected to the original image encoding, so that the original image portion of the second frame is displayed. The macroblock in the lower half is encoded with the pseudo image data, so that the gray color is gray. Is displayed, and the image is not matched as a whole.

Japanese Patent No. 2871316 describes a method of skipping the encoding of the entire image data of one field or one frame (hereinafter referred to as skipping a picture). FIG.
1 shows a configuration of a video encoding device described in the above issue. To explain the outline of the video encoding apparatus, encoding processing is performed on input moving image data by an orthogonal transformation circuit 6 or the like, and pictures are stored in a buffer memory 21. When the transmission rate excess determination circuit 24 determines that the transmission rate exceeds a certain value, the SKIP code in the SKIP code storage memory 22 is output, and when it is determined that the transmission rate does not exceed , The picture in the buffer memory 20 is output.

As described above, according to this method, when skipping, not only some macroblocks of a picture are skipped macroblocks, but also the rest of the picture except for the first and last slice layers. All of the macroblocks are skipped macroblocks. As a result, in the decoding device, the same image as the whole of the image decoded in the past is displayed, and the inconsistent image as described above is not displayed.

[0014]

However, the method described in the above-mentioned patent publication has the following problems. The first problem is to determine whether or not the amount of generated pictures exceeds the transmission rate, as shown in FIG.
This requires a buffer memory 20 for temporarily storing the generated pictures. That is, as described above, when determining whether or not to skip for each macroblock, a small-capacity buffer for storing encoded data in units of macroblocks should have been provided. Since it is determined whether or not to skip the entire picture, a large buffer capacity for storing the code amounts of all macroblocks in the picture is required.

A second problem is that, when one frame is assigned to one picture (frame structure), if the picture is skipped, the display order may be reversed when displayed by the interlaced scanning method. It is. FIG.
(A) shows a normal display screen in which pictures are not skipped. 1t and 1b represent a top field and a bottom field of the first frame, respectively. In the frame structure, encoding processing is performed in units of one frame including a top field and a bottom field. In this case, the decoding device performs the decoding process every 1/30 second in units of one frame. Since the display is performed in the interlaced mode, each field in the frame is displayed every 1/60 second. That is, 1t, 1b, 2t, 2
b, 3t, 3b, 4t, and 4b.

FIG. 23B shows a display screen when a picture skip occurs. If a skip occurs in B (3), the same image as I (1) referred to by B (3) is displayed in frame 2 encoded in B (3). Therefore, the top field of the frame 2 is the same as the top field 1t of the frame 1, and the bottom field of the frame 2 is the same as the bottom field 1b of the frame 1. Then, in order to display in an interlaced manner, every 1/60 second,
It is displayed in the order of 1t, 1b, 1t, 1b, 3t, 3b, 4t, 4b. Thereby, 1t is displayed after 1b, and the display order is reversed.

Therefore, the present invention eliminates the need for a special buffer for determining whether or not to skip the entire picture and, even when the entire picture is skipped,
It is an object of the present invention to provide a video encoding device whose display order is not reversed.

[0018]

SUMMARY OF THE INVENTION To achieve the above object, the present invention provides a video encoding system for encoding one frame or one field of image data while estimating a storage amount of a receiving buffer in a decoding device. The apparatus, before encoding the image data, compares a predicted accumulation amount predicted to be accumulated in a reception buffer at a decoding time of the encoded data of the image data with a predetermined threshold. And comparing means, when the predicted accumulation amount is less than the threshold, stops encoding of the image data, and as data taken out from the reception buffer at the decoding time, instead of the encoded data of the image data. And a skip unit that uses a proxy code that specifies that the same image data as previously decoded image data should be displayed.

Further, the present invention relates to a video encoding apparatus for encoding one frame of image data in a frame structure, wherein when encoding the image data, the amount of storage in an output buffer or the reception buffer of the decoding apparatus. Comparing means for comparing the predicted accumulation amount with a predetermined reference value, and encoding the image data in a frame structure when the accumulation amount or the predicted accumulation amount is determined to have reached a predetermined reference value. It is determined that two fields which are the same as the top field or the bottom field constituting the image data decoded in the past should be displayed instead of the encoded data of the top field and the bottom field of the image data. And a skip unit using a proxy code.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. <First Embodiment> In this embodiment, before generating a picture, whether to skip a picture based on whether or not a predicted code amount based on an empirical value for each picture type exceeds a predicted storage amount of a reception buffer. The present invention relates to a video encoding device for determining whether or not to perform the determination. (Configuration of Video Encoding Apparatus) FIG. 1 shows a configuration of a video encoding apparatus 100 according to the present embodiment. The video encoding device 100 includes a screen rearrangement unit 110, a DCT unit 113, a quantization unit 114, and a rate control unit 115
, Variable length coding means 116, buffer 117, inverse quantization means 119, inverse DCT means 120, video memory 122, motion compensation prediction means 123, adder 11
1, 112, a switch 118, a receiving buffer storage amount predicting unit 124, a comparing and determining unit 125, a SKIP picture storage memory 126, and a threshold setting unit 127.

The screen rearranging means 110 rearranges the screens in an encoding order according to the picture type. FIG.
(A) shows the original order of the video images and thus the order in which they are displayed on the decoding device. FIG. 2 (b) shows the order of encoding, and thus the order of input to the decoding device. B
A picture is encoded using a picture of a preceding or succeeding I or P picture, and thus is encoded after encoding a future P picture.

The DCT means 113 performs a discrete cosine transform (DCT) operation on a macroblock basis and outputs DCT coefficients. Here, for an I (Intra coded) picture, a DCT operation is performed in an intra coding mode, and a P (Predictive coded) picture and a B (Bidirectionally pred) are obtained.
For an coded picture, a DCT operation is performed by selecting an intra coding mode or a motion compensation prediction mode in macroblock units. In the case of the intra coding mode, the input original image is directly subjected to DCT.
Calculate. In the case of the motion compensation prediction mode, the difference between the original image and the prediction image obtained by the motion compensation prediction unit 123 is subjected to DCT calculation.

The quantization means 114 performs quantization of DCT coefficients by changing the quantization scale for each macroblock. The variable-length coding unit 116 performs variable-length coding on the quantized DCT coefficients together with the motion vector and coding prediction mode information to generate coded data of image data.

The buffer 117 stores coded data of variable-length coded image data or all-skip pictures. Inverse quantization means 119 and inverse DCT means 120
Since a decoded image of an I or P picture needs to be used as a reference image for motion compensation prediction, a decoded image obtained by performing inverse quantization and inverse DCT from a quantized DCT coefficient is restored. And outputs it to the video memory 122.

The video memory 122 stores a decoded image of an I or P picture as a reference image. The motion compensation prediction unit 123 outputs a prediction image using a motion vector from a reference image stored in the video memory 122 for encoding a P or B picture. Rate control means 15
Is the quantization means 1 according to the estimated accumulation amount of the reception buffer.
14 is instructed to change the quantization scale. That is, the rate control unit 115 instructs the quantization unit 114 to decrease the quantization scale when the predicted accumulation amount of the reception buffer is equal to or more than a certain amount, and to increase the quantization scale when the predicted accumulation amount is equal to or less than the certain amount. I do.

The SKIP picture storage memory 126 stores P
Stores all skip pictures for pictures and B pictures. Here, the all-skip picture means a picture in which all the remaining macroblocks except the first and last slice layers in one picture are composed of skipped macroblocks. Instead of a B picture and a P picture, a B picture and a P picture (all of which are all skipped B pictures and all skips, respectively) in which all the remaining macroblocks except the beginning and end of the slice layer are skipped macroblocks are used. (Called a P-picture). Since an I picture cannot be skipped (a macroblock cannot be a skipped macroblock), an all-skip P picture is transmitted instead of an I picture. These all skip pictures are transmitted in place of the encoded data of the image data, and are so-called proxy codes of the encoded data of the image data.

The reception buffer storage amount predicting means 124
Every time a picture is output, the predicted accumulation amount of the reception buffer before decoding at the decoding time of the next picture is calculated. FIG. 3A shows a predicted accumulation amount of the reception buffer in a normal state where a picture is not skipped. The predicted accumulation amount of the reception buffer before decoding at the time of DTSn is VBV (n), the bit generation amount of the B picture decoded at the time of DTSn is Db, and the transmission bit amount for each DTS interval is Rb.
Then, VBV (n + 1) = VBV (n) -Db + R
Is calculated.

FIG. 3B shows the predicted accumulation amount of the reception buffer when a picture is skipped. VBV (n) is the predicted accumulation amount of the reception buffer before decoding at the time of DTSn.
And all skip B decoded at the time of DTSn
Assuming that the bit generation amount of the picture is Dbskip and the bit amount transmitted from the encoding device at each DTS interval is R, VBV (n + 1) = VBV (n) -Dbskip +
R is calculated.

The threshold setting means 127 sets thresholds (Ti, Tp, Tb1, Tb2) for each picture type (I, P, B1, B2 picture). FIG.
An example of setting a threshold will be described. As shown in the following equation, the code amount (predicted code amount) Ei that is empirically known for each picture type,
Ep and Eb are threshold values Ti, Tp, Tb1, and Tb, respectively.
Set to 2. Ti = Ei Tp = Ep Tb1 = Eb Tb2 = Eb In the present embodiment, the resolution of the image data is 720 pixels ×
Ei = 400 kbits, Ep = 200 kbits, E, which are empirically known as the optimum values for 480 pixels.
It is assumed that b = 160 kbit is used. Further, since the resolution of image data and the predicted code amount have a substantially proportional relationship, even when the resolution is different from the above, the predicted code amount may be calculated in proportion to the resolution.

