JPH04200191A - Moving compensation prediction inter-frame coder - Google Patents

Abstract

PURPOSE:To improve the picture quality of a reproduced picture by applying numeral conversion taking an orthogonal transformation coefficient as a quantization step size so as to prevent that an orthogonal transformation coefficient having a size of some degree or larger after quantization becomes '0'. CONSTITUTION:A quantization section 23 uses a quantization step size 21 to quantize an orthogonal transformation coefficient 22 subject to numeral conversion to calculate an orthogonal transformation quantization coefficient 24. When the orthogonal transformation coefficient and the quantization step size are equal, the quantization section 23 calculates '1' or '-' as an orthogonal transformation quantization coefficient 24. A prediction error coding section 33 codes a coding system selection signal 10, a quantization step size 21 and an orthogonal transformation quantization coefficient 24 to calculate a prediction code 34. At the quantization of the orthogonal transformation coefficient, since the orthogonal transformation coefficient is smaller slightly than the quantization step size, the orthogonal transformation coefficient having been quantized to a numeral '0' is left after the quantization, then the picture quality of the reproduced picture is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a motion-compensated predictive interframe coding device for television signals.
ti video conference system CD-ROM.

Although a motion-compensated predictive interframe coding device has been developed as a high-efficiency coding device for color video images used in digital VTRs, etc. Newspaper publisher Chapter 7 High Efficiency Code 4#:, pp213
Although the motion-compensated predictive interframe coding device described in ``-pp291'' is known, the motion-compensated predictive interframe coding device is capable of realizing video coding at a constant frame rate when the amount of generated code is large. Even if the amount of generated code is limited by increasing the prediction error or the quantization step size of the image value of the input television signal, the conventional method for determining the quantization step size is based on the ``Description of Reference Modes Source'' ( C, C, I, T, T,
5GxV document #525” title: Dis
cri-ption of Ref, Mod
el 8 (RM8), 5source: Wo
rkingParty XV/4 5speciali
stGroup On Coding for
V-isual Te1ephony, versi
-on: June, 9. A motion-compensated predictive interframe coding device described in 1989'' is known.Hereinafter, the conventional motion-compensated predictive interframe coding device will be described with reference to FIG. 2. In FIG. 53 is an input terminal to which an input television signal is input; 53 is a motion vector calculation unit that compares the image signal of the encoded block of the current frame with the reproduced image signal of the previous frame to calculate the motion vector of the encoded block; 54 is the current frame 58 is a motion compensation prediction m which performs motion compensation prediction on the reproduction image signal of the previous frame.
60 is an inter-frame code that determines whether a block to be encoded is to be intra-frame encoded or inter-frame encoded;
The intra-frame judgment test 62 is 2 for the motion compensated prediction signal.
In-loop filter for dimensional low-pass filtering 6
4 is a prediction error calculation function that calculates a prediction error by calculating the difference between the original picture signal and the predicted signal of the coding block. 66 is a coding method selection symbol that selects a signal to be orthogonally transformed and a signal for calculating a reproduced image. The switch that makes the selection
68 is an orthogonal transform function that orthogonally transforms a signal to be orthogonally transformed. 70 is a quantization function that quantizes orthogonal transform coefficients. 73 is a quantization step size calculation function that calculates a quantization step size. 74 is a transmission frame 76 is an inverse orthogonal transform processor that performs inverse orthogonal transform on the quantized orthogonal transform coefficients; 78 is a reproduced image calculator that calculates the reproduced image of the current frame; and 82 is a prediction error that encodes the prediction error into channel coding. Coding block 84 is a motion vector coding block which encodes a motion vector for communication.86 is a multiplexer block which configures a transmission frame from a prediction code and a motion vector code.
