JPH05308662A - High efficiency encoder - Google Patents

Abstract

(57) [Summary] [Purpose] A high-efficiency coding device that performs motion-compensated prediction.
As with the conventional motion compensation prediction for the luminance signal, the color difference signal uses the motion vector of the luminance signal to create a motion vector for the color difference signal, and by performing motion compensation prediction,
A high-efficiency encoder for reducing hardware scale. [Structure] The conventional motion-compensated prediction is performed on the Y signal. Here, in order to determine the motion vector of the C signal, the absolute value of the error between the reference block in this case and the output of the Y signal blocking circuit 51 together with the motion vector for the Y signal obtained by the Y signal motion compensation circuit 55. The sum is input to the C signal motion vector calculation circuit 64. The C signal motion vector calculation circuit 64 uses the C signal blocking circuit 6
Y corresponding to the block for the C signal output from 0
Of the signal motion vectors, the smaller sum of absolute values of the error signals is used as the motion vector of the color difference signal to perform motion compensation prediction.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital signal recording / reproducing apparatus such as a video tape recorder (hereinafter abbreviated as VTR) for digitally recording and reproducing a video signal and an audio signal, and a video signal recording / reproducing apparatus. The present invention relates to an apparatus for subjecting a signal to motion compensation prediction and compression encoding.

[0002]

2. Description of the Related Art FIG. 5 shows a block diagram of a high-efficiency coding apparatus using unidirectional motion-compensated interframe prediction. Reference numeral 1 is a digital video input terminal, 2 is a blocking circuit for blocking a digital video input signal into a block, 3 is a subtracter for outputting an error signal of an input block and a prediction block as an error block, and 7 is an encoding based on a determined mode. A first switch circuit for selectively outputting a block, and 8 is a discrete cosine transform (hereinafter abbreviated as DCT) which is an orthogonal transform in a coding block.
Is a DCT circuit for performing the above, 9 is a quantization circuit for quantizing DCT coefficients, 10 is a first encoding circuit for performing encoding suitable for a transmission path, and 11 is a transmission path.

Reference numeral 12 is an inverse quantization circuit for inversely quantizing the quantized DCT coefficient, 13 is an inverse DCT circuit for performing inverse DCT on the inversely quantized DCT coefficient, and 14 is an inverse DCT.
Adder that adds a prediction block to the decoding block that is the output signal of the above to generate an output block, 15 is an image memory that stores the output block for performing motion compensation prediction, and 16 is a past image stored in the image memory 15. An MC circuit for performing motion compensation prediction by performing motion detection from the motion compensation search block cut out from the current input block and the current input block.
A second encoding circuit for encoding the output of the circuit, and a second switch circuit 19 for switching the prediction block according to the mode of the discrimination circuit.

Next, the operation will be described. The input digital video signal is m [pixels] × n [lines] (m and n are positive integers) by the blocking circuit 2 regardless of an intra field for which motion compensation prediction is not performed and a prediction field for which motion compensation prediction is performed. Is divided into input blocks each having a unit of. In the input block, the difference in pixel units from the prediction block is calculated in the subtractor 3 to obtain the error block. In this way, the input block and the error block are input to the first switch circuit 7, respectively.

The first switch circuit 7 operates so that an input block is output when the processing screen is an intra field and an error block is output when the processing screen is a prediction field. The input block or error block is output as a coded block. The coding block selected by the first switch circuit 7 is converted into a DCT coefficient by the DCT circuit 8, and further weighting processing or threshold processing is performed by the quantizing circuit 9 to obtain a predetermined bit corresponding to each coefficient. Is quantized to a number. The quantized DCT coefficient is converted into a code suitable for the transmission line by the first encoding circuit 10 and output to the transmission line 11.

The quantized DCT coefficient enters the decoding loop 20 and reproduces an image for the next motion compensation prediction. The quantized DCT coefficient that has entered the decoding loop 20 is subjected to inverse weighting processing and inverse quantization in the inverse quantization circuit 12, and is further converted from the DCT coefficient into a decoded block in the inverse DCT circuit 13. The decoded block is added by the adder 14 to the prediction block output from the second switch circuit 19 pixel by pixel to restore the image. This prediction block is the same as that used in the subtractor 3. The output of the adder 14 is written in a predetermined position of the image memory 15 as an output block.