The comparing and judging means 125 compares the predicted accumulation amount VBV of the reception buffer with the threshold value for each image data before the image data is encoded, and skips the image encoding process. It is determined whether or not to perform. That is, the comparison determination means 125 determines to skip if VBV <Ti when encoding the image as an I picture, and determines to skip if VBV <Tp when encoding the image as a P picture. When encoding the image as a B1 picture, VBV <Tb
If it is 1, it is determined to skip, and if the image is encoded as a B2 picture, it is determined to skip if VBV <Tb2.

FIG. 5 shows an example in which a picture is skipped. In FIG. 5A, DT is satisfied by VBV <Ti.
In Sn, the I picture is skipped. In FIG. 5B, because of VBV <Tb1, B1 in DTSn + 1
Skip a picture.

In FIG. 5C, since VBV <Tb2,
BTS picture is skipped in DTSn + 2. In FIG. 5D, the P picture is skipped in DTSn + 3 due to VBV <Tp. Comparison determination means 125
Is the DC when it is determined that the picture is to be skipped.
It causes the T means 113 to stop the encoding process and output an all-skip picture according to the picture type from the SKIP picture storage memory 126.

FIG. 6A shows an example of a screen normally displayed by the decoding device. In the first, second, third, and fourth frames, I (1), B (3), B (4), P
The decoded image of (2) is displayed. The numbers in parentheses indicate the encoding order. FIG. 6B shows an example of a screen displayed by the decoding device when a picture is skipped. When the B (3) picture of the second frame is skipped,
B (3) uses the decoded images of I (1) and P (2) as reference images, but in the second frame, the decoded image of I (1) whose display order is closer is displayed. (Operation) Next, an operation related to the skip processing of the video encoding device according to the present embodiment will be described.

FIG. 7 is a flowchart showing the operation procedure of the video encoding device. First, the threshold setting means 127
A threshold value Ti for skip determination for each picture type,
Tp, Tb1, and Tb2 are set (step S50).
1). Next, when the original image is encoded as an I picture, the comparison determination unit 125 compares and determines the size of the predicted accumulation amount VBV of the reception buffer with the threshold value Ti. VBV ≧
In the case of Ti, DCT means 113, quantization means 114
Then, the encoding process is performed by the variable length encoding unit 116 assuming that the original image is an I picture (steps S502 and S5).
03, S504).

Then, the reception buffer accumulation amount predicting means 12
4 calculates the predicted accumulation amount VBV of the reception buffer before decoding at the next DTS time point based on the generated bit amount of the I picture and the transmission bit amount R for each DTS time interval (step S505). Then, if there is the next image data, the processing is continued, and if there is no next image data, the processing is ended (step S506).

On the other hand, if the comparison and determination means 125 determines that VBV <Ti in step S503, the comparison and determination means 125 further compares the magnitudes of VBV and Tp. When VBV ≧ Tp, the encoding process is performed with the original image as a P picture. When VBV <Tp, the SKIP picture storage memory 126 outputs an all-skip P picture. Then, the reception buffer accumulation amount prediction means 124 calculates the bit generation amount of the all-skip P picture and the transmission bit amount R for each DTS time interval.
, The predicted accumulation amount VBV of the reception buffer before decoding at the next DTS is calculated (steps S508 and S508).
509, S510, S505).

When the original image is coded as a P picture, the comparing and determining means 125 compares and determines the magnitude of the predicted accumulation amount VBV of the receiving buffer and the threshold value Tp. V
When BV ≧ Tp, the encoding process is performed as usual with the original image as a P picture, and when VBV <Tp,
An all-skip P-picture is output from the SKIP picture storage memory 126. Then, the reception buffer storage amount prediction means 124 calculates the bit generation amount of the generated P picture or the bit generation amount of the all-skip P picture,
The next DT is determined by the transmission bit amount R for each TS time interval.
Predicted accumulated amount VBV of reception buffer before decoding at time S
(Steps S507, S508, S509,
S510, S505).

When the original image is coded as a B1 picture, the comparing and judging means 125 compares and judges the magnitude of the predicted accumulation amount VBV of the receiving buffer with the threshold value Tb1. If VBV ≧ Tb1, the original image is
An encoding process is performed as a picture, and when VBV <Tb1, an all-skip B picture is output from the SKIP picture storage memory 126. Then, the receiving buffer storage amount predicting unit 124 decodes at the time of the next DTS by using the generated bit amount of the B picture or the generated bit amount of the all-skip B picture and the transmission bit amount R for each DTS time interval. Calculate the predicted accumulation amount VBV of the previous reception buffer (steps S511, S51)
2, S513, S514, S505).

When the original image is encoded as a B2 picture, the comparing and determining means 125 compares and determines the size of the predicted accumulation amount VBV of the receiving buffer with the threshold value Tb2. When VBV ≧ Tb2, the original image is
An encoding process is performed as a picture, and when VBV <Tb2, an all-skip B picture is output from the SKIP picture storage memory 126. Then, the receiving buffer storage amount predicting unit 124 decodes at the time of the next DTS by using the generated bit amount of the B picture or the generated bit amount of the all-skip B picture and the transmission bit amount R for each DTS time interval. Calculate the predicted accumulation amount VBV of the previous reception buffer (steps S511, S51)
5, S513, S514, S505). (Summary) As described above, in the video encoding apparatus according to the present embodiment, a picture is once generated, and then the picture is skipped depending on whether or not the bit generation amount exceeds the predicted storage amount of the reception buffer. Instead of determining whether or not to skip a picture before generating a picture, it is determined whether or not the predicted code amount based on the empirical value for each picture type exceeds the predicted storage amount of the reception buffer. Since the determination is made, a special buffer temporarily stored for determining whether to skip the generated picture can be made unnecessary.

Although the first embodiment has been described above, the present invention is, of course, not limited to the above embodiment. That is, it goes without saying that the following modifications are also included in the present invention. (Modification 1) In the present embodiment, a fixed value is used as the threshold, but the threshold may be determined for each original image according to the complexity.

Let P _k (k = 1 to 64) be the pixel value of the original image in the block of 8 × 8 pixels. Average value E in block
If (P _k ), then E (P _k ) = 1/64 × ΣP _k . If the variance value is V (P _k ), V (P _k ) = 1
/ 64 × Σ (P _k −E (P _k )) ² .

[0042] The minimum value of V (P _k) of the 8 blocks included in the macroblock j (4 blocks in the four blocks and field DCT mode in the frame DCT mode) When _{VAR j, VAR j = MIN [} V
(P _k )]. If the activity of the macro block j is act _j , act _j = 1 + VAR _j .

If the activity ACT of the original image is the sum of the activities of all the macroblocks, AC
T = Σact _j . The activity reflects the variance of the pixel values of the original image. If the activity is large, the complexity in the frame of the original image is high, and it is considered that the bit generation amount of the picture is large. Therefore, it can be considered that the higher the activity, the higher the threshold value should be set.

Further, a threshold value may be set based on the amount of dispersion. For example, a VAR = ΣVAR _j,
When the VAR is large, the threshold may be set high, and when the VAR is small, the threshold may be set low. In addition, the variance of pixels of the entire image may be used instead of the variance of each block. That is, assuming that the number of pixels of the entire image is N and P _l (l = 1 to N) is the pixel value of the original image, the average value E of the pixels of the entire image is E = 1 / N × ΣP _l The overall variance value V is given by V = 1 / N × _l (P _l −E
(P _l )) ² . If V is large, the threshold may be set high, and if V is small, the threshold may be set low. (Modification 2) In the present embodiment, when the predicted code amount for each picture type exceeds the predicted storage amount of the reception buffer, the picture is skipped regardless of the picture type. However, the decoded image of an I picture or a P picture is used as a reference image of another picture, so that a skip in an I picture or a P picture affects other pictures as described later. Therefore, according to the picture type, the picture is skipped, the quantization scale is increased, only the DC component after the DCT is used, or the coding amount is suppressed such that the pseudo image data is coded. May be performed.

For example, when the predicted code amount exceeds the predicted storage amount of the receiving buffer, the picture is skipped only in the case of a B picture, and in the case of an I or P picture, the coding is performed with the code amount suppressed. It may be. Alternatively, skipping of a picture may be performed only in the case of a B or P picture, and encoding in which the code amount is suppressed may be performed in the case of an I picture. (Modification 3) In the present embodiment, underflow can be largely prevented, but in order to completely prevent underflow, when the predicted code amount is larger than the predicted accumulation amount of the reception buffer, as described above, the macroblock After encoding in units, the code amount of the macro block is compared with the accumulation amount of the reception buffer, and if an underflow occurs, the macro block may be set as a skipped macro block. <Second Embodiment> This embodiment relates to a video encoding apparatus in which skipping is more likely to occur in a B picture than in a P or I picture. (Configuration) The configuration of the video encoding apparatus according to the present embodiment is substantially the same as that of the video encoding apparatus according to the first embodiment, but the manner of setting the threshold by the threshold setting means 127 is different. Hereinafter, a method of setting a threshold by the threshold setting unit 127 will be described.