Reference numeral 9 denotes an output terminal for outputting a transmission signal, and in the configuration shown above, its operation will be explained below.It is converted into a digital signal by an analog-to-digital conversion circuit (not shown).M pixels in the horizontal direction and N in the vertical direction. Television signal C divided into blocks of lines
-t The 4-motion vector calculation unit 53, which receives the input television signal 52 from the input terminal 5I, compares the input television signal 52 with the reproduced television signal 55 of the previous frame stored in the image memory unit 54, and encodes it. The motion of the block is calculated as a motion vector and output as a motion vector signal 56. At the same time, the motion vector calculation unit 534 determines whether motion compensation prediction is valid or invalid for the encoded block using the evaluation at the time of motion vector calculation. The result is output as a motion compensation prediction control signal to the motion vector signal 56. Therefore, even if the motion vector and the motion compensation prediction signal are superimposed on the motion vector signal 56, the motion compensation prediction unit 58 performs (1) motion compensation prediction control. (2) If the signal instructs to enable motion compensation prediction, perform motion compensation prediction on the reproduced television signal 55 of the previous frame using a motion vector; (2) If the motion compensation prediction control signal instructs to disable motion compensation prediction; If so, the inter-frame/intra-frame determining unit 60 outputs the reproduced television signal 55 of the previous frame as it is as a motion-compensated predicted signal 59. Then, determine the effectiveness/invalidity of motion compensation prediction, and (1
) If the effectiveness of motion compensation prediction is small, it is determined that intraframe coding is effective for the corresponding block; (2) If the effectiveness of motion compensation prediction is large, it is determined that interframe coding is effective for the corresponding block; The result is output as the encoding method selection signal 61.By switching the encoding method between the intraframe encoding method and the interframe signaling method in block units, the following is compared to the case where encoding is performed using only the interframe encoding method. (1) When a scene change occurs, intra-frame coding is selected, so it is possible to improve the image quality after the scene change. (2) When a large movement of a moving object occurs, it is hidden behind the moving object. A background area appears, and in this case, intra-frame coding is selected to improve the image quality.
In the storage media encoding method used for the
It is necessary to insert a refresh frame (referred to as refresh frame-me)J), and the insertion of a refresh frame can be realized by equipping the motion compensation predictive interframe coding device with an intraframe coding machine.The in-loop filter unit 62 Two-dimensional low-pass filter processing is not performed on the encoded block subjected to motion compensation prediction using the motion vector. A prediction error calculation unit 64 calculates the predicted signal 63. The input television signal 52 and the predicted signal of the encoded block are 63 does not perform the difference calculation\ The result is output as the prediction error signal 65. Method (1) If the encoding method selection signal 61 selects intra-frame encoding, as the signal 67 to be orthogonally transformed. Select input television signal 52 and press (
2) When the encoding method selection signal 61 selects inter-frame signaling, the orthogonal transform unit 68 selects the prediction error signal 65 as the signal 67 to be orthogonally transformed. The orthogonal transform coefficient 69 is calculated by removing the high correlation between neighboring pixels of the signal 67 to be orthogonally transformed.In most cases, the orthogonal transform method has high transform efficiency.About hardware implementation Even if a feasible discrete cosine transform is used, the quantization unit 70 uses the quantization step size 71,
The orthogonal transform coefficient 69 is quantized and the orthogonal transform quantized coefficient 72 is
If the orthogonal transform coefficient and the quantization step size are equal, the quantization step size calculation unit 73 calculates "1" or "-1" as the orthogonal transform quantization coefficient 72. The quantization step size 71 is calculated from the residual code amount 75 in the code memory section 74 using the method shown in FIG. John signal 4 top horizontal direction 3
It has a size of 52 strokes and 288 lines in the vertical direction, and an area of 16 strokes in the horizontal direction and 16 lines in the vertical direction (in this conventional example, it is called "micro block").
Divide into 4 quantization step size Qb4in microblock period, and calculate from the formula shown in equation (1) at the start of quantization: Qb = 2X'rNT [Bcot-;
+ 2 However, in formula (1), define as follows (a) INT [*] L'L Function that rounds down the decimal places Example: INT [1, 5] = 1, INT [1,'3
] = 1, INT [1, 6] = 1 (b) Bcont indicates the remaining amount of the code memory section 74. (c) q is the encoding speed parameter, and the relationship between the encoding speed V and equation (2) is V = qX64 k b
i t / s e c −−−−−−−−−−−(
2) Example: When V = 64 kbit/sec, q = 1, and as is clear from equation (1), the residual code amount Bcon
As t increases, the quantization step size Qb increases and the generated code amount is limited, making it possible to encode a video signal at a constant frame rate.For example, when calculating the quantization step size Qb, the residual code amount Bconc = 700bi
When t, the quantization step size Qb = 8, and the residual code amount Qb = 62 h Bcont = 6100bi
When t, the quantization step size Qb=62.However, from the first microblock to the (n-1)th microblock, quantization is performed with a predetermined quantization step size Qb.