The image memory 15 has a different required memory amount depending on the prediction method. Assuming that it is composed of a plurality of field memories, the output block restored in the decoding loop is written in a predetermined position. From the image memory 15 to the MC circuit 16, a block that is a search range for motion detection that is cut out from a screen reconstructed from past output blocks is output.

The size of the search range block for motion detection is i [pixel] × j [line] (i ≧ m, j ≧ n:
i and j are positive integers. Data in the search range from the image memory 15 and the input block from the blocking circuit 2 are input to the MC circuit 16 as reference data, and the motion vector is extracted. There are various methods for extracting the motion vector, such as a full search block matching method and a tree search block matching method, which are well known and will not be described here.

The motion vector extracted by the MC circuit 16 is converted into a code suitable for the transmission line by the second encoding circuit 18, and is output to the transmission line together with the corresponding encoded block. Further, from the MC circuit 16, the prediction block is of the same size as the input block from the search range (m [pixel] ×
The block-shaped signal cut out to n [line]) is output. The prediction block output from the MC circuit 16 is generated from past image information. This prediction block is input to the second switch circuit 19. Here, a prediction block is output from the one output of the second switch circuit 19 to the subtractor 3 according to the processing field. The prediction block is output from the other output according to the decoding mode at that time.

As a prediction method performed by such a circuit block, for example, the one shown in FIG. 6 can be considered. In this method, an intra field is inserted every four fields and three fields in between are used as prediction fields. In FIG. 6, 40 is an intrafield, 41,
42 and 43 are prediction fields. The prediction in this method is performed from the first field 40 to the second field 40 of the intra field.
The field 41 is predicted, and similarly, the first field 40 to the third field 42 are predicted. Then, the reconstructed second field 41 to the fourth field 43 are predicted.

First, the first field 40 is blocked in the field and DCT is performed. Further, a weighting process and a threshold process are performed and quantized, and then encoded. In the decoding loop, the quantized first field signal is decoded / reconstructed. This reconstructed image is used for motion compensation prediction of the next second field 41 and third field 42. Next, the second field 41
Motion-compensated prediction is performed using the field 40, the obtained error block is subjected to DCT, and then encoded as in the first field 40. Further, the second field 41 is decoded / reconstructed in the decoding loop according to the mode signal of each block, and is used for the motion compensation prediction of the fourth field 43.
On the other hand, similarly to the second field 41, the third field 42 is also motion-compensated and predicted using the first field 40 and encoded. The fourth field 43 performs motion compensation prediction using the second field 41 reconstructed in the image memory 15.
The encoding is performed similarly to the third field 42.

[0012]

In such a conventional high-efficiency coding apparatus, when the sampling frequencies of the luminance signal and the color difference signals are different, such as a component digital signal of 4: 2: 2, a luminance signal block is generated. Since the area sizes of the color difference signal blocks are different, it is impossible to share the same motion vector, and it is necessary to obtain the motion vector for each of the luminance signal and the color difference signal. That is, two systems are required for the block configuration shown in FIG. 5, and there is a drawback that the hardware size becomes large. Further, if the area size of the blocks of the luminance signal and the color difference signal is the same, the luminance signal becomes twice as large as the block of the color difference signal, so that the prediction error of the motion compensation prediction becomes large, or in the orthogonal transformation due to the expansion of the orthogonal transformation block. The calculation error increases, and the image quality deteriorates.

[0013]

According to the present invention, a motion vector of a color difference signal is converted into a motion vector of a luminance signal in a high-efficiency coding apparatus which performs motion compensation prediction when the area size of a block of a luminance signal is different from that of a block of a color signal. A high-efficiency coding device with a small hardware scale is realized by obtaining the vector.

[0014]

According to the present invention, in motion compensation prediction when the area size of a block of a luminance signal is different from that of a block of a chrominance signal, the motion vector of the chrominance signal is obtained using the motion vector of the luminance signal, so that the image quality can be reduced on a small hardware scale. Performs information compression without noticeable deterioration.

[0015]

EXAMPLES Example 1. A first embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram of a first embodiment of the present invention. In FIG. 1, reference numeral 1 is a digital video input terminal, 51 is a block circuit for blocking the brightness signal input from the digital video input terminal 1, and 57 is a brightness signal for storing a reproduced image of the brightness signal as a reference image of the brightness signal. The image memory 55 for motion compensation performs the motion compensation prediction between the block output from the luminance signal blocking circuit 51 and the reference pattern output from the image memory 57 for luminance signal and outputs the motion vector and the prediction error. A circuit, 52 is a switch circuit for switching the output of the luminance signal blocking circuit 51 and the luminance signal motion compensation circuit 55 between the intra mode and the prediction mode, and 53 is orthogonal to the luminance signal block output from the switch circuit 52. An orthogonal transformation circuit that performs transformation, 54 is a quantization circuit that quantizes the output of the orthogonal transformation circuit 53, and 58 Inverse quantization circuit inverse-quantizes the output of the quantization circuit 54, 5
Reference numeral 9 is an inverse orthogonal transform circuit that performs an inverse orthogonal transform on the output of the inverse quantization circuit 58.

