JPH0451689A - Motion compensating predictive interframe coding device - Google Patents

Abstract

PURPOSE:To improve picture quality by measuring the fineness of the input television signal of a block to be intraframe-coded, and making quantization step size small for the block whose average value is larger than a definite value, and in addition, which is fine in a continuous block group to be quantized by the same quantization step size. CONSTITUTION:The fineness of every block the luminance signal of the input television signal has is measured. Then, as for the block which is the contiuous block to be quantized by first quantization step size 24 to be the same standard, and has a fine picture, the average value of whose picture element value is larger than the definite value, an orthogonal transformation factor 18 is quantized by second quantization step size 20 which is the first quantization step size 24 made small according to the average and the variance of the picture element value of the corresponding block and a threshold corrected by the first quantization step size 24. Thus, the improvement of the picture quality of the block with the fine picture can be attained while limiting the quantity of generated codes.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a motion compensated predictive interframe coding device for television signals.

Conventional technologyIn recent years, with the development of video encoding technology,
Video conference system, aD-rtoM, 7' Digital V
A motion compensated predictive interframe coding device has been developed as a highly efficient coding device for color video images used in TR and the like. For example, Multidimensional Signal Processing of rTV Images J by Takahiko Fukinuki (published November 16, 1988, Nikkan Kogyo Shimbunsha, Chapter 7 High Efficiency Coding, I) Il) 213-1) I)
A motion-compensated predictive interframe coding device described in No. 291) is known.

In order to achieve video encoding at a constant frame rate, the motion compensated predictive interframe coding device increases the prediction error or the quantization step or size of the pixel values of the input television signal when the amount of generated code is large. The amount of code generated is limited.

As a conventional method for determining the quantization step size, the
C, I, T, GXF Documents
s;s (C, C, 1, T, T, 5GXV document @625
Title: Description of
Ref, Mode18 (几M8), 5ourc
e: Working Party XV/4Sp
ecialist Group On Codi
ng for Visual Telephony
, version: June, 9.198
9'') is known.

Hereinafter, a conventional motion compensation predictive interframe coding device will be explained with reference to FIG.

In FIG. 2, 51 is an input terminal into which an input television signal is input, and 63 is a motion vector that calculates the motion vector of the encoded block by comparing the image signal of the encoded block of the current frame with the reproduced image signal of the previous frame. 54 is an image memory unit that stores the reproduced image signals of the current frame and the previous frame; 58 is a motion compensation prediction unit that performs motion compensation prediction on the reproduced image signal of the previous frame; and 60 is the block to be encoded within the frame. An inter-frame/intra-frame determination unit that determines whether to perform encoding or interframe encoding; 62 is an in-loop filter unit that performs two-dimensional low-pass filter processing on the motion compensation prediction signal; 64 is an original signal and prediction of the encoded block; a prediction error calculation unit that calculates a prediction error by calculating a difference between signals;
66 is a switch unit that selects a signal to be orthogonally transformed and a signal for calculating a reproduced image based on the encoding method selection signal, 68 is an orthogonal transform unit that orthogonally transforms the signal to be orthogonally transformed, and 0 is an orthogonal transform coefficient. 73 is a quantization step size calculation unit that calculates the quantization step size, 74 is a code memory unit that temporarily stores transmission frames, and 76 is an inverse orthogonal transform that performs inverse orthogonal transformation on the quantized orthogonal transform coefficients. An orthogonal transform unit; 78 is a reproduced image calculation unit that calculates a reproduced image of the current frame; 82 is a prediction error encoding unit that encodes a positive prediction difference through a channel; 84 is a motion vector encoder that encodes a motion vector through a channel. 8e is a multiplexer unit which configures a transmission frame from the prediction code and the vector code; 89 is an output terminal for outputting a transmission signal.

The operation of the above configuration will be explained below.

A television signal converted into a digital signal by an analog-to-digital conversion circuit (not shown) and divided into blocks of M pixels in the horizontal direction and N lines in the vertical direction is input as an input television signal 62 from an input terminal 61.