For example, CL V==64kbit/sec (q=1)
In case (7), when Qb=32, in the conventional example of Nobomoto, the calculation period n of the quantization step size Qb
Assuming that cL n=12, the inverse orthogonal transform unit 76 performs inverse orthogonal transform on the orthogonal transform and quantization number 72, and calculates the orthogonal transform signal 77 including the quantization error.4 Switch unit 664:L (1) Encoding method Selection signal 6
1 selects intraframe encoding, selects the numerical value "0" signal 80 as the reproduced image calculation signal 79, and (
2) If the encoding method selection signal 61 selects interframe coding, the predicted signal 63 is selected as the reproduced image calculation signal 79. and the reproduced image calculation signal 79,
The 4-image memory 54 that calculates the reproduced image signal 81 of the encoded block stores the reproduced image signal 81 of the current frame, and the prediction error encoder 82 that outputs the reproduced image signal 55 of the previous frame calculates the orthogonal transform quantization number 72. , quantization step size 71, a quantization step size 7 that encodes the encoding method selection signal 61 and calculates the prediction error code 83.
1 is encoded when the value of the quantization step size 71 changes, that is, only once for n microbolts.The motion vector encoding unit 84 encodes the motion vector 56,
A multiplexer unit 86 that calculates a motion vector code 85
calculates a transmission frame 87 in a predetermined format from the prediction error code 83 and the motion vector code 85. The code memory section 74 stores the transmission frame 87 over time.
The code memory section 74 outputs the transmission code 88 from the output terminal 89 in synchronization with a clock signal inputted from an external source (not shown).At the same time, the code memory section 74 calculates the amount of codes remaining in the memory as the residual code amount 75. Problems to be Solved by the Invention However, in the above configuration, the quantization step size Q
In order for all orthogonal transform coefficients less than or equal to b to become rQJ after quantization (1) The quantization step size Qb is large,
(2) When there are many orthogonal transform coefficients with small differences between the orthogonal transform coefficients and the quantization step size, "block distortion" and "mosquido noise", which are deterioration of the image quality of the reproduced image due to quantization, occur. There was a problem because if the orthogonal transform coefficient has a visually important meaning and its value is even slightly smaller than the quantization step size Qb, the orthogonal transform coefficient after quantization becomes "0". Especially when the quantization step size Qb is large, many orthogonal transform coefficients become rOJ after quantization, so the quality of the reproduced image deteriorates significantly. ~ The present invention takes into consideration the above-mentioned problems. Before quantizing orthogonal transform coefficients, each orthogonal transform coefficient value 1Coef1 and quantization step size Qb are compared, and among the orthogonal transform numbers that become "0" after quantization, For orthogonal transform coefficients for which the difference between the orthogonal transform coefficient and the quantization step size is less than or equal to a predetermined threshold Th, by performing numerical transformation using the orthogonal transform system value 1coefl as the quantization step size, it is possible to convert the orthogonal transform coefficient to a size larger than a certain level after quantization. Means for solving the problem In order to achieve the above object, the technical solution CL of the present invention is to prevent the orthogonal transform coefficient from becoming "0" and improve the image quality of the reproduced image.
an analog-to-digital conversion means for converting a television signal from an analog signal to a digital signal; a blocking means for dividing one frame or one field of the digitized input television signal into blocks of a predetermined size; and individual blocks. a motion vector calculation means for calculating a motion vector which is the motion of a television image; a motion compensation determination means for determining whether to perform motion compensation prediction using the motion vector for each block; A motion compensation predicted pixel calculation means that performs motion compensation prediction on a reproduced image of a frame using a motion vector and calculates a predicted pixel value, and a prediction error that calculates the difference between the pixel value of the input television signal and the predicted pixel value as a prediction error value. a calculation means, a coding method determination means for determining whether to perform interframe encoding or intraframe encoding for each block, and an orthogonal encoding method based on the intraframe encoding/interframe encoding determination result for each block. orthogonal transformation signal selection means for switching a signal to be converted into pixel values of an input television signal or prediction error values; and orthogonal transformation of the pixel values or prediction error values of the input television signal to calculate orthogonal transformation coefficients. An orthogonal transform means, a quantization step size calculation means for calculating a quantization step size from the generated code amount, and a quantization step size calculation means that compares each orthogonal transform coefficient with the quantization step size, For orthogonal transform coefficients that become zero after quantization, orthogonal transform coefficient converting means for numerically converting the corresponding orthogonal transform coefficient value 1coef1 into a value equal to the quantization step size, and all orthogonal transforms including the numerically converted orthogonal transform coefficients. An orthogonal transform coefficient quantization means for quantizing coefficients using a quantization step size, information on intra-frame encoding or inter-frame encoding, a motion vector value, a quantization step size, and a quantized orthogonal transform coefficient. an encoding means for encoding; an inverse orthogonal transformation means for inverse orthogonal transforming the quantized orthogonal transform coefficients and calculating an inverse quantized signal; The power to make the pixel value to be used the predicted pixel value obtained by motion compensation prediction＼