The threshold setting means 127 sets thresholds (Ti, Tp, Tb1, Tb2) for each picture type (I, P, B1, B2 picture). FIG.
An example of setting a threshold will be described. As in the following equation, Ti and Tp are set to the predicted code amounts Ei and Ep, as in the first embodiment. Ti = Ei Tp = Ep In the case of skipping with an I picture or a P picture, the B picture that refers to the skipped picture also has the same screen as the skipped picture. In this case,
For example, when M (appearance cycle of I or P picture) = 3, four frames are continuously the same screen. Therefore, the threshold value of the B picture is set to be larger than that in the first embodiment so that skipping occurs in the B picture immediately before the P picture as much as possible.

FIG. 9 is a diagram for explaining a threshold value of a B picture immediately before a P picture. VBV at the time of DTSn + 1
Assuming (n + 1), at the time of DTSn + 2, VBV
(N + 2) = VBV (n + 1) -Eb + R. VB
If V (n + 2) ≧ Tp, at the time of DTSn + 2, P
It can be predicted that picture skipping will not occur. When the equation is transformed, VBV (n + 2) = VBV (n + 1) −Eb +
R ≧ Tp = Ep, and VBV (n + 1) ≧ Eb + (E
p-R).

As in the first embodiment, VBV (n
+1) ≧ Eb, so that VB
It is necessary to satisfy the expression V (n + 1) ≧ MAX (Eb, Eb + (Ep−R)). Therefore, the threshold value Tb2 = MAX
(Eb, Eb + (Ep-R)), and when VBV is less than Tb2 at the time of DTSn + 1, the B2 picture can be skipped, and the P picture can be prevented from skipping in prediction.

When (Ep-R) ≧ 0, as shown in FIG. 9, a new threshold value Tb2 is obtained by adding (Ep-R) (an amount corresponding to FIG. 9) to the original Tb2. It will be. When (Ep−R) <0, the threshold value Tb
2 remains at the original Tb2. Similarly, assuming that VBV (n) at the time of DTSn, VBV (n + 2) = VBV (n) −Eb + R−Eb at the time of DTSn + 2.
+ R = VBV (n) -Eb- (Eb-2R). V
If BV (n + 2) ≧ Tp, at the time of DTSn + 2,
It can be predicted that skipping of the P picture will not occur. When the equation is transformed, VBV (n + 2) = VBV (n) −Eb−
(Eb-2R) ≧ Tp = Ep, and VBV (n) ≧ E
b + (Ep + Eb-2R) = Eb + (Ep-R) + (E
b−R) = Eb + (Tb2−R).

Also, as in the first embodiment, VBV
(N) Since it is necessary that ≧ Eb, VB
It is necessary to satisfy the formula of V (n) ≧ MAX (Eb, Eb + (Tb2-R)). Therefore, the threshold value Tb1 = MAX (E
b, Eb + (Tb2-R)), and at the time of DTSn,
When VBV is less than Tb1, it is possible to skip the B1 picture and prevent the P picture from skipping in prediction. Also, this is (Tb2
When −R) ≧ 0, as shown in FIG. 9, the new threshold value Tb1 is obtained by adding (Ep−R) (the amount corresponding to FIG. 9) and (Eb−R) (FIG. 9) to the original Tb1. ). When (Tb2-R) <0, the threshold value Tb1 remains at the original value Tb1.

The threshold value of the B picture immediately before the I picture is the same except that Ep is replaced by Ei. I
In order to distinguish between a picture and a P picture, the threshold of the B1 picture immediately before the P picture is Tb2 (p), the threshold of the B2 picture immediately before the P picture is Tb1 (p), and the threshold of the B2 picture immediately before the I picture is Tb2 ( i), assuming that the threshold value of the B2 picture immediately before the I picture is Tb1 (i), the following equation is established.

Tb2 (p) = MAX (Eb, Eb + (Ep−R)) Tb1 (p) = MAX (Eb, Eb + (Tb2 (p) −
R)) Tb2 (i) = MAX (Eb, Eb + (Ei−R)) Tb1 (i) = MAX (Eb, Eb + (Tb2 (i) −
R)) FIG. 10A shows the transition of the estimated accumulation amount of the reception buffer at the threshold value in the first embodiment. DTSn, DTS
At the time point of n + 1, the predicted accumulation amount VBV of the reception buffer is equal to or larger than the threshold values Tb1 and Tb2, and no skip of the B picture occurs. At the time point of DTSn + 2, the VBV becomes the threshold value Tb.
Since it is less than p, skipping of the P picture occurs.

FIG. 10B shows the transition of the estimated accumulation amount of the reception buffer at the threshold value in the present embodiment. As shown in the figure, since the threshold value Tb1 is set high, the DTS
At the time point n, a B1 picture skip occurs. As a result, the predicted accumulation amount VBV of the reception buffer increases by the difference (Db1-Dskip) between the bit generation amount of the B1 picture and the all-skip B picture amount. Therefore, skipping of the P picture does not occur.

FIG. 11A shows an example of a screen normally displayed by the decoding device. In the first, second, third, and fourth frames, I (1), B (3), B (4), P
The decoded image of (2) is displayed. The numbers in parentheses indicate the encoding order. FIG. 11B illustrates an example of a screen displayed by the decoding apparatus when skipping a picture is determined using the threshold according to the first embodiment. With the threshold value set in the first embodiment, skipping is likely to occur in a B picture preceding a P picture, and a skip is likely to occur in a P picture since no threshold value is set. In this case, since the P (2) picture is skipped, P (2) becomes I
Since the decoded image of (1) is used as the reference image, the decoded image of I (1) is displayed in the fourth frame. Since B (3) and B (4) refer to P (2), the decoded image of I (1) is displayed even in the second and third frames.

FIG. 11C shows an example of a screen displayed by the decoding apparatus when a picture is skipped at the threshold according to the present embodiment. In the threshold value set in the present embodiment, a skip is likely to occur in a B picture because the threshold value is easily set in a B picture before a P picture. In this case, since the B (3) picture of the second frame is skipped, B (3) becomes I
Although the decoded images of (1) and P (2) are used as reference images, in the second frame, the decoded image of I (1) whose display order is closer is displayed. Since there is no picture referring to the B (3) picture, decoded pictures of other pictures are displayed as usual.

FIG. 11D shows an example of a screen displayed by the decoding apparatus when a picture is skipped at the threshold according to the present embodiment. In this case, B of the third frame
(4) B (4) due to skipped picture
Uses the decoded images of I (1) and P (2) as reference images, but in the third frame, P
The decoded image of (2) is displayed. Since there is no picture referring to the B (4) picture, decoded pictures of other pictures are displayed as usual. (Operation) The operation of the video encoding apparatus according to the present embodiment is the same as the operation of the first embodiment in FIG. 7 except that the processing in step S501 is different. Therefore, description of the processing procedure is omitted. (Summary) As described above, according to the video encoding apparatus according to the present embodiment, the threshold value of a B picture is set high so as to satisfy a prediction condition that a skip does not occur in a P picture. Even in the case where a skip occurs in a P picture, skipping of a P picture can be avoided by skipping first in the immediately preceding B picture. (Modification 1) The video encoding apparatus according to the present embodiment has a unique effect of avoiding skipping in a P picture, which is not provided in the first embodiment. Is different from that disclosed in Japanese Patent Application Laid-Open Publication No. H08-28,279, in that a picture is generated and then temporarily stored in a buffer memory as described in Japanese Patent No. 2871316.

That is, a picture is generated by encoding image data, the picture is temporarily stored in a buffer memory, and based on the bit generation amount of the picture, a threshold value for avoiding skipping in a P picture as follows: May be set to determine whether to skip. The threshold value Ti in the case where image data is encoded into an I picture is set to the following equation, where the bit generation amount of the I picture is Di. Ti = Di Further, the threshold value Tp when the image data is encoded into a P picture is set to the following equation, where Dp is the bit generation amount of the P picture.

Tp = Dp The threshold value Tb (i) when the image data is encoded into the B picture immediately before the I picture is such that the bit generation amount of the B picture is Db, and the prediction code amount of the I picture is Eb.
Assuming that i is i and the amount of transmission bits per decoding time interval of one picture is R, the following equation is set to avoid skipping in I pictures.

Tb (i) = MAX (Db, Db + (Ei−R)) The threshold value Tb (p) when image data is encoded into a B picture immediately before a P picture is the bit generation amount of the B picture. Db, the prediction code amount of the P picture is Ep,
Assuming that the transmission bit amount per decoding time interval of one picture is R, the following equation is set to avoid skipping in a P picture.

Tb (p) = MAX (Db, Db + (Ep−R)) (Modification 2) FIG. 12 shows another setting example of the threshold value.
Ti, Tp, Tb2 (p), and Tb2 (i) are set to the same values as in the second embodiment as in the following equations.

Ti = Ei Tp = Ep Tb2 (p) = MAX (Eb, Eb + (Ep−R)) Tb2 (i) = MAX (Eb, Eb + (Ei−R)) In the second embodiment, Since the threshold value of the B1 picture is set high as described above, skipping is likely to occur. However, when the actual bit generation amounts of the B1, B2, and P pictures are smaller than the predicted amounts, the P picture may not be skipped even if the B1 picture is not skipped. Therefore, it is desirable to wait as long as possible for the B2 picture immediately before the P picture and confirm whether or not to skip, in order to avoid unnecessary skipping.