The motion vector calculation unit 63 compares the input television signal 62 with the reproduced television signal 66 of the previous frame stored in the image memory unit 54, calculates the motion of the encoded block as a motion vector, and generates a motion vector signal 66. Output to. At the same time, the motion vector calculation unit 63 uses the evaluation value at the time of motion vector calculation to determine whether motion compensation prediction is valid or invalid for the encoded block, and outputs the result as a motion compensation prediction control signal to the motion vector signal 66. . Therefore, the motion vector signal 66 has a motion vector and a motion compensated prediction signal superimposed thereon.

The motion compensation prediction unit 58 (1) performs motion compensation prediction on the reproduced television signal 66 of the previous frame using a motion vector when the motion compensation prediction control signal instructs to enable motion compensation prediction, and (2) When the motion compensation prediction control signal instructs invalidation of motion compensation prediction, the reproduced television signal 65 of the previous frame is output as is as the motion compensation prediction signal 69.

The inter-frame/intra-frame determining unit 60 compares the input television signal 52 and the motion compensation prediction signal 69 in units of blocks, determines the effectiveness of the motion compensation prediction, and if the effectiveness of the motion compensation prediction is small, It is determined that intraframe coding is effective for the relevant flock, and if the effectiveness of motion compensation prediction is large, it is determined that interframe coding is effective for the relevant flock, and the result is output as the encoding method selection signal 61. By switching the encoding method between intra-frame encoding and inter-frame encoding on a per-flock basis,
The following improvements can be made compared to encoding using only the interframe encoding method.

That is, (1) since intra-frame encoding is selected when a scene change occurs, the image quality after the scene change can be improved. (2) When a large movement of the moving object occurs, a background area hidden behind the moving object appears, and in this case, intra-frame encoding is selected, so that the image quality can be improved. Also, C
In the storage media encoding method used for D-ROM etc.,
In order to realize the editing function and reverse playback function of the reproduced image, a frame in which all blocks are intra-frame encoded at a fixed frame period (this frame is called a refresh frame J).
It is necessary to insert a refresh frame, and the insertion of a refresh frame can be realized by providing a motion compensated predictive interframe coding device with an intraframe coding function.

The in-loop filter section 62 performs two-dimensional low-pass filter processing on the encoded block subjected to motion compensation prediction using the motion vector, and calculates a prediction signal 63. The prediction error calculation unit 64 calculates the input television signal 6 of the encoded block.
2 and the prediction signal 63, and outputs the result as a prediction error signal 66.

(1) If the encoding method selection signal 61 selects intraframe encoding, the switch unit 66 selects the input television signal 62 as the signal 67 to be orthogonally transformed;
2) When the encoding method selection signal 61 selects interframe encoding, the prediction error signal 65 is selected as the signal 67 to be orthogonally transformed.

The orthogonal transform unit 68 performs orthogonal transform on the signal 6γ to be orthogonally transformed, removes high correlation between neighboring pixels of the signal 67 to be orthogonally transformed, and calculates orthogonal transform coefficients 69.

As the orthogonal transform method, in many cases, discrete cosine transform is used, which has high transform efficiency and can be implemented in hardware.

The quantization unit To quantizes the orthogonal transform coefficient 69 using the quantization step size 71 and calculates the orthogonal transform quantization coefficient 72.

The quantization step size calculation unit 73 calculates the quantization step size R1 from the residual code amount 76 in the code memory unit 74 using the method shown below.

The method for calculating the quantization step size 71 in this conventional example will be described below.

In this conventional example, the input television signal has a size of 35,252 pixels in the horizontal direction and 288 lines in the vertical direction, as shown in FIG. Block
) It is called l. ). The quantization step size Qb is calculated from equation (1) at the start of quantization with a period of n macroblocks. Qb=2xINT[Bcont+200q]+2-(
1) However, in equation (1), it is defined as follows.

(a) INT [' ] is a function that truncates the decimal point.

