JPH0490218A - Coder and adaptive decoder - Google Patents

Abstract

PURPOSE:To select a predictive system suitable for a video signal in a short time unit by deciding which predictive system is selected based on a data resulting from coding a predictive error data and outputting a relevant coding data by each stored prediction system as the coded data. CONSTITUTION:An adaptive changeover switch 10 encodes a predictive error data being an input data, stores the coded data by one field, calculates a parameter deciding which data of an input data is to be selected, selects an optimum predictive system according to the parameter and outputs a coded data (c) corresponding to the predictive system. An adaptive decoder 11 decodes the coded data (c) into a predictive error data and uses a data by three preceding field data to be stored to decode the field data. Then the field data 1 of one preceding field corresponding to the coded data (c) and a frame data (m) of one preceding frame time composed of the data of two-preceding field of the field data 1 are outputted.

Description

[Detailed Description of the Invention] <Industrial Application Field> The present invention relates to an encoding device for encoding a video signal with high efficiency, which is used in an image transmission device, an image recording/reproducing device, etc. .

<Conventional technology> As a method for realizing highly efficient encoding of video signals,
There is predictive encoding. Predictive encoding predicts a current pixel value from past pixel values, and quantizes and encodes the difference value (prediction error) between the predicted pixel value and the actual pixel value. Since the distribution of prediction errors concentrates around zero, highly efficient encoding can be achieved by using its statistical properties.

Predictive coding includes intraframe predictive coding that performs prediction using adjacent pixel values within the same frame, and interframe predictive coding that performs prediction using pixel values of a plurality of frames before and after. Intraframe prediction removes spatial redundancy within the same frame, and interframe prediction removes temporal redundancy across multiple frames. In video images, inter-frame prediction is generally more efficient than intra-frame prediction because the correlation between frames is large in static areas with little movement, but it is less efficient in dynamic areas with rapid movement. Intraframe prediction has better prediction efficiency than interframe prediction.

From this, we examine the relationship between frames, determine whether the frame is a frame in a dynamic domain or a frame in a static domain, and use intra-frame prediction for frames in a dynamic domain and A method using inter-frame prediction has been proposed for frames in the region (Television Society of Japan, Technical Report, Vol. 13, No. 26. method"
).

<Problems to be Solved by the Invention> In the above-mentioned method that adaptively uses intra-frame prediction and inter-frame prediction, a parameter is used to determine whether the area is a dynamic area or a static area based on the pixel values of the previous frame and the current frame. After calculating, if the area is determined to be dynamic, intra-frame prediction is performed, and if it is determined to be a static area, inter-frame prediction is performed. In other words, there is a problem in that a time delay occurs because prediction for that frame is performed after data input for one frame has been completed. In addition, because the size relationship of the parameters used to determine whether the area is dynamic or static does not necessarily match the size relationship of the amount of code when actually encoded, some areas may be determined to be dynamic areas. There was also the problem that better results were obtained using interframe prediction for frames as well.

The present invention was made in view of the above-mentioned problems, and is capable of adaptively selecting an optimal prediction method from among intra-field prediction, inter-field prediction, intra-frame prediction, and inter-frame prediction. It is an object of the present invention to provide an encoding device that does not cause a time delay when encoding a prediction error and can switch to each prediction method for each arbitrary field.

<Means for Solving the Problems> In order to achieve the above object, the encoding device according to the present invention generates current frame data from the current field data and the field data l field time ago every field time. an intra-field predictor that performs intra-field prediction from the current field data and outputs an intra-field prediction error; and an inter-field predictor that performs intra-frame prediction from the current frame data output from the framing circuit and outputs an intra-frame prediction error of a portion corresponding to the current field data. An intra-frame predictor performs inter-frame prediction from frame data one frame time before and the current frame data output from the framing circuit, and outputs an inter-frame prediction error of a portion corresponding to current field data. an inter-frame predictor, each encoding the intra-field prediction error, the inter-field prediction error, the intra-frame prediction error, and the inter-frame prediction error and storing them as encoded data;
an adaptive switcher that appropriately switches and outputs the optimal encoded data for each field from among the four types of stored encoded data based on the encoded data; and encoded data output from the adaptive switcher. The field data corresponding to the encoded data is restored from the encoded data and outputted as the field data of one field time ago, and the field data is stored, and based on the already stored two field data of two field time ago. The present invention is characterized by comprising an adaptive decoder that configures frame data of the one frame time before.