Reference numeral 60 is a color difference signal blocking circuit for dividing the color difference signal into a block, 64 is a color difference signal motion vector calculation circuit for obtaining a motion vector of the color difference signal from the motion vector of the luminance signal output from the brightness signal motion compensation circuit 55, 66
Is a color difference signal image memory that stores a reproduced image of the color difference signal as a reference image of the color difference signal, and 65 is a color difference that outputs an error between the output of the color difference signal blocking circuit 60 and the reference pattern output from the color difference signal image memory 66. A signal error calculating circuit 61 is a color difference signal blocking circuit 60 and a color difference signal error calculating circuit 6 in the intra mode and the prediction mode.
5 is a switch circuit for switching the output of 5; 62 is an orthogonal transform circuit for performing orthogonal transform on the color difference signal block output from the switch circuit 61; 63 is a quantization circuit for quantizing the output of the orthogonal transform circuit 62; An inverse quantizer circuit for inverse quantizing the output of the quantizer circuit 63, 68 is an inverse quantizer circuit 67
An inverse orthogonal transform circuit for performing an inverse orthogonal transform on the output of
Is an encoder for variable-length encoding the outputs of the luminance signal quantization circuit 54 and the color difference signal quantization circuit 63, and 11 is an encoder 69.
Output terminal of.

As a prediction method performed by such a circuit block, for example, the one shown in FIG. 6 can be considered. In this method, an intra field is inserted every four fields and three fields in between are used as prediction fields. In FIG. 6, 40 is an intrafield, 41,
42 and 43 are prediction fields. The prediction in this method is performed from the first field 40 to the second field 40 of the intra field.
The field 41 is predicted, and similarly, the first field 40 to the third field 42 are predicted. Then, the reconstructed second field 41 to the fourth field 43 are predicted.

First, the first field 40 is blocked in the field and DCT is performed. Further, a weighting process and a threshold process are performed and quantized, and then encoded. In the decoding loop, the quantized first field signal is decoded / reconstructed. This reconstructed image is used for motion compensation prediction of the next second field 41 and third field 42. Next, the second field 41
Motion-compensated prediction is performed using the field 40, the obtained error block is subjected to DCT, and then encoded as in the first field 40. Further, the second field 41 is decoded / reconstructed in the decoding loop according to the mode signal of each block, and is used for the motion compensation prediction of the fourth field 43.
On the other hand, similarly to the second field 41, the third field 42 is also motion-compensated and predicted using the first field 40 and encoded. The fourth field 43 performs motion compensation prediction using the second field 41 reconstructed in the image memory 15, and is coded in the same manner as the third field 42.

Next, the operation will be described. A luminance signal (Y signal) and two color difference signals (RY, BY) are input to the digital video input terminal 1, and a Y signal blocking circuit 51 and a color difference signal (C signal) blocking circuit 6 are provided, respectively.
At 0, regardless of whether it is an intra field or a prediction field, for example, block formation is performed with 8 pixels × 8 lines as one unit. In the Y signal motion compensation circuit 55, for the input block output from the Y signal blocking circuit 51 in the case of the prediction field, the intra field reproduced image data stored in the Y signal image memory 57 is used as a reference image for the motion vector. To detect.

The operation of motion vector detection in the Y signal motion compensation circuit 55 will be described. The motion vector detection is performed by the input block Xi (i, j) (i, i) in units of 8 pixels × 8 lines output from the Y signal blocking circuit 51.
j = 1 to 8), the motion vector search range is set to 16
The size of pixels × 16 lines is used. That is, when the input block Xi (i, j) and the position information in the field of the input block are input from the Y signal blocking circuit 51 to the Y signal motion compensating circuit 55, the Y signal motion compensating circuit 55 causes the Y signal motion compensating circuit 55 to operate. The position information in the field of the input block is output to 57. Intra-field reproduced image data is stored in the Y-signal image memory 57, and a reference becomes a motion vector detection range for the input block based on position information in the field of the input block output from the Y-signal motion compensation circuit 55. The block is output to the Y signal motion compensation circuit 55. Here, when the input block is Xi (i, j) (i, j = 1 to 8), the motion vector detection range is 16 pixels × 16 lines.
The signal image memory 57 uses Xr (i +
x, j + y) (x, y = -8 to 7; i, j = 1 to 8) is output.