Example: INT [1,5] =1, INT [1,3]
=1, INT [1,6] =1 (b) Bcont is 0 indicating the residual code amount in the code memory section 74 (c) q is the encoding speed parameter, and the encoding speed ■ and Equation (2) There is a relationship between

V = q x 64 kbit / sec...
...(2) Example: V: Q x 64 kbit
/ sec, q=1.

As is clear from equation (1), the residual code amount BCont
As the number increases, the quantization step size Qb increases, the amount of generated code is limited, and video signal encoding at a constant frame rate can be realized. For example, quantization step size Q
When calculating b, residual code amount □B'ConL = 700
When the residual code amount Bcont is 6100 bits, the quantization step size Qb is 62. However, from the first macroblock to the (n-1)th macroblock, quantization is performed with a predetermined quantization step size Qb. For example, V = 64 kbit/5ec(
q=1), Qb=32.

In this conventional example, the calculation period n of the quantization step size Qb
Here, n=12.

The inverse orthogonal transform unit 76 performs inverse orthogonal transform on the orthogonal transform quantization coefficient 72 and calculates an orthogonally transformed signal 77 containing a quantization error.

The switch unit 66 (1) selects the numerical value "o" signal 8o as the reproduced image calculation signal 79 when the encoding method selection signal 61 selects intraframe encoding, and (2) selects the encoding method. When the signal 61 selects interframe coding, the predicted signal 63 is selected as the reproduced image calculation signal 79.

The reproduced image calculation unit 8 adds the orthogonally transformed signal 77 containing the quantization error and the reproduced image calculation signal element 9 to calculate the reproduced image 81 of the encoded block.

The image memory 44 stores the reproduced image signal 81 of the current frame and outputs the reproduced image signal 66 of the previous frame. The prediction error encoding unit 82 encodes the orthogonal transform quantization coefficient 72, the quantization step size 71, and the encoding method selection signal 61,
A prediction error code 83 is calculated. Quantization step size 7
1 is encoded only when the value of the quantization step size 71 changes, that is, once every n macroblocks.

The motion vector encoding unit 84 encodes the motion vector signal 56 and calculates a motion vector code 86. The multiplexer unit 86 calculates a transmission frame 8 in a predetermined format from the prediction error code 83 and the motion vector code 86.

The code memory unit 74 temporarily stores the transmission frame 8A, and outputs it from an output terminal 89 as a transmission code 88 in synchronization with a clock signal input from an external device (not shown). At the same time, the code memory unit 74 calculates the amount of codes remaining in the memory as the remaining code amount 76.

Problem to be Solved by the Invention However, in the above configuration, the quantization step size Qb is fixed because the block cycle period (n macroblock cycle period in the conventional example) for calculating the quantization step size is fixed. , the quantization step size Q is the same between n consecutive blocks regardless of the characteristics of the input television signal.
The orthogonal transform coefficients quantized by b are quantized. That is, within consecutive blocks belonging to the same quantization step size cycle, a block with a fine pattern to be intra-frame encoded has the same quantization step size Q as other blocks.
Since the block is quantized by b, there is a problem in that the image quality of intra-frame encoded blocks having minute patterns deteriorates.

In other words, by quantizing the orthogonal transform coefficients generated from intra-frame encoded blocks with fine nodules using a large quantization step size, the fineness of the original image is lost and the blocks become flat. "Block distortion" occurred, which was visually recognized as a major deterioration in image quality.

On the other hand, in order to reduce the amount of code generated by encoding the quantization step size, the number of consecutive blocks to be quantized with the same quantization step size must be greater than a certain value (in the conventional example, n macroblocks). , it is not possible to calculate and change the quantization step size for each block.

In view of the above-mentioned problems, the present invention provides a block that is a luminance signal block and is subjected to intra-frame encoding.
In consecutive blocks that are quantized using the standard first quantization step size calculated from the generated code amount, when the average value of pixel values is greater than a certain value, it is proportional to the image precision of the corresponding block. Then, a second quantization step size is calculated from the first quantization step size, and the orthogonal transform coefficients are quantized using the second quantization step size, thereby maintaining the precision of the original image. The purpose is to improve image quality as a result.