Reproduction pixel selection means for setting or switching the numerical value "0";
The above objective can be achieved by a motion compensated predictive interframe coding device comprising a reproduced pixel calculation means for calculating a reproduced image from a predicted pixel value or numerical value "0" and a dequantized signal, and a reproduced pixel accumulation means for accumulating the reproduced image. The direct light conversion coefficient Coef, which is generally subjected to quantization processing, is
When we measure the distribution of the orthogonally transformed images, we find that the intra-frame image is L, the inter-frame image is L (1), many images with large coefficient values 1coefl are distributed in the low frequency component region, and (2
) There are many small coefficient values 1coef l distributed in the high frequency component region.Therefore, the orthogonal transform coefficient value 1coef1 that is converted to the numerical value "0" after quantization is the orthogonal transform coefficient value belonging to the high frequency component region. In many cases, it is 1coefl ~ If the high frequency component of the orthogonal transform coefficient 1coe-fl of the image within the frame becomes a value of 0.J due to quantization, the reproduced image is considered to be a blurred image, and the a frame If the high-frequency component of the orthogonal transform coefficient 1coefl of the intermediate image becomes the numerical value rOJ by quantization, it is thought that image quality deterioration such as "monoskeet noise" will occur in the reproduced image. Before quantizing the orthogonal transform coefficients, compare each orthogonal transform coefficient value I C0efl with the quantization step size Qb for the orthogonal transform number that becomes "0" after quantization. For the orthogonal transform coefficients for which the difference between the two is less than or equal to the predetermined threshold Th, the orthogonal transform coefficient value 1coef
By performing numerical conversion with l as the quantization step size, this method prevents orthogonal transform coefficients having a certain size or more from becoming "O" after quantization, and improves the image quality of reproduced images. Embodiment Hereinafter, an embodiment of the present invention will be described with reference to FIG. 1. FIG. 1 is a block diagram of a motion compensated predictive interframe coding apparatus in an embodiment of the present invention. In, 1 is an input terminal into which an input television signal is input, and 3 is an input terminal to which an input television signal is input, and 3 is an input terminal that compares the image signal of the encoded block of the current frame and the reproduced image signal of the previous frame to calculate the motion vector of the encoded block and the motion compensation prediction control signal. The motion vector calculation unit 4 uses an image memory memory that stores the reproduced image signal of the previous frame.
7 is a motion compensation prediction unit that performs motion compensation prediction on the reproduced picture signal of the previous frame; 9 is an inter-frame/intra-frame judgment ftL that determines whether a block to be encoded is to be intra-frame encoded or intra-frame encoded; 11 is an inter-frame/intra-frame determination ftL An in-loop filter unit 13 that performs two-dimensional low-pass filter processing on a motion-compensated prediction signal is a prediction error calculation unit 1 that calculates a prediction error by calculating a difference between the original signal and the prediction signal of the encoded block.
5 is a switch that selects a signal to be orthogonally transformed and a signal for calculating a reproduced image according to a coding method selection signal; 17 is an orthogonal transform switch that performs orthogonal transform; and 19 is an orthogonal transform coefficient transform that numerically transforms orthogonal transform coefficients. 20 is a quantization size calculation unit that calculates the quantization step size; 23 is an orthogonal coefficient quantization unit that quantizes numerically converted orthogonal transform coefficients; and 25 is a code memory unit 2 that temporarily stores transmission frames.
7 is an inverse orthogonal transform function that performs inverse orthogonal transform on quantized orthogonal transform coefficients. 31 is a reproduced image calculation unit that calculates a reproduced image of the current frame. 33 is a coding method selection signal, a prediction error, and a quantization step size as a channel code. prediction error encoding
35 is a motion vector encoding unit that encodes a motion vector through a channel; 37 is a multiplexer B that configures a transmission frame from a prediction code and a motion vector code; and 40 is an output terminal that outputs a transmission signal; , the operation is explained below.