In the Y signal motion compensation circuit 55, the reference data Xr (i + x, j) output from the Y signal image memory 57 is output.
+ Y) is used to find the motion vector by full search. Here, the motion vector detection method is as follows: input data Xi (i, j) and reference data Xr (i + x, j +) in a search range (x, y = -8 to 7) of 16 pixels × 16 lines, respectively.
The prediction error E (x, y) with respect to y) is calculated by the following equation.

[0022]

[Equation 1]

Further, in the search range (x, y = -8 to 7), E
(X0, y0) that minimizes (x, y) is obtained as a motion vector. Further, the Y signal motion compensation circuit 55 calculates the motion vector Mv (x0, y0) and the prediction error Xe in this case.
(I, j) Xe (i, j) = Xi (i, j) -Xr (i + x0, j
+ Y0) is output to the switch circuit 52.

In the switch circuit 52, the output Xi (i, j) of the Y signal blocking circuit 51 in the intra mode.
In the prediction mode, the Y signal motion compensation circuit 55 is selected.
The output Xe (i, j) of is selected and output to the orthogonal transformation circuit 53. The orthogonal transform circuit 53 performs, for example, two-dimensional discrete cosine transform on each input 8 × 8 block. The quantization circuit 54 quantizes the orthogonal transformation coefficient output from the orthogonal transformation circuit 53. Here, the quantization circuit 54
Then, in the intra mode, only the orthogonal transform coefficient is output to the encoder 69, but in the prediction mode, the motion vector information is output to the encoder 69 in addition to the orthogonal transform coefficient. The quantizing circuit 54 outputs to the encoder 69 a signal at the beginning of each field for identifying whether the field is in the intra mode or the prediction mode.

Further, the output of the quantizing circuit 54 is decoded by the inverse quantizing circuit 58 to be used as the reference data in the prediction mode. The inverse orthogonal transform circuit 59 performs two-dimensional discrete inverse cosine transform on the output of the inverse quantization circuit 58 to restore image data used as reference data. In the Y signal image memory 57, each block restored by the inverse orthogonal transform circuit 59 is used as reference data in the prediction mode to store two fields of the restored image. Further, the reference image data in the motion vector detection range is output to the Y signal motion compensation circuit 55.

Next, the motion vector detection method of the color difference signal will be described. Originally, when there is motion in the video signal, the luminance signal and the color difference signal move simultaneously, so if the sampling frequencies of the luminance signal and the color difference signal are the same, the motion vector of the color difference signal becomes the same as the motion vector for the luminance signal. Should be. However, in the case of a 4: 2: 2 component digital signal, the sampling frequency of the two color difference signals RY and BY is the same as the sampling frequency of the luminance signal.
It is 1/2. Therefore, as shown in FIG. 2, RY, B-
When 8 pixels × 8 lines are blocked for the Y signal, each area size is twice as large as the Y signal. Therefore, one C signal block has two Y signals. You can see that the blocks correspond.
However, the motion vector for the C signal block has a strong correlation with the corresponding motion vector for the two Y signal blocks. Therefore, even if motion compensation prediction is performed using one of the motion vectors for these two Y signals as the motion vector of the C signal, it can be said that the prediction error of the C signal is sufficiently small. Further, of the two Y signal motion vectors corresponding to the C signal block, the smaller prediction error for the Y signal block is considered to be more appropriate as the motion vector for the C signal block. Therefore, in the present invention, of the two motion vectors for the Y signal obtained by the Y signal motion compensation circuit 55, the one having the smaller prediction error for the Y signal block is selected as the motion vector of the C signal.