In other words, for blocks in which the average pixel value of the luminance signal block to be intra-frame encoded is a certain value or more and the precision is high, the actual quantization step size should be set smaller than the standard quantization step size. In this way, precision can be maintained while limiting the amount of generated code, and as a result, the image quality of the block can be improved.

Furthermore, when the first quantization step size is equal to or larger than a certain value, by correcting the threshold value for calculating the second quantization step size, the code amount of the luminance signal is reduced and the code amount of the color difference signal is reduced. It is possible to prevent deterioration of the color of the encoded block caused by an extremely small amount of code of the color difference signal. As a result, the image quality of the entire image can be improved.

Means for Solving the Problems To achieve this object, the present invention provides an A/D conversion means for analog-to-digital conversion of a television signal and a method for defining one frame or one field of the digitized input television signal. a blocking means for dividing each block into blocks of a certain size, a motion vector detection means for calculating a motion vector representing the motion of a television image for each block, and a determination as to whether to perform motion compensated prediction using the motion vector for each block. a motion compensation prediction means for performing motion compensation prediction on a reproduced image of a previous frame using a motion vector and calculating a predicted pixel value for a block to be motion compensated predicted; an error calculation means for calculating the difference between the predicted error value and the predicted error value; an encoding method determining means for determining whether to perform interframe encoding or intraframe encoding for each block; and an encoding method determination means for determining the encoding method for each block. a switching means for switching whether the signal to be orthogonally transformed is the pixel value of the input television signal or the prediction error value, based on the determination result of the means; orthogonal transform means for calculating a transform coefficient; first quantization step size determining means for calculating a first quantization step size from the generated code amount; and an average of pixel values for each block of a luminance signal of an input television signal. An average/variance calculation means for calculating variance, and a first quantization step size that is a predetermined threshold for classifying blocks that are luminance signal blocks and are to be intra-frame encoded based on the average or variance. For blocks that are luminance signal blocks and are to be intraframe encoded, the blocks are divided into classes based on the mean, variance, and corrected threshold, and for each class, the second quantization step size is determined based on the first quantization step size. A quantization step size is calculated, and for other blocks, using a second quantization step size determination means that sets the first quantization step size to a second quantization step size, and the second quantization step size, Quantization means for quantizing orthogonal transform coefficients and calculating quantized orthogonal transform coefficients, information on intra-frame coding or inter-frame coding, and information on the first quantization step size and quantization classification. and an encoding means for encoding the quantized orthogonal transform coefficients. A predicted pixel value obtained by motion-compensated prediction of a pixel value to be used when calculating a reproduced pixel value based on the determination result of the inverse quantization means that performs inverse orthogonal transformation on the orthogonal transform coefficients and calculates an inverse quantized signal, and the encoding method determination means. a switching means for switching whether to use the predicted pixel value or the numerical value "0" and the inverse quantization signal; and an image storage means for accumulating the reproduced image. , and motion vector encoding means for encoding the motion vectors.

Effects According to the present invention, with the above configuration, the precision of each block of the input television signal is determined by the variance σ of pixel values within the block.
It is thought that it can be measured with 2. For example, the variance σ2 of a block with a fine pattern is considered to be smaller than the variance σ2 of a "coarse" block with sharp changes in pixel values. Consider that σ2 becomes smaller.

Further, even if the variance is small, if the average value of pixel values is smaller than a certain value, it can be said that the block is hardly visually noticeable even if the precision is high.

Therefore, with the above configuration, the present invention measures the precision of each block of the luminance signal of the input television signal using the variance σ2, and measures the precision of each block of the luminance signal of the input television signal using consecutive blocks that are quantized using the same reference first quantization step size. , for a block that has a fine image to be intra-frame encoded and whose average pixel value is greater than a certain value, the first quantization step size is determined by the average and variance of the pixel values of the block and the first quantization step size. By quantizing the orthogonal transform coefficients with a second quantization step size that is smaller according to the threshold corrected by the quantization step size, the image quality of blocks with detailed images is improved while limiting the amount of generated code. It was made so that it could be done.