The television signal is converted from analog to digital by a signal processing unit not shown in Figure 1.
It is divided into blocks of N lines in the vertical direction and inputted as input television signal 2 from input terminal 1.Next, motion vector calculation section 3 receives input television signal 2 and the reproduced image of the previous frame read from image memory section 4. 5, calculate the motion vector, and output it as the 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 as motion compensation prediction control information to the motion vector signal 6. Motion compensation prediction unit 7 top If motion compensation prediction is to be performed using the motion vector signal 6 for the reproduced image 5 of the previous frame at the same position as the encoded block, motion compensation prediction is performed using the motion vector, and if no motion compensation prediction is to be performed, nothing is done. The inter-frame/frame white determination unit 9 outputs the motion-compensated prediction signal 8 as the motion-compensated prediction signal 8. It compares the input television signal 2 and the motion-compensated prediction signal 8 block by block, determines the effectiveness of the motion-compensated prediction, and If the effectiveness of compensated prediction is small, it is determined that intraframe coding is effective for the corresponding block, and if the effectiveness of motion compensation prediction is large, it is determined that interframe coding is effective for the corresponding block, and the result is encoded. Method selection signal 1
Output as o and intra-frame/inter-frame determination unit 9
In the case of an encoding device that requires the insertion of a refresh frame, the encoding method selection signal 1o is outputted so that all blocks are intra-frame encoded at a constant frame period. If the encoded block is a block to be predicted with motion compensation, in-loop filter processing, which is two-dimensional low-pass filter processing, is performed on the motion-compensated prediction signal 8. In other cases, the in-loop filter process is not performed, and the prediction signal 12 is The prediction error calculation unit 13 calculates the difference between the input television signal 2 of the encoded block and the prediction signal 12. The switch unit 15 outputs the result as the prediction error signal 14. (1) Encoding method If the selection signal 10 selects intraframe coding, select the input television signal 2 as the signal 16 to be orthogonally transformed, and (2
) If the encoding method selection signal 10 selects interframe coding, the prediction error signal 14 is selected as the signal 16 to be orthogonally transformed. No line＼ Calculate the orthogonal transform coefficient 18 by removing the high correlation between neighboring pixels of the signal 16 to be orthogonally transformed. In most cases, the orthogonal transform method has high conversion efficiency. Realization of hardware implementation Even if a discrete cosine transform is used, the orthogonal transform coefficient transform unit 19 compares the coefficient value 1coefl of the orthogonal transform coefficient 18 with the quantization step size 21 calculated by the quantization step size calculation unit 20, and determines the difference between the two in advance. If the value is not larger than the threshold Th, then perform numerical transformation using orthogonal transformation system numerical value 1coef1 as quantization step size Qb (\Numerically transformed orthogonal transformation coefficient 22
In other words, IF 1coef l (Qb) THENIF
(Qb-lcoef l (Th) THENlcoe
fl=Qb END I F NDIF Summata Threshold T h 41 Using a typical image, the quantization unit 23 uses the quantization step size 21 to calculate
The quantization unit 23 quantizes the numerically converted orthogonal transform coefficient 22 and calculates the orthogonal transform quantized coefficient 24. If the orthogonal transform coefficient and the quantization step size are equal, the orthogonal transform quantized coefficient 24 is set to "1" or "-1'', the quantization step size 2tc;
6, it is obtained by the method described in the conventional example above. Inverse orthogonal transform section 27, top. Muswitch section that performs inverse orthogonal transform on the orthogonal transform quantization coefficients 24 and calculates the signal 28 containing the quantization error. 15: (1) When the encoding method selection signal 10 selects intra-frame encoding, the numerical value rQJ signal 3° is selected as the reproduced image calculation signal 29, and (2) when the encoding method selection signal 10 selects intra-frame encoding, If inter-coding is selected, the predicted signal 12 is selected as the reproduced image calculation signal 29.The reproduced image calculation unit 31 adds the signal 28 containing the quantization error and the reproduced image calculation signal 29, and performs encoding. The image memory 4 that calculates the reproduced image 32 of the block stores the reproduced image signal 32 of the current frame, and the prediction error encoding unit 33 that outputs the reproduced image signal 5 of the previous frame receives the encoding method selection signal 10, the quantum The motion vector encoding unit 35 encodes the quantization step size 21 and the orthogonal transform quantization coefficient 24 and calculates the prediction error code 34. The motion vector encoding unit 35 encodes the motion vector signal 6 of the block subjected to motion compensation prediction and calculates the motion vector code 36. The sum multiplexer unit 37 calculates a transmission frame 38 in a predetermined format from the prediction error code 34 and the motion vector code 36.The sum code memory unit 25 stores the transmission frame 38 once.