Further, the adaptive decoder includes a decoder that decodes encoded data in which a prediction error is encoded into prediction error data, and a decoder that decodes encoded data in which a prediction error is encoded, and a decoder that decodes the prediction error data, field data two field times before, and l which is delayed by one field time. an adaptive restorer for restoring field data from one field time before from frame data from a frame time before, and a field delay circuit for delaying the field data from one field time before and outputting field data from two field times before. It is preferable to have a frame forming means for outputting frame data from one frame time before and frame data from one frame time before, which is delayed by the l field time, from the field data from two field times before. .

Furthermore, the framing means includes a field delay circuit that delays the field data of the two field times before and outputs the field data of the three field times before;
a framing circuit that outputs frame data of one frame time before from field data of a field time before;
It is preferable that the image forming apparatus further includes a second field delay circuit that delays the frame data of the previous frame time and outputs the frame data of the previous frame time delayed by the one field time. The framing means also includes a field delay circuit that delays the field data of the two-field time song and outputs the field data three field times before and the field data four field times before, and the two field time delay circuit. a first framing circuit that outputs frame data from one frame time ago from the field data from the field data and the field data from the three field time periods; 1 delayed by one field time from
It is preferable that the image forming apparatus further includes a second framing circuit that outputs frame data of a frame time.

<Operation> In the present invention configured as described above, the intra-frame predictor, inter-frame predictor, intra-field predictor, and inter-field predictor perform intra-frame prediction, frame prediction before deciding which prediction method to select. Inter-field prediction, intra-field prediction, and inter-field prediction are performed in parallel, and prediction error data is also encoded at the same time. When field data input is completed, the prediction error data and prediction error of the field are encoded. Create and store converted data. After that, the prediction error data is encoded, it is determined which prediction method to select based on the f-co data, and the data corresponding to the selected method is selected from among the already stored encoded data of each prediction method. is output as encoded data.

<Example> The details of the invention will be described below.

First, the frame configuration method adopted in the present invention will be explained.

When a prediction method is selected adaptively by adding intra-field prediction and inter-field prediction to intra-frame prediction and inter-frame prediction, two types of methods shown in FIG. 2 can be considered.

In general, intra-frame prediction and inter-frame prediction are performed in units of one frame, and intra-field prediction and inter-field prediction are performed in units of one field. Therefore, from intra-field prediction and inter-field prediction to intra-frame prediction, When adaptively switching to interframe prediction, a method of switching at the end field of the corresponding frame can be considered.

Figure 2 (a) shows this situation. FIG. 2(a) shows that frame 1 is composed of field la and field lb. Since the end point of field 1b is also the end point of frame 1, the prediction is switched between intra-frame prediction and inter-frame prediction at this point. That is, intra-frame prediction and inter-frame prediction are performed starting from frame 2, and at the end of field lai%, there is no switching to intra-frame prediction and inter-frame prediction in the middle of frame 1.

However, in this method, the timing of switching the prediction method is limited to each frame time, which limits the effectiveness of optimization by switching the prediction method.

Therefore, in the present invention, a method is adopted in which one frame is always constructed from the acid field and the current field. This situation is shown in Figure 2 (b). Figure 2 (b) shows the field l
This indicates that when inputting b, frame IB is created from field la and field 1b, and when field 2a is inputted, frame 2A is created from field 1b and field 2a. That is, every field time (
Frame data is created every time it takes to input one minute of data. With this method, the prediction method can be switched for each field time, so the optimization effect can be fully demonstrated.

FIG. 3 is an example of an apparatus for encoding and decoding interlaced analog signals using the present invention. Figure 3 (
(a) is a video signal encoding section, and (b) in the same figure is a decoding section. In the figure, reference numeral 1 is an A/D converter, and 2 is an adaptive predictive encoder according to the present invention. Further, 3 is a decoder, and 4 is a D/A converter. The interlaced analog signal a is converted into digital data b by an A/D converter l.

The resolution of the digital data may be 16 bits or 8 hits. In any case, the description of the invention is not affected. The adaptive predictive encoder 2 first performs prediction on digital data b using four methods: intra-field prediction, inter-field prediction, intra-frame prediction, and inter-frame prediction, and obtains a prediction error that is a difference from the original data. Next, the prediction error is encoded, the encoded data is compared, and the optimal prediction method is selected. Then, encoded data C based on the selected prediction method is output. On the other hand, in the video signal decoding section, the decoder 3 decodes the encoded data d into prediction error data, restores interlaced data e from the prediction error data, and applies it to the D/A converter 4. D/A
The converter 4 converts the input signal into the original analog signal,
A reproduced signal f is output.