The motion vector for the C signal is calculated by the C signal motion vector calculation circuit 64. The operation of the C signal motion vector calculation circuit 64 will be described below.
In the Y signal motion compensation circuit 55, together with the motion vector component Mv (x0, y0) of the Y signal, the Y signal blocking circuit 5
The absolute value sum Er of the error component between the output Di (i, j) of 1 and the reference block Dr (i + x0, j + y0) indicated by the motion vector of the Y signal is

[0028]

[Equation 2]

Then, the absolute value sum Er of the error components is calculated by C
It is output to the signal motion vector calculation circuit 64. Where C
In the signal motion vector calculation circuit 64, RY,
Of the motion vectors of the two Y signals corresponding to the block of the BY signal, the sum of absolute values of prediction errors is compared, and the smaller motion vector is set as the motion vector of the C signal. That is, the motion vectors of the Y signal corresponding to the blocks of the C signals are Mv1 and Mv2, and the absolute value sums Er1 and E of the error components thereof.
If r2, then Mv1 is Er1 when Er1> Er2.
When ≦ Er2, Mv2 is selected as the motion vector of the entire block of the C signal.

The C signal error calculation circuit 65 predicts the reference block output from the C signal image memory 66 and the block output from the C signal blocking circuit 60 according to the motion vector obtained by the C signal motion vector calculation circuit 64. The error is obtained and output to the switch circuit 61 together with the motion vector. Here, the motion vector of the C signal is Y
Since either one of the two motion vectors Mv1 and Mv2 for the signal is used, Mv1 or Mv for the C signal is used.
Only a control signal indicating which motion vector of v2 is selected is output as a motion vector. Further, the operations up to the switch circuit 61, the orthogonal transformation circuit 62, and the quantization circuit 63 for the C signal and the operations of the inverse quantization circuit 67 and the inverse orthogonal transformation circuit 68 are the same as in the case of the Y signal, and therefore will be omitted.

Next, the output of the quantizing circuit 54 for the Y signal and the output of the quantizing circuit 63 for the C signal are the encoder 69.
Entered in. In the encoder 69, the data of the Y and C signals are variable length coded and output to the transmission line 11.

In the above embodiment, the motion vector of the C signal is selected as the motion vector of the C signal from the motion vectors of the two Y signals, whichever has the smaller prediction error in each Y signal block. However, it is also possible to calculate the respective prediction errors using the motion vectors of the two Y signals and use the smaller prediction error as the motion vector of the C signal. Also, the average of the two motion vectors for Y may be selected as the motion vector of the C signal. However, in this case, the motion vector of the C signal is Y in the decoding system.
Since it can be combined from the motion vector of the signal, the motion vector information for the C signal need not be transmitted.

In the above embodiment, the sampling frequency of the two color difference signals of the input signal is 1/2 times that of the luminance signal. However, the sampling frequency is not necessarily 1/2 times, and the sampling frequency is 1 / n times. You can do it in. For example, when the color difference signal is 1/4 times the luminance signal, one color difference signal block corresponds to four luminance signal blocks as shown in FIG.
Therefore, the C signal motion vector calculation circuit 64 selects the optimum one of the four Y signal motion vectors corresponding to each C signal as the C signal motion vector.

Although the C signal is not line-sequential in the above embodiment, the C signal may be line-sequential at an arbitrary interval. For example, when the color-difference signal is subjected to line-sequencing every two lines, one color-difference signal block corresponds to four luminance signal blocks as shown in FIG. Therefore, the C signal motion vector calculation circuit 64 may select the optimum one of the four Y signal motion vectors corresponding to each C signal as the C signal motion vector.

In the above embodiment, the block size of the orthogonal transformation is set to 8 pixels × 8 lines, but it is not necessarily 8 pixels × 8 lines, and the block size of n pixels × m lines is used. Good. Similarly, the detection range of the motion vector does not have to be 16 pixels × 16 lines, but may be k pixels × l lines (k ≧ n, l ≧ m). Although the predictive coding is completed every four fields, the predictive coding is not necessarily limited to four fields, and the predictive coding may be completed every arbitrary field.

[0036]

As described above in detail, in the high-efficiency coding apparatus of the present invention, the motion vector of the color signal is used to perform the motion-compensated prediction.
The hardware size of the arithmetic circuit or the like for obtaining the motion vector can be reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is a diagram showing block sizes of a luminance signal and a color difference signal in the first embodiment of the present invention.

FIG. 3 is a diagram showing block sizes of a luminance signal and a color difference signal according to another embodiment of the present invention.

FIG. 4 is a diagram showing block sizes of a luminance signal and a color difference signal in another embodiment of the present invention.

FIG. 5 is an explanatory diagram of motion compensation prediction in a conventional encoding device.

FIG. 6 is an explanatory diagram of motion compensation prediction.