On the other hand, even if it is determined to skip after waiting for the B2 picture, if the predicted accumulation amount of the reception buffer is considerably small, skipping of the P picture can be avoided only by skipping the B2 picture. There are times when you can't. Therefore, the B1 picture is skipped even if it is skipped in the B2 picture and only when skipping occurs in the P picture.
A threshold value of the B1 picture for that purpose is set.

FIG. 13 is a diagram for explaining a threshold value of a B picture immediately before a P picture. At the time of DTSn, VBV
(N), at the time of DTSn + 2, VBV (n
+2) = VBV (n) -Eb + R-Dbskip + R =
VBV (n) -Eb- (Dbskip-2R).
Here, Dbskip is a bit generation amount of an all-skip B picture. If VBV (n + 2) ≧ Tp,
At the time of DTSn + 2, no P picture skip occurs. When the equation is transformed, VBV (n + 2) =
VBV (n) −Eb− (Dbskip−2R) ≧ Tp =
Ep, and VBV (n) ≧ Eb + (Ep + Dbski
p-2R) = Eb + (Ep-R) + (Dbskip-
R) = Dbskip + (Tb2-R).

Further, similarly to the first embodiment, VBV
(N) Since it is necessary that ≧ Eb, VB
V (n) ≧ MAX (Eb, Dbskip + (Tb2-
R)) must be satisfied. Therefore, the threshold value Tb1 =
MAX (Eb, Dbskip + (Tb2-R)),
At the time of DTSn, if VBV is less than Tb1, B
By skipping one picture, it is possible to prevent skipping of a P picture in prediction.

This is because Dbskip + (Tb2-R) ≧
In the case of Eb, as shown in FIG.
1 means that (Ep-R) (the amount corresponding to FIG. 13) was added to the original Tb1, and (R-Dbskip) (the amount corresponding to FIG. 13) was reduced. Also, Dbskip
When + (Tb2-R) <Eb, the threshold value Tb1 remains at the original Tb1.

The threshold value of the B picture immediately before the I picture is the same except that Ep is replaced by Ei. Therefore, the following equation is established. Tb1 (p) = MAX (Eb, Dbskip + (Tb2
(P) -R)) Tb1 (i) = MAX (Eb, Dbskip + (Tb2
(I) -R)) <Third Embodiment> In the present embodiment, when one frame is assigned to one picture (frame structure), and when encoding of image data is skipped, interlaced scanning is performed. The present invention relates to a video encoding device in which the display order is not reversed even when the video encoding is performed. (Configuration) The configuration of the video encoding apparatus according to the present embodiment is substantially common. Hereinafter, different points will be described.

The screen rearranging means rearranges the image data in frame units. The DCT unit 113 encodes image data in frame units. As in the first embodiment, when the predicted accumulation amount of the reception buffer is smaller than the threshold, the comparison determination unit 125 causes the DCT unit 113 to stop the encoding process of the image data of the next frame, and
An all skip picture is output from the KIP picture storage memory 126.

An all-skip picture P-picture and an all-skip B-picture for a field structure with a reference to the bottom field of the forward (display order) or the top field of the backward (display order) are stored in the SKIP picture storage memory. Is stored. Of these all-skip pictures, the one having the display order closest to the field of the skipped image data as the reference destination is selected by the comparison determination means 125 and transmitted.

FIG. 14 shows the picture header and picture coding extension of the picture layer of the all-skip B picture. As shown in the figure, the picture structure is specified in the top field. FIG. 15 shows fields to be referred to when frame 2 is skipped. FIG.
(A) shows a case where M (appearance period of I or P picture) = 1. Since frame 2 is a P picture, the fields that can be used as reference fields are 1t and 1b. Since the display order of both 2t and 2b is closer to 1b, 1b becomes the reference field of 2t and 2b.

Therefore, at the time of encoding 1t and 1b, an all-skip P-picture that refers to the forward bottom field is output. FIG.
(B) shows a case where M (appearance cycle of I or P picture) = 2. Since frame 2 is a B picture, fields that can be used as reference fields are 1t, 1b, 3
t and 3b. 2t is closer to 1b in display order,
1b becomes a reference field of 2t. Since 2b has a display order close to that of 3t, 3t becomes a reference field of 2b.

Therefore, at the time of encoding 2t, an all-skip B-picture that refers to the forward bottom field is output. Therefore, at the time of encoding 2b, an all-skip B-picture that refers to the top field of the backward is output. FIG. 15C shows a case where M (appearance period of I or P picture) = 3. Since frame 2 is a B picture, the fields that can be used as reference fields are 1t, 1b, 4t, and 4b. Since the display order of 2t and 2b is close to that of 1b, 1b is a reference field.

Therefore, at the time of encoding 2t and 2b, all-skip B pictures each having the forward bottom field as a reference are output. FIG. 16 shows a field to be referred when the frame 3 is skipped. FIG. 16A shows M (I or P
The case where (appearance period of picture) = 1 is shown. Frame 3
Is a P-picture, the fields that can be used as reference fields are 2t and 2b. Among them, 3t,
Since the display order of both 3b is closer to 2b, 2b is 3
Reference fields t and 3b.

Therefore, at the time of encoding 3t and 3b, an all-skip P-picture which refers to the bottom field of the forward is output. FIG.
(B) shows a case where M (appearance cycle of I or P picture) = 2. Since frame 3 is a P picture, the fields that can be used as reference fields are 1t and 1b. Since the display order of 3t and 3b is closer to 1b, 1b is a reference field for 3t and 3b.

Accordingly, at the time of encoding 3t and 3b, an all-skip P-picture which refers to the forward bottom field is output. Also, 3
By displaying 1b at time t, 2t and 2b are also
In order to prevent the display order from being reversed, skip processing is performed for 2t and 2b with the reference field set to 1b. Also in this case, an all-skip picture with the forward bottom field as a reference destination is output.

FIG. 16C shows a case where M (appearance period of I or P picture) = 3. Since frame 3 is a B picture, the fields that can be used as reference fields are 1b, 1t, 4t, and 4b. Among them, 3t,
In 3b, the display order is close to 4t, so 4t is 3t, 3t
b becomes the reference field. Therefore, an all-skip B-picture that refers to the top field of the backward is output.

FIG. 17 shows a reference field when frame 4 is skipped when M (appearance period of I or P picture) = 3. Since frame 4 is a P picture,
Fields that can be used as reference fields are 1t,
1b. Since the display order of 4t and 4b is close to that of 1b, 1b becomes the reference field of 4t and 4b.

Therefore, at the time of encoding 4t and 4b, an all-skip P-picture whose forward bottom field is a reference is output. Also, 4
By displaying 1b at time t, 2t, 2b, 3
t, 3b are also 2t, 2b to prevent the display order from being reversed.
For b, 3t, and 3b, skip processing is performed with the reference field set to 1b. Also in this case, an all-skip picture with the forward bottom field as a reference destination is output. (Operation) Next. The operation of the video encoding device according to the present embodiment will be described.

FIGS. 18 and 19 are flowcharts showing the operation procedure when M (appearance period of I or P picture) = 3 in the video encoding apparatus according to the present embodiment. Setting of threshold (S501) and determination of whether to skip (S502, S503, S507, S508, S5
11, S512, S515), encoding processing as an I picture (S504), encoding processing as a P picture (S509), encoding processing as a B picture (S51)
The operation of 3) is the same as that of the first embodiment, and the description is omitted.

This embodiment is different from the first embodiment in the method of skipping pictures shown in S801 to S810, so these steps will be described. In the present embodiment,
When skipping, since the frame structure is switched to the field structure, two all-skip pictures of a top field and a bottom field for one screen are output.
That is, the picture is skipped twice.

In step S512, VBV <Tb
In the case of 1, steps S807 and S808
, An all-skip B picture with the forward bottom field as a reference is output twice. (Refer to FIG. 15C for an example of the reference destination.) If VBV <Tb2 in step S515, step S809 is executed.
Then, in step S810, an all-skip B picture having the backward top field as a reference destination is output twice. (Refer to FIG. 16C for an example of the reference destination.)

In step S508, VBV <Tp
In the case of, in steps S801 and S802,
An all-skip P picture with the forward bottom field as a reference is output twice (see FIG. 17 for an example of the reference). Then, in steps S803 to S806, instead of the B picture between the skipped P picture and the referenced I picture, an all-skip B picture with the forward bottom field as a reference is output, respectively (see References). See FIG. 17 for the previous example).

In step S 505, the reception buffer accumulation amount prediction means 124 calculates the bit generation amount of the generated picture or the bit generation amount of two or six all-skip pictures and the transmission bit amount for each DTS time interval. Based on R, a predicted accumulation amount VBV of the reception buffer before decoding at the next DTS is calculated. (Summary) As described above, according to the video encoding device according to the present embodiment, when one frame is assigned to one picture (frame structure), and the predicted accumulation amount of the reception buffer becomes smaller than the threshold value When the encoding of the image data is skipped and the encoding of the image data is skipped, and the encoding of the image data is skipped, the encoding of the image data is skipped, and the encoding of the image data is skipped. Even when the image is displayed by the scanning method, the display order can be prevented from being reversed. (Modification) In the video encoding device according to the present embodiment,
Although it was determined whether or not to skip a picture based on the estimated storage amount of the reception buffer, the present invention is not limited to this. For example, whether or not to skip a picture may be determined based on the amount of storage in the output buffer of the video encoding device.