EXAMPLE Hereinafter, an example of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a motion compensated predictive interframe coding device according to the present invention.

1 is an input terminal into which an input television signal is input, and 3 is an input terminal that compares the picture signal of the coded block of the current frame with the reproduced picture signal of the previous frame to calculate the motion vector of the coded block and the motion compensation prediction control signal. A vector calculation unit; 4 is an image memory unit that stores the reproduced image signals of the current frame and the previous frame; 7 is a motion compensation prediction unit that performs motion compensation prediction for the reproduced image signal of the previous frame; 9 is a frame for the block to be encoded; An inter-frame/intra-frame determination unit that determines whether to perform inter-coding or intra-frame encoding; 11 is an in-loop filter unit that performs two-dimensional low-pass filter processing on the motion-compensated prediction signal; 13 is an original image signal of the encoded block; a prediction error calculation unit that calculates a prediction error by calculating a difference between the prediction signal and the prediction signal;
6 is a switch unit that selects a signal to be orthogonally transformed and a signal for calculating a reproduced image based on the encoding method selection signal; 17 is an orthogonal transform unit that performs orthogonal transform; and 19 is a quantizer that quantizes orthogonal transform coefficients. unit, 21 is a second quantization step size calculation unit, 2, which calculates the second quantization step size;
2 is a first quantization size calculation unit that calculates the first quantization step size; 26 is an average/variance value calculation unit that calculates the average and variance of the input television signal; and 28 is a code memory that temporarily stores transmission frames. 3o is an inverse orthogonal transform unit that performs inverse orthogonal transform on the quantized orthogonal transform coefficients, 34 is a reproduced image calculation unit that calculates a reproduced image of the current frame, and 36 is an encoding method selection signal, prediction error, and first quantization. step size,
and a prediction error encoding unit that encodes the quantization class information through a channel; 38, a motion vector encoding unit that encodes a motion vector through a channel; 4o, a multi-flexer unit that configures a transmission frame from a prediction code and a motion vector code; 43 is an output terminal for outputting a transmission signal.

The operation of the above configuration will be explained below.

The television signal is converted from analog to digital by a signal processing unit not shown in FIG. 1, and is divided into blocks of M pixels in the horizontal direction and N lines in the vertical direction.
The input television signal 2 is input as the input television signal 2.

Next, the motion vector calculation section 3 compares the input television signal 2 with the reproduced image 6 of the previous frame read out from the image memory section 4, calculates a motion vector, and outputs it as a motion vector signal 6. At the same time, the motion vector calculation unit 3 uses the evaluation value at the time of motion vector calculation to determine whether motion compensation prediction for the encoded block is valid or invalid, and outputs the result to the motion vector signal 6 as motion compensation prediction control information. do.

The motion compensation prediction unit 7 performs motion compensation prediction using a motion vector when motion compensation prediction is to be performed using the motion vector signal 6 on the reproduced image 6 of the previous frame at the same position as the encoded block, and does nothing when motion compensation prediction is not to be performed. motion compensation prediction signal 8
Output as .

The inter-frame/intra-frame determination unit 9 compares the input television control signal 2 and the motion compensation prediction signal 8 on a per-flock basis,
The effectiveness of motion compensation prediction is determined, and if the effectiveness of motion compensation prediction is small, intraframe coding is determined to be effective for the block, and if the effectiveness of motion compensation prediction is high, interframe coding is performed for the block. is determined to be valid, and outputs the result as the encoding method selection signal 1o.

In addition, in the case of an encoding device that requires insertion of refresh frames, the intra-frame/inter-frame determining unit 9 outputs an encoding method selection signal 10 so as to intra-frame encode all blocks at a constant frame period. .