Synchronized with a clock signal input from an external device (not shown), the output signal is output from the output terminal 40 as a transmission code 39.
At the same time, the code memory unit 25 calculates the amount of codes remaining in the memory as the amount of residual codes 26.As is clear from the above explanation, according to this embodiment, when quantizing orthogonal transform coefficients, the quantization step size is Since the orthogonal transform coefficient power that was quantized to a value slightly smaller than the numerical value [○ Before quantizing the orthogonal transform coefficients, each orthogonal transform coefficient value lc for the orthogonal transform coefficients that become rOJ after quantization.

efl and quantization step size Qb are compared, and for orthogonal transform coefficients where the difference between the two is less than or equal to a predetermined threshold Th,
By performing numerical conversion using the orthogonal transform coefficient value ICoef1 as the quantization step size, the image quality of the reproduced image is improved by preventing the orthogonal transform number having a certain size or more from becoming "0" after quantization. The effect is strong.

[Brief explanation of the drawing]

FIG. 1 is a block wiring diagram of a motion compensated predictive interframe coding device according to an embodiment of the present invention. FIG. 2 is a block wiring diagram of a conventional motion compensated predictive interframe coding device. ... Input terminal, 3, 53... Motion vector calculation ability 4, 54... Image memory capacity 7.58... Motion compensation prediction ability 9, 60... Intra-frame/inter-frame determination ability 11, 62 ...loop filter 13.6
4...Prediction error calculation @15.66...Switch test 1
7.68... Orthogonal transform ability 19... Orthogonal transform coefficient conversion ability 23.70... Quantization ability 20.73... Quantization step size calculation ability 25.74... Code memo 1 person
27.76... Inverse orthogonal transform performance 31.78... Reproduction image calculation ratio 33, 82... Prediction error coding paper 35.84
...Motion vector encoding capacity 37.86...Multiplexer 40, 89...Output terminal.

Claims (1)

[Claims] Analog-to-digital conversion means for converting a television signal from an analog signal to a digital signal, and blocking for dividing one frame or one field of a digitized input television signal into blocks of a predetermined size. a motion vector calculation means for calculating a motion vector representing a motion of a television image for each block; and a motion compensation determination means for determining whether to perform motion compensation prediction using the motion vector for each block. , motion compensation prediction pixel calculation means for performing motion compensation prediction on the reproduced images of all frames using motion vectors for blocks to be motion compensated predicted and calculating predicted pixel values, and predicting the difference between the pixel value of the input television signal and the predicted pixel value. A prediction error calculation means for calculating an error value; a coding method determination means for determining whether to perform interframe encoding or intraframe encoding for each block; orthogonal transformation signal selection means for switching a signal to be orthogonally transformed as a pixel value of an input television signal or a prediction error value according to a determination result; An orthogonal transform means that calculates transform coefficients, a quantization step size calculation means that calculates a quantization step size from the generated code amount, and compares each orthogonal coefficient with the quantization step size to determine whether orthogonal transform system values are predetermined. For orthogonal transform coefficients that become zero after quantization within the range, an orthogonal transform coefficient conversion means that numerically converts the corresponding orthogonal transform value to a value equal to the quantization step size, and all orthogonal transform coefficients including the numerically converted orthogonal transform coefficients. Orthogonal transform coefficient quantization means for quantizing orthogonal transform coefficients using a quantization step size, information on intra-frame encoding or inter-frame encoding, a motion vector value, a quantization step size, and a quantized orthogonal transform An encoding means for encoding coefficients, an inverse orthogonal transformation means for performing inverse orthogonal transformation on the quantized orthogonal transformation coefficients and calculating an inverse quantized signal, and reproduction based on the interframe encoding/intraframe encoding determination result. Reproduction pixel selection means for switching whether a pixel value to be used when calculating a pixel value is a motion-compensated predicted pixel value or a numerical value "0"; A motion compensated predictive interframe coding device comprising a reproduced image calculating means for calculating a reproduced image, and a reproduced pixel accumulating means for accumulating the reproduced image.