[0083]

In order to achieve the above object, the present invention provides:
A video encoding device that encodes one frame or one field of image data while predicting a storage amount of a reception buffer in a decoding device, wherein encoded data of the image data is encoded before encoding the image data. Comparing means for comparing a predicted accumulation amount predicted to be accumulated in the reception buffer at a decoding time of the reception buffer with a predetermined threshold value, and a code of the image data when the predicted accumulation amount is less than the threshold value. And that the same image data as the previously decoded image data should be displayed instead of the coded data of the image data as the data taken out from the reception buffer at the decoding time. And a skip unit using a proxy code.

As a result, it is determined whether or not to stop the coding before coding the image data, so that the coded data of the image data is determined in order to determine whether or not the predicted storage amount of the reception buffer is less than the threshold value. This eliminates the need for a special buffer for temporarily storing data. Here, the video encoding apparatus further includes a threshold setting unit that sets a threshold for each type of picture (I picture, P picture, or B picture) in which image data is encoded. I do.

Thus, I picture, P picture, B picture
Since the encoding method differs for each picture, the order of the code amount also differs, so that a threshold value can be appropriately set according to the picture type. Here, when the picture type is a B picture, the skip means is an all-skip B picture in which all the remaining macroblocks except the first and last of the slice layer are B pictures composed of skipped macroblocks. Is used as a proxy code, and when the picture type is I or P picture,
The present invention is characterized in that an all-skip P-picture, which is a P-picture in which all the remaining macroblocks except the first and last slice layers are composed of skipped macroblocks, is used as a proxy code.

Thus, it is possible to create a proxy code that specifies that the same image data as the previously decoded image data should be displayed using the MPEG skipped macroblock. Here, the threshold value of the threshold value setting means is a predicted code amount. Thereby, since the prediction code amount of the picture for each of the I, P, and B picture types is set to the threshold value, the underflow of the reception buffer can be prevented with high accuracy.

Here, the threshold value setting means calculates a variance value of a pixel value for each image data, and sets a larger threshold value as the variance value increases.
Thus, the larger the variance of the pixels of the image data,
In general, the code amount of the image data becomes large, and by setting a large threshold value for the image data having a large variance value, it is possible to more accurately prevent the underflow of the reception buffer.

Here, the threshold value setting means calculates the variance value of the pixel value of the image data in each of eight blocks (four blocks in the frame DCT mode and four blocks in the field DCT mode) included in the macroblock j. VAR _j , the activity of macro block j is act _j, and act _j = 1 + V
And AR _j, the activity ACT of the image data when the sum of the activities of all the macro blocks, the more activity ACT of image data is large image, and sets a larger threshold.

Thus, the larger the activity of the image data, the larger the code amount of the image data in general. Therefore, by setting a larger threshold value for the image data having a higher activity, the underflow of the reception buffer can be increased. Accuracy can be prevented. here,
The threshold value setting means sets a predicted code amount for each of I or P pictures as a threshold value, and sets a threshold value for a B picture larger than the predicted code amount of the B picture.

As a result, the threshold value of the B picture is set higher than the predicted code amount, and the skip in the B picture is easily performed. Therefore, the skip in the I or P picture referred to by another picture is less likely to occur. As a result, it is possible to prevent the same image from being displayed continuously many times. Here, the threshold value setting means is configured to set M (I
Or the appearance period of the P picture) ≧ 2, the predicted code amount of the I picture is Ei, the predicted code amount of the P picture is Ep,
Assuming that the predicted code amount of the B picture is Eb and the transmission bit amount per decoding time interval of one picture is R, the threshold value of the I picture is Ti = Ei, the threshold value of the P picture is Tp = Ep, and the coding order is that of the I picture. Set the threshold value of the immediately preceding B picture to Tb
Assuming that (i), Tb (i) = Eb + (Ei−R) (Ei−R) ≧ 0, Tb (i) = Eb (Ei−R) <0, and the encoding order of the P picture Let the threshold of the immediately preceding B picture be T
When b (p), Tb (p) = Eb + (Ep−R) (Ep−R) ≧ 0, Tb (p) = Eb (Ep−R) <0, I do.

Thus, the threshold value of the B picture immediately before the I or P picture can be set based on the predicted code amount of the I or P picture so that the skip does not occur in the I or P picture. Here, the threshold setting means includes:
If M (appearance period of I or P picture) ≧ 3, the predicted code amount of the I picture is Ei, the predicted code amount of the P picture is Ep, the predicted code amount of the B picture is Eb, and the decoding time interval of one picture is Let the transmission bit amount of R be R, the threshold value of the I picture Ti = Ei, the threshold value of the P picture Tp = E
p, the B picture (B2
Assuming that the threshold value of (i)) is Tb2 (i), Tb2 (i)
= Eb + (Ei-R) When (Ei-R) ≥0, When Tb2 (i) = Eb (Ei-R) <0, the encoding order of the B picture (B1) immediately before B2 (i)
If the threshold of (i)) is Tb1 (i), Tb1 (i) = Eb + (Tb2 (i) -R) (Tb2 (i) -R) ≧ 0, Tb1 (i) = Eb (Tb2 ( i) -R) <0, the encoding order is B picture immediately before P picture (B2
If the threshold of (p)) is Tb2 (p), Tb2 (p) = Eb + (Ep−R) (Ep−R) ≧ 0, and Tb2 (p) = Eb (Ep−R) <0 , The encoding order of the B picture (B1) immediately before B2 (p)
Assuming that the threshold value of (p)) is Tb1 (p), when Tb1 (p) = Eb + (Tb2 (p) -R) (Tb2 (p) -R) ≧ 0, Tb1 (p) = Eb (Tb2 ( p) -R) <0, where

Thus, the threshold value of the B picture immediately before the I or P picture and the threshold value of the B picture immediately before the B picture are set based on the predicted code amount of the I or P picture so that skipping does not occur in the I or P picture. can do. Here, when M (appearance period of I or P picture) ≧ 3, the threshold setting means sets the threshold of the B picture (B2) immediately before the I picture in the coding order, It is characterized in that it is set higher than the threshold value of the B picture (B1) immediately before (B2).

Thus, the skipping of the B picture to avoid the skipping of the I or P picture is performed by the I or P picture.
A point that can easily occur immediately before a picture and has a high degree of uncertainty as to whether skipping is necessary for an I or P picture, that is, the amount of storage in a reception buffer at the decoding time of a P picture based on a large number of prediction items Can be avoided at the point of time when prediction is made.

Here, when M (appearance period of I or P picture) ≧ 3, the threshold setting means sets the predicted code amount of the I picture to Ei, the predicted code amount of the P picture to Ep, and the predicted code amount of the B picture. When the code amount is Eb, the transmission bit amount R per one picture decoding time interval, and the code amount of the all-skip B picture is Dbskip, the threshold value T of the I picture
i = Ei, threshold value Tp = Pp of P picture, threshold value T of B picture (B2 (i)) immediately before the encoding order of I picture
Assuming that b2 (i), Tb2 (i) = Eb + (Ei−R) (Ei−R) ≧ 0, Tb2 (i) = Eb (Ei−R) <0, and the encoding order is B2 ( i) B picture immediately before (B1
Assuming that the threshold value of (i)) is Tb1 (i), Tb1 (i) = Dbskip + (Tb2 (i) −R) Dbskip + (Tb2 (i) −R) ≧ Eb, Tb1 (i) = Eb Dbskip + ( When Tb2 (i) -R) <Eb, the encoding order is the B picture (B2
If the threshold of (p)) is Tb2 (p), Tb2 (p) = Eb + (Ep-R) When (Ep-R) ≧ 0, Tb2 (p) = Eb (Ep-R) <0 , The encoding order of the B picture (B1) immediately before B2 (p)
If the threshold of (p)) is Tb1 (p), Tb1 (p) = Dbskip + (Tb2 (p) -R) Dbskip + (Tb2 (p) -R) ≧ Eb Tb1 (p) = Eb Dbskip + (Tb2) (P) -R) <Eb.

As a result, in order to avoid skipping in the I or P picture, only when skipping in the B2 picture, the B1 picture is also skipped.
At a time when there is a high degree of uncertainty as to whether or not a picture requires skipping, that is, at a time when the amount of storage in the reception buffer at the decoding time of a P picture is predicted based on a large number of prediction items, the use of unnecessary B1 pictures Skips can be avoided.

Further, the present invention predicts one frame or one frame while predicting the accumulation amount of the reception buffer in the decoding device.
A video encoding device for encoding image data of a field, wherein after encoding the image data, M (appearance period of I or P picture) ≧ 2 according to the picture type
In the case of, when the image data is encoded into an I picture, the threshold value Ti is set to Ti = Di where the code amount of the I picture is Di, and when the image data is encoded into a P picture. Assuming that the code amount of the P picture is Dp, the threshold value Ti is set to Tp = Dp.
The threshold value Tb (i) when image data is encoded in a B picture immediately before a picture is represented by D
b, the estimated code amount of the I picture is Ei, and the amount of transmission bits per decoding time interval of one picture is R. When Tb (i) = Db + (Ei−R) (Ei−R) ≧ 0, When Tb (i) = Db (Ei-R) <0, the threshold value Tb (p) is set to when the image data is encoded in the B picture immediately before the P picture. Let Db be the code amount of P, E be the predicted code amount of a P picture, and R be the transmission bit amount per decoding time interval of one picture: Tb (p) = Db + (Ep−R) (Ep−R) ≧ When 0, Tb (p) = Db (Ep−R) <0, a threshold setting means for setting the following equation, and after the encoding of the image data, the reception buffer at the decoding time of the encoded data of the image data The amount that is expected to be accumulated Comparing means for comparing with the set threshold value, and, when the predicted accumulation amount is less than the threshold value, decoding in the past instead of the encoded data of the image data extracted from the reception buffer at the decoding time. And a skip unit that uses a proxy code that specifies that the same image data as the converted image data should be displayed.