The in-loop filter unit 11 performs in-loop filter processing, which is two-dimensional low-pass filter processing, on the motion compensation prediction signal 8 when the encoded block is a block to be predicted with motion compensation, and in other cases performs in-loop filter processing. It is output as the predicted signal 12 with .

The prediction error calculation unit 13 calculates the difference between the input television signal 2 and the prediction signal 12 of the encoded block, and outputs the result as a prediction error signal 14.

The switch unit 15 (1) selects the input television signal 2 as the signal 16 to be orthogonally transformed when the encoding method selection signal 10 selects intraframe encoding; (2)
When the encoding method selection signal 10 selects interframe encoding, the prediction error signal 14 is selected as the signal 16 to be orthogonally transformed.

The orthogonal transform unit 17 performs orthogonal transform on the signal 16 to be orthogonally transformed, removes high correlation between neighboring pixels of the signal 16 to be orthogonally transformed, and calculates orthogonal transform coefficients 18 .

As the orthogonal transform method, in many cases, discrete cosine transform is used, which has high transform efficiency and can be implemented in hardware.

The quantization unit 19 quantizes the orthogonal transform coefficients 18 with a second quantization step size of 2o. The method for calculating the second quantization step size 20 will be described below.

(1) When the block to be encoded is not a luminance signal block or is an inter-frame encoded block The second quantization step size section 21 uses the first quantization step size calculation section 22 according to the conventional example. The first quantization step size calculated from the residual code amount 23 using the described method is output as the second quantization step size.

(2) The block to be encoded is a luminance signal block,
If the block is an intra-frame encoded block, calculate the average and variance of the pixel values of the encoded block using the method shown below, classify the block based on the calculated average and variance, and use the standard for each class. A second quantization step size used in actual quantization is calculated from the first quantization step size.

The average/variance value calculation unit 26 calculates the average value μ and variance σ2 of the pixel values of the input television signal 2 of the encoded block by the (3rd
) and equation (4), and output as an average signal 44 and a variance signal 26. The average μ is used to determine whether to change the first step size depending on the precision of the block.

The variance σ2 has a small value in a block in which the input television signal 2 has high precision, and has a small value in degrees Celsius in a block in which the input television signal 2 has low precision.

However, equations (3) and (4) are defined as follows.

(a) M indicates the number of pixels in the horizontal direction of the block.

(b) N indicates the number of vertical lines of the block.

(c) P(i, j) indicates the pixel value of address (i, i) within the block.

The second quantization step size calculation unit 21 calculates a second quantization step size 20 and quantization class information 2 from the average signal 44 of the encoded block, the variance signal 26, and the first quantization step size 24. do. The first quantization step size 24 is calculated by the first quantization step size calculation unit 22
This is calculated from the code residual amount 23 in the code memory unit 28 using the method described in the conventional example. The second quantization step size calculation unit 21 first corrects a predetermined threshold value using the first quantization step size. The method of correction is as follows.

The first quantization step size Qb and a predetermined threshold thq are compared.

(1) Threshold values tb1, th2. used for classification when Qb <= th q. th 3゜t
Leave h4 unchanged without correcting it.

(2) When Qb > tllq Threshold th1 used for classification Correct th4 using the following formula.

th2. th3゜'''' + 2.3.4 max is the maximum value of the quantization step size Qb and the thresholds th1, th2, th3, th4
Figure 4 shows the relationship diagram.

Next, regarding a block which is a luminance signal and is to be intra-frame encoded, the average signal 44, the variance signal 26 and the threshold value th
1, th2, th3, th4, each block is 4
The second quantization step size 2o is calculated from the first quantization step size 24 according to the quantization class.

However, the first quantization step size is Qb, and the second quantization step size is Qstep.