FIG. 1 shows an example of the adaptive predictive encoder 2. b in the same figure
and C correspond to digital data b and encoded data C in FIG. 3, respectively. In the same figure, numeral 5 is a framing circuit, 6 is an intra-field predictor, 7 is an inter-field predictor, 8 is an intra-frame predictor, 9 is an inter-frame predictor, IO is an adaptive switch, and it is an adaptive decoder. It is.

Digital data b (current field data) is applied to a framing circuit 5, an intrafield predictor 6, and an interfield predictor 7. In the framing circuit 5, current frame data g is constructed from current field data b and field data l of a field one field time before, and is applied to an intraframe predictor 8 and an interframe predictor 9. Therefore, in the framing circuit 5, one frame of data is constructed from the current field data and the field data one field time ago every l field times. In order to prevent the accumulation of coding distortion, the data l of the previous field is not original data, but data obtained by decoding coded data by the adaptive decoder 11.

In the intra-field predictor 6, the current field data b
The inter-field predictor 7 predicts the current field data b from the field data l delayed by l field time and outputs the inter-field prediction error i. Uchisensokuki 8 predicts the entire frame from the current frame data g, and then predicts the current field data b.
The interframe predictor 9 predicts the entire frame from the frame data m delayed by l frame time, and outputs the part corresponding to the current field data b of the current frame data g as the intraframe prediction error j. Output as the prediction error between. Prediction error data h, i, and jk are each applied to an adaptive switch lO.

Here, as the data m of the previous frame, data obtained by decoding encoded data by the adaptive decoder 11 is used.

The adaptive switch IO encodes the prediction error data that is input data, accumulates one field of encoded data, and calculates parameters to determine which data of the input data to select.According to the parameters, The optimum prediction method is selected and encoded data C corresponding to the selected prediction method is output. Here, the encoded data C is
Since one field's worth of data is output after input is completed,
l field time delay has occurred. As the parameter for selecting the above-mentioned data, any of the variance and entropy of prediction errors, the amount of code when actually encoded, the error between decoded data and original data, etc. may be used.

For this purpose, an appropriate one that matches the actual encoding method may be adopted.

The adaptive decoder 11 decodes the encoded data C into prediction error data and restores the field data using the stored data for the previous three fields. Then, field data 1 of 1 field time period corresponding to encoded data C and frame data m of 1 frame time ago, which is composed of data of the previous two fields of field data 1, are output.

Figure 4 shows each signal in Figure 1 (b, l, g, m, hi, j,
shows the time chart of c). In the figure, the time on the horizontal axis is expressed in units of time for one field. la,
1b represents the field number, 1B and 2A represent the frame number, frame 1B represents field Ia and field Ib1 frame 2A represents field ib and field 2a.
It consists of Taking the case of time t=4 in FIG. 4 as an example, the data of field 2b is input to b, and the data of field 2a delayed by one field time is input to 1. field 2a in g
The data of frame 2B consisting of field 2b and field 2b are input, and the previous frame data m contains frame I
Data of B has been input. Prediction error data h, i,
Field 2b is added to j and k from the respective input data.
The prediction error from the original data is output. The encoded data C is delayed by one field time in the adaptive switch IO, and data 2a' which encodes the prediction error of field 2a is output.

FIG. 5 shows an example of a framing circuit. In the figure, numerals 39 and 42 are buffer controllers, and 40 and 41 are FI
FO buffer 43 is a switch. Signal names 11g in the figure correspond to the signal names in FIG. Z is a control signal included in field data b and is not shown in FIG.

The current field data b is stored in the FIFO buffer 40.
Field data I from one field time ago is added to the FIFO buffer 41. FIFO buffers 40 and 4
1 can store field data for l lines. Field data within the FIFO buffer is controlled by a buffer controller. Buffer controller 39
and 42 and switch 43, the above-mentioned control signal 2 is applied, respectively. In the control signal Z included in field data b, field data b is the first signal of the frame.
There is a control signal 1 indicating that the line is a line, and a control signal 2 indicating that it is the second line of the frame. Note that the time required to input one line of field data is defined as one unit.