This makes it difficult for skips to occur in I or P pictures, and since the threshold is always set to the actual code amount rather than the predicted code amount of the picture, it is possible to reliably prevent the underflow of the reception buffer. Can be. The present invention also relates to a video encoding apparatus for encoding one frame of image data in a frame structure, wherein when encoding the image data, the accumulated amount of an output buffer or the predicted accumulation of a reception buffer of a decoding apparatus. A comparing unit that compares the amount with a predetermined reference value, and when it is determined that the accumulation amount or the estimated accumulation amount has reached a predetermined reference value, stops encoding the image data in a frame structure, A proxy code that specifies that two fields that are the same as the top field or the bottom field constituting the previously decoded image data should be displayed instead of the encoded data of the top field and the bottom field of the image data And a skip means using

As a result, when the encoding of the image data is skipped, the method of allocating the picture is switched to the field structure, and each field of the skipped image data can be referred to in both the top field and the bottom field. Even when the frame structure is maintained, it is possible to prevent the display order from being reversed.

Here, the skipping means sets the display order to each field of the image data, of which the encoding in the frame structure has been stopped, out of the top field and the bottom field constituting the image data decoded in the past. It is characterized by using a proxy code that specifies that two fields that are the same as the nearest field should be displayed.

As a result, in each field of the image data for which the encoding is skipped, the field having the closest display order is set as the reference destination field. Therefore, the reference destination field of the top field of the image data for which the encoding is skipped is skipped. Since the display order cannot be later than the reference destination field of the bottom field of the data, the inversion of the display order can be reliably avoided.

Here, when the picture type in which the picture data whose coding in the frame structure has been stopped was to be coded is a B picture, the skip means sets the beginning and end of the slice layer to the B picture. All the skipped macroblocks except for all the remaining macroblocks are two skipped macroblocks, two of which are all-skip B-pictures, as the proxy code, and the image data, which has been coded in a frame structure, is coded. If the type of picture that was to be deleted is an I or P picture, all the remaining macroblocks except for the beginning and end of the slice layer are P pictures composed of skipped macroblocks. Are used as the proxy code.

Thus, it is possible to create a proxy code specifying that the same image data as the previously decoded image data should be displayed, using the MPEG skipped macroblock. Here, the skip unit is configured to use the all-skip P used when the type of picture in which the image data whose coding in the frame structure has been stopped is to be coded is an I or P picture.
In the picture, the bottom field of the image data of the I or P picture whose display order comes first is used as a reference field,
The all-skip B picture used when the picture type in which the image data whose coding in the frame structure has been stopped is to be coded is a B picture (B1) coded immediately after an I or P picture. In the above, the bottom field of the image data of the I or P picture whose display order is earlier is used as a reference field, and the type of the picture in which the image data whose coding in the frame structure has been stopped is to be coded is I or P. In the all-skip B picture used in the case of the B picture (B2) encoded immediately before the P picture, the top field of the image data of the I or P picture whose display order is later is used as a reference field. I do.

As a result, an all-skip picture in which an appropriate reference field is specified is selected according to the picture type, so that inversion of the display order can be easily and appropriately avoided. Here, when the picture type in which the image data whose coding in the frame structure has been stopped is to be coded is an I or P picture, the skip means is coded into a B picture. It is assumed that the encoding of the image data to be stopped is stopped, and two all-skip B pictures whose display order is the first bottom field as a reference field are used instead of the encoded data of the top field and the bottom field of the image data. Features.

Thus, when a skip occurs in a P or I picture, the field of the picture data of the B picture which refers to the same picture as the field referred to by the field of the skipped picture data is also used. By using a reference destination field, the display order can be prevented from being reversed. Also, the present invention predicts the amount of storage in the reception buffer in the decoding device,
A video encoding device for encoding one frame or one field of image data, which is stored in a reception buffer at the decoding time of the encoded data of the image data before encoding the image data. Comparing means for comparing a predicted accumulation amount predicted to be waxed with a threshold value for each picture type in which the image data is encoded; and a case in which the estimated accumulation amount is less than the threshold value, wherein the picture data is encoded. If the type of the image data is a B picture, the encoding of the image data is stopped, and in place of the encoded data of the image data extracted from the reception buffer at the decoding time, the image data decoded in the past is used. And control means for using a proxy code defining that the same image data should be displayed.

Thus, even if the predicted accumulation amount of the reception buffer is less than the threshold value, the picture is skipped only in the case of the B picture, so that the I or P picture referred to by another picture can be encoded with the code amount suppressed. By doing so, the number of times the same image is displayed continuously can be reduced. Further, the present invention estimates one frame or one
A video encoding device for encoding image data in a field, which is predicted to be stored in a reception buffer at the decoding time of the encoded data of the image data before encoding the image data. Comparing means for comparing the predicted accumulation amount with a threshold value for each picture type in which the image data is encoded; and a case in which the predicted accumulation amount is less than the threshold value and the type of the picture in which the image data is encoded is B Or, in the case of a P picture, the encoding of the image data is stopped, and instead of the encoded data of the image data extracted from the reception buffer at the decoding time, the same as the previously decoded image data is used. And control means for using a proxy code that specifies that the image data should be displayed.

As a result, even if the predicted storage amount of the reception buffer is less than the threshold value, the picture is skipped only in the case of a B or P picture, so that the I-picture referred to by other pictures can be encoded with the code amount suppressed. By doing so, the frequency with which the same image is continuously displayed can be reduced. Further, the present invention is a video encoding method for encoding one frame or one field of image data while predicting the accumulation amount of a reception buffer in a decoding device, wherein the video data is encoded before encoding the image data. A step of comparing a predicted storage amount predicted to be stored in the reception buffer at the decoding time of the encoded data of the image data with a predetermined threshold, and when the predicted storage amount is less than the threshold, Stop the encoding of the image data, and display the same image data as the previously decoded image data, instead of the encoded data of the image data, as the data extracted from the reception buffer at the decoding time. And a step of using a proxy code that defines what should be done.

As a result, it is determined whether or not the coding is stopped before the image data is coded. Therefore, the coded data of the image data is determined in order to determine whether or not the predicted storage amount of the reception buffer is less than the threshold value. This eliminates the need for a special buffer for temporarily storing data. The present invention also relates to a video encoding method for encoding one frame of image data in a frame structure, wherein when encoding the image data, the storage amount of an output buffer or the prediction accumulation of a reception buffer of a decoding device. Comparing the amount with a predetermined reference value, and when it is determined that the storage amount or the predicted storage amount has reached a predetermined reference value, stop encoding the image data in a frame structure, Instead of the encoded data of the top field and the bottom field of the image data, a proxy code that specifies that two fields that are the same as the top field or the bottom field constituting the image data decoded in the past should be displayed. And a step of using.

As a result, when the encoding of the image data is skipped, the method of allocating the picture is switched to the field structure, and each field of the skipped image data can be referred to by both the top field and the bottom field. Even when the frame structure is maintained, it is possible to prevent the display order from being reversed.

Further, the video encoding program of the present invention uses a computer to encode one frame or one field of image data while predicting the amount of storage in the receiving buffer in the decoding device. Before performing the operation, the comparing unit compares the predicted storage amount predicted to be stored in the reception buffer at the decoding time of the encoded data of the image data with a predetermined threshold value, If less than, the encoding of the image data is stopped, and as the data extracted from the reception buffer at the decoding time, instead of the encoded data of the image data, the image data decoded in the past and It functions as a skip unit that uses a proxy code that specifies that the same image data should be displayed.

As a result, it is determined whether or not to stop encoding before encoding the image data. Therefore, the encoded data of the image data is determined in order to determine whether or not the predicted storage amount of the reception buffer is less than the threshold. This eliminates the need for a special buffer for temporarily storing data. The video encoding program according to the present invention may be configured such that the computer encodes one frame of image data in a frame structure, and when encoding the image data, stores the amount of output buffer or the reception buffer of the decoding device. Comparing means for comparing the predicted storage amount with a predetermined reference value; and stopping the encoding of the image data in a frame structure when the storage amount or the predicted storage amount is determined to have reached a predetermined reference value. Then, in place of the encoded data of the top field and the bottom field of the image data, two fields that are the same as the top field or the bottom field constituting the image data decoded in the past are to be displayed. Function as a skip unit using a proxy code.

As a result, when the encoding of the image data is skipped, the method of allocating the picture is switched to the field structure, and each field of the skipped image data can be referred to in both the top field and the bottom field. Even when the frame structure is maintained, it is possible to prevent the display order from being reversed.

[Brief description of the drawings]

FIG. 1 shows a configuration of a video encoding device 100.