(1) When o≦σ2〈thl and μ〉th4, Q 5t
ep' C1ass = 1Q 5tep = -
Q b (2) When th1≦σ2<th2 and μ>th4 Q
5step C1ass22 Qstep = -Qb (s) Q if th2≦σ2 th3 and μ> th4
5step C1ass=:3Qstep”
-Qb (4) If th3≦σ2 and μ)th4, then Q 5tep
C! 1ass = 4Qstep = Qb By doing the above, the second quantization step size 2o of the block that is a luminance signal block and is to be intraframe encoded is such that the average value of the pixels of the input television signal 2 is a constant value. For blocks that are above and have high precision, the first quantization step size is smaller than 24. In addition, when the first quantization step size exceeds a certain value, by correcting the classification threshold, the number of blocks that are quantized with a quantization step size smaller than the first quantization step size is reduced, and the luminance signal It is possible to reduce the code amount of the color difference signal and prevent deterioration of the color of the encoded block caused by an extremely small code amount of the color difference signal.

The quantization unit 19 quantizes the orthogonal transform coefficients 14 with a second quantization step size of 2o, and calculates orthogonal transform quantized coefficients 29. The inverse orthogonal transform unit 30 converts orthogonal transform quantization coefficients 2
9 is subjected to inverse orthogonal transformation, and a signal 31 containing a quantization error is calculated.

The switch unit 16 (1) selects the numerical value rOJ signal 33 as the reproduced image calculation signal 32 when the encoding method selection signal 10 selects intraframe encoding, and (2) selects the numerical value rOJ signal 33 as the reproduction image calculation signal 32. If interframe coding is selected, the predicted signal 12 is selected as the reproduced image calculation signal 32.

The reproduced image calculation unit 34 adds the signal 31 containing the quantization error and the reproduced image calculation signal 32 to calculate a reproduced image 36 of the encoded block. The image memory 4 stores the reproduced image signal 36 of the current frame and outputs the reproduced image signal 5 of the previous frame.

The prediction error encoding unit 36 includes an encoding method selection signal 10, a first quantization step size 24, and quantization class information 27.
, and the orthogonal transform quantization coefficients 26 to calculate a prediction error code 37.

The motion vector encoding unit 38 encodes the motion vector signal 6 of the block subjected to motion compensation prediction and calculates a motion vector code 39.

The multiplexer unit 4o calculates a transmission frame 41 in a predetermined format from the prediction error code 37 and the motion vector code 39.

The code memory section 28 stores the transmission frame 41 -times,
It is output from an output terminal 43 as a transmission code 42 in synchronization with a clock signal input from an external source (not shown). At the same time, the code memory unit 28 calculates the amount of codes remaining in the memory as the remaining code amount 23.

As is clear from the above description, according to this embodiment, when encoding a luminance signal block, intra-frame encoding is performed in a group of consecutive blocks belonging to the same quantization step size, and the average value of pixel values is For blocks that are above a certain value, the first reference value is
A second quantization step size is calculated from the quantization step size of can be achieved.

Also, when the first quantization step size exceeds a certain value, the number of blocks quantized with a quantization step size smaller than the first quantization step size is reduced by correcting the threshold for classification. As a result, the amount of code for the luminance signal is reduced, and as a result, the amount of code for the color difference signal is increased.
Deterioration of the color of the encoded block caused by an extremely small amount of code of the color difference signal can be prevented.

Note that in the above explanation, the variance value 26 calculated by the mean/variance value calculation unit 26 is defined by equation (3), but other measurement scales may be used as long as the variance can measure the fineness of the input television signal 2. But that's fine. For example, since the size of a block (horizontal pixel count: M, vertical line count: 2N) is generally a fixed value, the numerical value shown in equation (6) is available as a measure that is easy to calculate. However, p(i, D indicates the pixel value of address (i, j) in the block, and μ indicates the average pixel value of the block.

Further, in the above description, whether or not to change the first quantization step size is determined based on the average value of pixel values, but other measures expressing the characteristics of the entire pixel values may be used.

Further, in the above description, the quantization step size is classified into four classes, but the other classes may be divided into numbers.

Furthermore, in the above explanation, the first standard is used for each class.
The second quantization step size was determined by equally dividing the quantization step size, but if the second quantization step size is calculated to be smaller for blocks with small variance, another method But that's fine.