In the buffer controller 39, when the control signal l is applied,
One line of field data is read from the FIFO buffer 40 in 0.5 units of time, and no data is read out for the following 0.5 units of time. This operation is repeated until control signal 2 is applied. When control signal 2 is applied, field data is not read for the first 0.5 units of time, and FIFO is read for the next <0.5 units of time.
One line of field data is read from the buffer 40. This operation is repeated until the next control signal l is applied. In the buffer controller 42, the buffer controller 3
Do the opposite of step 9. A control signal l is sent to the buffer controller 42.
is added, field data is not read for the first 0B unit time, and FI is read for the next 0.5 unit time.
One line of field data is read from the FO buffer 41. This operation is repeated until control signal 2 is applied. When control signal 2 is applied, one line of field data is read from the FIFO buffer 41 in 0.5 units of time, and no field data is read out in the following 0.5 units of time. This operation is repeated until the next control signal 1 is applied. As described above,
One line of field data is alternately output from the FIFO buffers 40 and 41 every 05 units of time and applied to the switch 43. The switch 43 is P.I.F.
The current frame data g is constructed by alternately switching and outputting the data output from the O-buffers 40 and 41. That is, after a control signal of 1 is applied, 0.
The field data from the FIFO buffer 40 is output for 5 units of time, and the field data from the FIFO buffer 41 is output for the following 0.5 unit time. This operation is repeated until control signal 2 is applied. When control signal 2 is applied, the first 05 units of time are FIFO
Output field data from buffer 41, followed by 0
．． Data from the FIFO buffer 40 is output for five units of time. This operation is repeated until the next control signal l is applied.

An example of previous value prediction is shown as an example of the intra-field predictor 6 in FIG. 6, the inter-field predictor 7 in FIG. 7, the intra-frame predictor 8 in FIG. 8, and the inter-frame predictor 9 in FIG. show.

In each figure, 44, 47, 49, 54 are predictors, 45 and 50 are decoders, 46 and 51 are encoders, 48 and 53 are FIFO buffers, and 52 and 55 are buffer controllers. represents. Signal names in each figure: h, 1. ig, j, and m correspond to the signal names in FIG.

Current field data b is input to the intra-field predictor in FIG. 6, and the difference between b and the predicted value b, that is, the difference value from the previous value, is output as a prediction error. In order to prevent the accumulation of encoding distortion, b is a value obtained when data one pixel before is encoded and decoded. h is also input to the encoder 46, and the encoded data is added to the decoder 45. The sum of the data decoded by the decoder 45 and the decoded data of the previous pixel, that is, the decoded value of the current pixel, is added to the predictor 44. The predictor 44 uses these decoded values to predict the current pixel and outputs the predicted value. Here, a delay circuit of one pixel time (time required to read field data of one pixel) is used as a predictor. Therefore, the predictor 44 outputs 1 for b.
The pixel-delayed decoded data is output as a predicted value.

The current field data b and field data l delayed by one field time are input to the interfield predictor shown in FIG. 1 is encoded and decoded data. (2) is added to the predictor 47, and the predicted value of b is output. Here, the predictor 47 outputs the input data 1 as is.

That is, the decoded value of data at the same position l fields before b is output as the predicted value. The difference between b and b is output as the prediction error l.

Frame data g is input to the intra-frame predictor in FIG. 8, and the difference between g and the predicted value g, that is, the difference value from the previous value, is output as a prediction error α.

In order to prevent the accumulation of coding distortion, t uses a value obtained when data one pixel before is coded and decoded.

α is applied to FIFO buffer 48 and encoder 51. Data encoded by encoder 51 is applied to decoder 50. The sum of the data decoded by the decoder 50 and the decoded data of the previous pixel, that is, the decoded value of the current pixel, is input to the predictor 49. Here, as a predictor, a delay circuit for the time required to read one pixel of frame data is used. Therefore, the predictor 49 outputs decoded data delayed by one pixel with respect to g as the predicted value g. Since one frame of frame data g is input in one field time, a prediction error corresponding to one frame data can also be obtained in one field time. Data input to the FIFO buffer 48 is controlled by a buffer controller 52. The buffer controller 52 controls the prediction error data α input to the FIFO buffer 48 so that it is only the prediction error of the current field data. The FIFO buffer 48 takes out one piece of data every pixel time in order from the data that was input first, and outputs it as prediction error data j.

The interframe predictor shown in FIG. 9 receives frame data m delayed by one field time from the current frame data g. m is encoded and decoded data. m is applied to a predictor 54, and a predicted value g of g is output. Here, the predictor 54 outputs the input data m as is. That is, the decoded value of data at the same position one frame before g is output as the predicted value ↑. The difference between g and g is output as prediction error β, and β is FI
It is added to the FO buffer 53. FIFO buffer 53
The data input to is controlled by a buffer controller 55. The buffer controller 55 controls so that the prediction error data input to the FIFO buffer is only the prediction error of the current field data. The FIFO buffer 53 outputs one piece of data as prediction error data every pixel time in order from the data that was input first.

FIG. 1O shows an example of the adaptive switch IO of FIG.

Here, the prediction method that minimizes the amount of code is selected.