FIG. 2 (a) shows the original order of the video images. FIG. 2B shows the encoding order of video images.

FIG. 3A shows a predicted accumulation amount of a reception buffer in a normal state where a picture is not skipped. FIG. 3 (b)
This shows the estimated accumulation amount of the reception buffer when a picture is skipped.

FIG. 4 shows a setting example of a threshold.

FIG. 5A shows a predicted accumulation amount of a reception buffer when an I picture is skipped. FIG. 5B shows the predicted accumulation amount of the reception buffer when the B1 picture is skipped. FIG. 5C shows the predicted accumulation amount of the reception buffer when the B2 picture is skipped. FIG. 5D shows the predicted accumulation amount of the reception buffer when the P picture is skipped.

FIG. 6A shows an example of a screen normally displayed by the decoding device. FIG. 6B shows an example of a screen displayed by the decoding device when a picture is skipped.

FIG. 7 is a flowchart showing an operation procedure of the video encoding device according to the first embodiment.

FIG. 8 shows a setting example of a threshold.

FIG. 9 is a diagram illustrating a threshold value of a B picture immediately before a P picture.

FIG. 10A shows a transition of a predicted accumulation amount of a reception buffer at a threshold according to the first embodiment. FIG.
(B) shows the transition of the estimated accumulation amount of the reception buffer at the threshold value in the second embodiment.

FIG. 11A shows an example of a screen normally displayed by the decoding device. FIG. 11B illustrates an example of a screen displayed by the decoding apparatus when skipping a picture is determined using the threshold according to the first embodiment. FIG. 11C shows the second
17 shows an example of a screen displayed by the decoding device when a picture is skipped at the threshold according to the embodiment. FIG.
(D) illustrates an example of a screen displayed by the decoding device when a picture skip occurs at the threshold according to the second embodiment.

FIG. 12 shows a setting example of a threshold.

FIG. 13 is a diagram illustrating a threshold value of a B picture immediately before a P picture.

FIG. 14 shows a picture header and picture coding extension of a picture layer of an all-skip B picture.

FIG. 15A shows a case where M (appearance period of I or P picture) = 1. FIG. 15B shows a case where M (appearance cycle of I or P picture) = 2. FIG.
(C) shows a case where M (appearance cycle of I or P picture) = 3.

16A shows a case where M (appearance period of I or P picture) = 1; FIG. 16B shows a case where M (appearance period of I or P picture) = 2; . FIG.
(C) shows a case where M (appearance cycle of I or P picture) = 3.

FIG. 17 shows a reference field when frame 4 is skipped when M (appearance period of I or P picture) = 3.

FIG. 18 is a flowchart showing an operation procedure of the video encoding device according to the third embodiment.

FIG. 19 is a flowchart showing an operation procedure of the video encoding device according to the third embodiment.

FIG. 20A shows a normal transition of a predicted accumulation amount of a reception buffer. FIG. 20B shows an example in which the predicted accumulation amount of the reception buffer underflows. FIG.
(C) shows an example in which the predicted accumulation amount of the reception buffer overflows.

FIG. 21A shows an example of a screen normally displayed. FIG. 21B shows an example of a screen displayed when a skipped macroblock is used. FIG.
(C) shows an example of a screen displayed when a pseudo image is used.

FIG. 22 shows a configuration of a video encoding device described in Japanese Patent No. 2871316.

FIG. 23A shows a normal display screen in which a picture is not skipped. FIG. 23B shows a display screen when a picture skip occurs.

[Explanation of symbols]

 6 Orthogonal Transformation Circuit 20 Buffer Memory 22 SKIP Code Storage Memory 24 Transmission Rate Excess Judgment Circuit 100 Video Encoding Device 110 Screen Rearrangement Unit 111 Adder 112 Adder 113 DCT Unit 114 Quantization Unit 115 Rate Control Unit 116 Variable Length Coding Means 117 Buffer 118 Switching device 119 Inverse quantization means 120 Inverse DCT means 121 Buffer memory 122 Video memory 123 Motion compensation prediction means 124 Reception buffer storage amount prediction means 125 Comparison judgment means 126 SKIP picture storage memory 127 Threshold setting means

 ────────────────────────────────────────────────── ─── Continuing from the front page (72) Inventor Masahiro Morishita 2-6-1, Sakae, Naka-ku, Nagoya City, Aichi Prefecture Shirakawa Building Annex 5F Matsushita Electric Information System Nagoya Laboratory F-term (reference) 5C059 KK35 LA05 LA09 MA23 MC11 ME01 NN21 PP05 PP06 PP07 TA03 TA25 TB07 TC10 TC16 TC19 TC38 TD04 TD12 UA02 UA05 UA33 5J064 BA09 BA16 BB03 BC01 BC08 BC14 BC16 BC22 BD01

Claims (23)