Furthermore, although the threshold value for classification is corrected using equation (6), the present invention has the effect of increasing the quantization step size. For intraframe-encoded blocks whose average value is a certain value or more within a group of consecutive blocks that are quantized with the same quantization step size and that have delicate patterns, By reducing the quantization step size,
Since video encoding can be performed while limiting the amount of generated code and without impairing the fineness of the original image, it is possible to improve the image quality.

In particular, the image quality improvement effect of refresh frames is remarkable.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a motion-compensated predictive interframe coding device according to an embodiment of the present invention, FIG. 2 is a block diagram of a conventional motion-compensated predictive interframe coding device, and FIG. 3 is a block diagram of a conventional motion-compensated predictive interframe coding device. FIG. 4 is a conceptual diagram showing the relationship between macro blocks and a characteristic diagram showing the relationship with the quantization step size of the motion compensated predictive interframe coding device in an embodiment of the present invention. 1.61... Input terminal, 3.63... Motion vector calculation unit, 4.64... Image memory unit, 58... Motion compensation prediction unit, 9.60... Intra-frame/inter-frame determination Section, 11.62... In-loop filter section, 13.64...
Prediction error calculation unit, 15.66... switch unit, 17.
68...Orthogonal transformation unit, 19.70...Quantization unit, 2
1 Second quantization step size calculation unit, 22.69... First quantization step size calculation unit, 26... Average/variance value calculation unit, 28, 4... Code memory unit, 30.76
... Inverse orthogonal transform unit, 34.78... Reproduction image calculation unit, 36.82... Prediction error encoding unit, 38.84...
・Input vector encoding unit, 41.86... Multiplexer unit, 43.89... Output terminal.

Claims (1)

[Claims] A/D converts television signals from analog to digital
a converting means, a blocking means for dividing one frame or one field of the digitized input television signal into blocks of a predetermined size, and a moving means for calculating a motion vector representing the movement of the television image for each block. a vector detection means, a motion compensation determination means for determining whether to perform motion compensation prediction using the motion vector for each block, and a block to be motion compensated predicted;
Motion-compensated prediction is performed on the reproduced image of the previous frame using a motion vector,
A motion compensation prediction means for calculating predicted pixel values; an error calculation means for calculating the difference between the pixel value of the input television signal and the predicted pixel value as a prediction error value; encoding method determining means for determining whether to perform inner encoding; and switching for each block whether the signal to be orthogonally transformed is to be a pixel value of an input television signal or a prediction error value based on the determination result of the encoding method determining means. a switching means; an orthogonal transform means for orthogonally transforming pixel values or prediction error values of an input television signal to calculate orthogonal transform coefficients; and a first quantizer for calculating a first quantization step size from the generated code amount. step size determining means; mean/variance calculation means for calculating the mean and variance of pixel values for each block of the luminance signal of the input television signal; means for correcting a predetermined threshold value by a first quantization step size in order to classify according to The blocks are divided into classes based on the corrected threshold value, and the second quantization step size is determined based on the first quantization step size for each class.
and for other blocks, a second quantization step size determination means that sets the first quantization step size to the second quantization step size; quantization means for calculating the quantized orthogonal transform coefficients, information on intra-frame coding or inter-frame coding, first quantization step size, and quantization classification. an encoding means for encoding the quantized orthogonal transform coefficients, an inverse quantization means for performing inverse orthogonal transformation on the quantized orthogonal transform coefficients and calculating an inverse quantized signal, and the encoding method determining means. A switching means for switching whether the pixel value to be used when calculating the reproduced pixel value is the predicted pixel value predicted by motion compensation or the numerical value "0" according to the determination result, and the predicted pixel value or the numerical value "0" and the dequantized signal. A motion compensated predictive interframe encoding device comprising an image reproduction means for calculating a reproduced image, an image storage means for accumulating the reproduced image, and a motion vector encoding means for encoding a motion vector.