[Explanation of symbols]

 1 Digital Image Input Terminal 11 Transmission Line 51 Luminance Signal Blocking Circuit 52 Luminance Signal Switch Circuit 53 Luminance Signal Quadrature Conversion Circuit 54 Luminance Signal Quantization Circuit 55 Luminance Signal Motion Compensation Circuit 57 Luminance Signal Image Memory 58 Luminance Signal Dequantization Circuit 59 Luminance signal inverse orthogonal transform circuit 60 Color difference signal blocking circuit 61 Color difference signal switch circuit 62 Color difference signal orthogonal conversion circuit 63 Color difference signal quantization circuit 64 Color difference signal motion vector calculation circuit 65 Color difference signal error calculation circuit 66 Color difference signal image Memory 67 Color difference signal inverse quantization circuit 68 Color difference signal inverse orthogonal transformation circuit

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] August 17, 1992

[Procedure Amendment 1]

[Document name to be amended] Statement

[Name of item to be corrected] 0002

[Correction method] Change

[Correction content]

[0002]

2. Description of the Related Art FIG. 5 shows a block diagram of a high-efficiency coding apparatus using unidirectional motion-compensated interframe prediction. Reference numeral 1 is a digital video input terminal, 2 is a blocking circuit for blocking a digital video input signal into a block, 3 is a subtracter for outputting an error signal of an input block and a prediction block as an error block, and 7 is an encoding based on a determined mode. A first switch circuit for selectively outputting a block, and 8 is a discrete cosine transform (hereinafter abbreviated as DCT) which is an orthogonal transform in a coding block.
DCT circuit which performs, 9 quantization circuit for quantizing the DCT coefficients, the 10 first encoding circuit which performs the coding suitable for a transmission path, 11 is a transmission path.

[Procedure Amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0005

[Correction method] Change

[Correction content]

[0005] When the first switch circuit 7 processes the screen intra field, the input block is output, the dynamic so that the error block is outputted in the case of predictors
To make. The coding block selected by the first switch circuit 7 is converted into a DCT coefficient by the DCT circuit 8, and further weighting processing or threshold processing is performed by the quantizing circuit 9 to obtain a predetermined bit corresponding to each coefficient. Is quantized to a number. The quantized DCT coefficient is converted into a code suitable for the transmission line by the first encoding circuit 10 and output to the transmission line 11.

[Procedure 3]

[Document name to be amended] Statement

[Correction target item name] 0018

[Correction method] Change

[Correction content]

First, the first field 40 is blocked in the field and DCT is performed. Further, a weighting process and a threshold process are performed and quantized, and then encoded. In the decoding loop, the quantized first field signal is decoded / reconstructed. This reconstructed image is used for motion compensation prediction of the next second field 41 and third field 42. Next, the second field 41
Motion-compensated prediction is performed using the field 40, the obtained error block is subjected to DCT, and then encoded as in the first field 40. Further, the second field 41 is decoded / reconstructed in the decoding loop according to the mode signal of each block, and is used for the motion compensation prediction of the fourth field 43.
On the other hand, similarly to the second field 41, the third field 42 is also motion-compensated and predicted using the first field 40 and encoded. The fourth field 43 performs motion compensation prediction using the second field 41 reconstructed in the image memory 15 .
The encoding is performed similarly to the third field 42.

Claims (3)

[Claims] 1. A device for dividing a digitized video signal into blocks of a predetermined size and coding by using motion compensation prediction, when the area sizes of the motion compensation blocks of a luminance signal and a chrominance signal are different. A high-efficiency coding device characterized in that a motion vector of a color signal is obtained from a motion vector of a luminance signal. 2. An apparatus for dividing a digitized video signal into blocks of a predetermined size and coding by using motion compensation prediction, wherein the area size of a motion compensation block of a chrominance signal is motion compensation of a luminance signal. When the block area size is n times (n ≧ 2), the motion vector of the chrominance signal is the motion vector for the n luminance signal blocks for each chrominance signal block that minimizes the prediction error in each luminosity signal block. A high-efficiency coding device characterized by: 3. An apparatus for dividing a digitized video signal into blocks of a predetermined size and encoding by using motion compensation prediction, wherein the area size of the motion compensation block of the chrominance signal is motion compensation of the luminance signal. When the block area size is n times (n ≧ 2), the motion vector of the chrominance signal is the motion vector of the n chrominance signal blocks that minimizes the prediction error of the chrominance signal block. A high-efficiency coding device characterized by obtaining.