[Claims] 1. A video encoding apparatus for encoding one frame or one field of image data while predicting a storage amount of a reception buffer in a decoding apparatus, wherein the image data is encoded before encoding the image data. A comparing unit that compares a predicted storage amount predicted to be stored in the reception buffer at the decoding time of the encoded data of the data with a predetermined threshold; and when the predicted storage amount is less than the threshold, Stop the encoding of the image data, and display the same image data as the previously decoded image data, instead of the encoded data of the image data, as the data extracted from the reception buffer at the decoding time. And a skip unit that uses a proxy code that defines what should be done. 2. The video encoding apparatus according to claim 1, further comprising: threshold setting means for setting a threshold for each type of picture (I picture, P picture or B picture) in which the image data is encoded. The video encoding device according to claim 1, wherein: 3. When the picture type is a B picture, the skip means is an all-skip which is a B picture in which all the remaining macroblocks except the first and last of the slice layer are B pictures composed of skipped macroblocks. B
When a picture is used as a proxy code and the picture type is an I or P picture, all the remaining macroblocks except the first and last of the slice layer are all skip P pictures that are P pictures composed of skipped macroblocks. 3. The video encoding apparatus according to claim 2, wherein the picture is used as a proxy code. 4. The video encoding apparatus according to claim 3, wherein the threshold value of said threshold value setting means is a predicted code amount. 5. The video according to claim 3, wherein said threshold value setting means calculates a variance value of a pixel value for each image data, and sets a larger threshold value as the variance value increases. Encoding device. 6. The macroblock j
8 blocks included in the frame (in the frame DCT mode
4 blocks and 4 blocks in field DCT mode
Of pixel values of image data in each block
VAR is the minimum of the values_jAnd the macro block j
Act activity_jAnd act _j= 1 + VAR_j
And the activity ACT of the image data
When the activity ACT of the image data is large, assuming the sum of the activities of the black blocks,
4. The method according to claim 3, wherein a larger threshold is set.
Video encoding device. 7. The threshold setting unit sets a predicted code amount for each of I or P pictures as a threshold,
4. The video encoding apparatus according to claim 3, wherein a value larger than a predicted code amount of the B picture is set as the threshold for the picture. 8. The method according to claim 1, wherein when M (appearance period of I or P picture) ≧ 2, the predicted code amount of the I picture is Ei, the predicted code amount of the P picture is Ep, and the predicted code amount of the B picture is When the amount is Eb, and the amount of transmission bits per decoding time interval of one picture is R, the threshold value of the I picture is Ti = Ei, the threshold value of the P picture is Tp = Ep, and the encoding order of the B picture immediately before the I picture is Threshold T
If b (i), Tb (i) = Eb + (Ei-R) (Ei-R) ≧ 0, Tb (i) = Eb (Ei-R) <0, the encoding order is P picture The threshold value of the B picture immediately before
When b (p), Tb (p) = Eb + (Ep−R) (Ep−R) ≧ 0, Tb (p) = Eb (Ep−R) <0, The video encoding device according to claim 7, wherein 9. The method according to claim 1, wherein when M (appearance period of I or P picture) ≧ 3, the predicted code amount of the I picture is Ei, the predicted code amount of the P picture is Ep, and the predicted code amount of the B picture is When the amount is Eb, and the amount of transmission bits per decoding time interval of one picture is R, the threshold value of the I picture Ti = Ei, the threshold value of the P picture Tp = Ep, and the encoding order of the B picture ( B2
If the threshold of (i)) is Tb2 (i), Tb2 (i) = Eb + (Ei-R) (Ei-R) ≧ 0, and Tb2 (i) = Eb (Ei-R) <0 , The encoding order of the B picture (B1
If the threshold of (i)) is Tb1 (i), Tb1 (i) = Eb + (Tb2 (i) -R) (Tb2 (i) -R) ≧ 0, Tb1 (i) = Eb (Tb2 ( i) -R) <0, the encoding order is B picture immediately before P picture (B2
If the threshold of (p)) is Tb2 (p), Tb2 (p) = Eb + (Ep−R) (Ep−R) ≧ 0, and Tb2 (p) = Eb (Ep−R) <0 , The encoding order of the B picture (B1) immediately before B2 (p)
Assuming that the threshold value of (p)) is Tb1 (p), when Tb1 (p) = Eb + (Tb2 (p) -R) (Tb2 (p) -R) ≧ 0, Tb1 (p) = Eb (Tb2 ( 8. The video encoding apparatus according to claim 7, wherein: p) -R) <0. 10. The threshold setting means sets a threshold of a B picture (B2) immediately before an I picture when M (appearance period of I or P picture) ≧ 3, B picture (B1) immediately before B picture (B2)
The video encoding apparatus according to claim 7, wherein the video encoding apparatus is set to be higher than the threshold value. 11. The threshold setting means, when M (appearance period of an I or P picture) ≧ 3, sets a predicted code amount of an I picture to Ei, sets a predicted code amount of a P picture to Ep, and sets a predicted code amount of a B picture to When the amount is Eb, the transmission bit amount R per one picture decoding time interval, and the code amount of the all-skip B picture is Dbskip, the threshold value of I picture Ti = Ei, the threshold value of P picture Tp = Ep, the coding order Is the B picture (B2) immediately before the I picture.
Assuming that the threshold value of (i)) is Tb2 (i), Tb2 (i) = Eb + (Ei-R) when (Ei-R) ≧ 0, and Tb2 (i) = Eb (Ei-R) <0 , The encoding order of the B picture (B1
Assuming that the threshold value of (i)) is Tb1 (i), Tb1 (i) = Dbskip + (Tb2 (i) −R) Dbskip + (Tb2 (i) −R) ≧ Eb, Tb1 (i) = Eb Dbskip + ( Tb2 (i) -R) <Eb When the encoding order is a B picture (B2
If the threshold of (p)) is Tb2 (p), Tb2 (p) = Eb + (Ep-R) When (Ep-R) ≧ 0, Tb2 (p) = Eb (Ep-R) <0 , The encoding order of the B picture (B1) immediately before B2 (p)
Assuming that the threshold of (p)) is Tb1 (p), Tb1 (p) = Dbskip + (Tb2 (p) −R) Dbskip + (Tb2 (p) −R) ≧ Eb, Tb1 (p) = Eb Dbskip + ( The video encoding device according to claim 10, wherein when Tb2 (p) -R) <Eb, 12. A video encoding device that encodes one frame or one field of image data while predicting a storage amount of a reception buffer in a decoding device, wherein the image type is changed to a picture type after encoding the image data. Accordingly, if M (appearance period of I or P picture) ≧ 2,
Threshold value Ti when image data is encoded into an I picture
Is Ti = D, where Di is the code amount of the I picture.
i, the threshold value Ti when the image data is encoded into a P picture
Is Tp = D, where Dp is the code amount of the P picture.
p, and the threshold value Tb (i) when the image data is encoded in the B picture immediately before the I picture, the code amount of the B picture is Db, and the predicted code amount of the I picture is Ei.
Tb (i) = Db + (Ei-R) When (Ei-R) ≧ 0, Tb (i) = Db (Ei-R) When <0, the threshold value Tb (p) when the image data is coded in the B picture immediately before the P picture is set as Code amount is Ep
Tb (p) = Db + (Ep−R) (Ep−R) ≧ 0, and Tb (p) = Db (Ep−R) where R is the transmission bit amount per decoding time interval of one picture. When <0, a threshold value setting unit that sets the following information; and, after encoding of the image data, a predicted accumulation amount that would be accumulated in the reception buffer at the decoding time of the encoded data of the image data. A comparing means for comparing the image data extracted from the reception buffer at the decoding time, in place of the encoded data of the image data taken out from the reception buffer at the decoding time, And a skip unit that uses a proxy code that specifies that the same image data as the data should be displayed. 13. A video encoding apparatus for encoding one frame of image data in a frame structure, wherein when encoding the image data, an accumulation amount of an output buffer or a predicted accumulation amount of a reception buffer of the decoding device. And a comparing means for comparing the storage amount or the predicted storage amount with a predetermined reference value.If it is determined that the storage amount or the predicted storage amount has reached a predetermined reference value, the encoding of the image data in a frame structure is stopped, Instead of the encoded data of the top field and the bottom field of the image data, a proxy code that specifies that two fields that are the same as the top field or the bottom field constituting the image data decoded in the past should be displayed. A video encoding device comprising: a skip unit to be used. 14. The skip unit according to claim 1, wherein, among the top field and the bottom field constituting the image data decoded in the past, the display order of each field of the image data for which the encoding in the frame structure has been stopped is the smallest. 14. The video encoding apparatus according to claim 13, wherein a proxy code that specifies that two fields that are the same as a near field are to be displayed is used. 15. The method according to claim 15, wherein said skipping means excludes the start and end of a slice layer when the picture type in which said image data whose coding in the frame structure has been stopped is to be coded is a B picture. All of the remaining macroblocks are all skipped B-pictures, which are B-pictures composed of skipped macroblocks. Two of the all-skip B-pictures are used as the proxy codes, and the image data whose coding in the frame structure is stopped is to be coded. If the picture type is I or P picture, all the other macroblocks except the first and last of the slice layer are all skipped P pictures, which are P pictures composed of skipped macroblocks. The video encoding device according to claim 14, wherein one of the proxy codes is the proxy code. 16. The all-skip P picture used when the picture type in which the image data whose coding in the frame structure has been stopped was to be coded is I or P picture, The bottom field of the image data of the I or P picture whose display order is earlier is used as a reference field, and the type of picture in which the image data whose coding in the frame structure has been stopped is to be coded is I or P picture In the all-skip B picture used in the case of the B picture (B1) encoded immediately after the frame picture, the bottom field of the image data of the I or P picture whose display order comes first is used as a reference field, and the encoding is performed in a frame structure. The type of picture in which the image data whose coding was to be stopped was to be I or P In the all-skip B picture used in the case of the B picture (B2) encoded immediately before the ch, the top field of the image data of the I or P picture whose display order is later is set as a reference field. The video encoding device according to claim 15. 17. The image processing apparatus according to claim 17, wherein the type of picture in which the image data whose coding in the frame structure has been stopped is to be coded is an I or P picture, and the skipping means then codes the B picture. Coding of the image data to be encoded is stopped, and instead of the encoded data of the top field and the bottom field of the image data, two all-skip B-pictures using the bottom field whose display order is earlier as a reference field are used. 18. The video encoding device according to claim 17, wherein: 18. A video encoding device that encodes one frame or one field of image data while predicting the amount of storage in a reception buffer in a decoding device, wherein the image data is encoded before encoding the image data. Comparing means for comparing a predicted storage amount predicted to be stored in a reception buffer at a decoding time of encoded data with a threshold value for each picture type in which image data is to be encoded; If the amount is less than the threshold value and the type of picture in which the image data is to be encoded is a B picture, the encoding of the image data is stopped and the image data is extracted from the reception buffer at the decoding time. In place of the encoded data of the image data, a proxy code that specifies that the same image data as previously decoded image data should be displayed is used. Video encoding apparatus characterized by comprising a means. 19. A video encoding device that encodes one frame or one field of image data while predicting the storage amount of a reception buffer in a decoding device, wherein the image data is encoded before encoding the image data. Comparing means for comparing a predicted storage amount predicted to be stored in a reception buffer at a decoding time of encoded data with a threshold value for each picture type in which image data is to be encoded; If the amount is less than the threshold value and the type of picture in which the image data is to be encoded is a B or P picture, the encoding of the image data is stopped and extracted from the reception buffer at the decoding time. A proxy code that specifies that the same image data as the previously decoded image data should be displayed is used in place of the encoded data of the image data. Video encoding apparatus characterized by comprising a control unit that. 20. A video encoding method for encoding one frame or one field of image data while predicting the storage amount of a reception buffer in a decoding apparatus, wherein the image data is encoded before encoding the image data. Comparing the predicted storage amount predicted to be stored in the reception buffer at the decoding time of the encoded data of the data with a predetermined threshold value, and when the predicted storage amount is less than the threshold value, The encoding of the image data is stopped, and the same image data as the image data decoded in the past is displayed instead of the encoded data of the image data as the data extracted from the reception buffer at the decoding time. Using a proxy code that defines what should be done. 21. A video encoding method for encoding one frame of image data in a frame structure, wherein when encoding the image data, a storage amount of an output buffer or a predicted storage amount of a reception buffer of a decoding device. And a predetermined reference value, and when it is determined that the accumulation amount or the predicted accumulation amount has reached a predetermined reference value, encoding of the image data in a frame structure is stopped, Instead of the encoded data of the top field and the bottom field of the data, a proxy code that specifies that the same two fields as the top field or the bottom field constituting the image data decoded in the past are to be displayed is used. And a video encoding method. 22. A computer for encoding one frame or one field of image data while predicting the amount of storage in a reception buffer in a decoding device, comprising the steps of: Comparing means for comparing a predicted storage amount predicted to be stored in the reception buffer at the decoding time of the coded data with a predetermined threshold value; and when the predicted storage amount is less than the threshold value, the image data That the same image data as the previously decoded image data should be displayed instead of the coded data of the image data as data taken out from the reception buffer at the decoding time. A video encoding program for functioning as a skip unit using a proxy code that defines the following. 23. A computer for encoding one frame of image data in a frame structure, wherein when encoding the image data, a computer stores a predetermined amount of storage in an output buffer or a predetermined amount of storage in a reception buffer of a decoding device. Comparing means for comparing a reference value; and when it is determined that the accumulation amount or the estimated accumulation amount has reached a predetermined reference value, encoding of the image data in a frame structure is stopped, and Skip means using a proxy code that specifies that two fields identical to the top field or the bottom field constituting the image data decoded in the past should be displayed instead of the encoded data of the top field and the bottom field And a video encoding program to function as.