In the figure, the signal names are i, j, k.

C corresponds to the signal name in FIG. Sign 12-15
is the encoder, 16 to 19 are the field buffers, 2
0 to 23 are code amount counters, 24 is a minimum value detector,
25 represents a switching circuit. Encoders 12 to 15 encode the input prediction error data, and send the encoded data to field buffers 16 to 19 and field buffers 20 to 20, respectively.
~23 is added to the code amount counter. Field buffers 16 to 19 store encoded data for one field. The code amount counters 20 to 23 count the total code amount for each field, and the results are added to the 24 minimum value detectors. The minimum value detector outputs the data number r corresponding to the minimum value among the four input values,
25 switching circuits. The data number r corresponds to the field buffer number. The switching circuit No. 25 selects one of the four input signals according to the contents of r.
Select one and set it as output C.

FIG. 11 shows an example of the adaptive decoder 11 of FIG. 1. Signal names c, 1. m corresponds to the signal name in FIG. 26 is a decoder, 27 is an adaptive restorer, 28
．． 29 and 31 are field delay circuits, and 30 is a framing circuit. The decoder 26 decodes the prediction error encoded data C into prediction error data and applies it to the adaptive restorer 27.

The encoded data C has a delay of one field time from the original data. The adaptive restorer 27 receives the prediction error data, the restored data S of the previous field used when creating the prediction error, and the restored data U of the previous frame, and uses this data to restore the predictive error data to field data. Output the time-delayed field data l. Field delay circuits 28, 29, and 31 each delay input data by one field time. The framing circuit 30 contains field data S delayed by 2 field times and field data t delayed by 3 field times from the original data.
is input, and frame data m of the previous frame is configured.

Figure 12 shows the current field data b in Figure 1 and field data 11.
Signal c in the figure, 1. A time chart of s, t, m, and u is shown. The description method is the same as that in FIG. Taking the case of time t = 4 in FIG. 12 as an example, the data in field 2b is input to the original data b, and the encoded data C is 1
Data 2a that encodes the prediction error of field 2a delayed by field time is output. The encoded data C is decoded and the restored field data 2 is stored in l.
a is output. Field data 1b obtained by delaying 1 by one field time is output to S, and field data la obtained by delaying S by one field time is output to t. Frame data IB composed of field data 1b of S and field data 1a of t is stored in m, and frame data IA is created by delaying m by one field time in U.
is being output.

Another example of the adaptive decoder 11 is shown in FIG.

Signal names c, l, and m in the figure correspond to the signal names in FIG. 1. In the figure, 32 is a decoder, 33 is an adaptive restorer, 34, 35, 36 is a field delay circuit, 37, 38
is a framing circuit. The decoder 32 decodes the prediction error encoded data C into prediction error data and applies it to the adaptive restorer 33. The encoded data C has 1 more than the original data.
Is there a field time delay? The adaptive restorer 33 receives prediction error data, restored data of the previous field v1 used when creating the prediction error, restored data y of the previous frame, and uses this data to restore the predicted error data to field data, and restores the prediction error data to field data. Delayed field data 1
to come out. Field delay circuits 34, 35, and 36 each delay input data by one field time.

Field data V delayed by 2 field times 17 and field data W delayed by 3 field times from the original data are input to the framing circuit 37, thereby forming frame data m of the previous frame. The frame circuit 38 contains field data W delayed by three field times from the original data.
Field data X delayed by four field times is input to form frame data y.

FIG. 14 shows the current field data b and the signals c in FIG. 13, 1. A time chart of v1w, m, x, and y is shown. Taking the case of time t=4 in the figure as an example, original data b is inputted with the data of field 2b, and encoded data C is delayed by l field time in field 2.
Data 2a that encodes the prediction error of a is output. 1, encoded data C is decoded and restored field data 2a is output. ■ is field data 1b, which is 1 delayed by one field time, and W is V.
Field data Ia which is delayed by one field time is output. m has field data of V lb and W
Frame data IB made up of field data la of , field data Ob obtained by delaying W by one field time are placed in X, and frame data IA made up of field data la of W and field data Ob of X are placed in y. is being output.

The above explanation has been given using the adaptive predictive encoder used in the video signal encoding section in FIG. 3(a) as an example, but the decoder 3 used in the decoding section in FIG. It can be realized with the same configuration as the adaptive decoder shown in FIG. 11th
When the configuration shown in the figure is used as the decoder 3, the signal l in FIG. 11 may be used as the signal e in FIG. 3 (b). Similarly, the decoder 3 can also be realized with the configuration shown in FIG. 13, in which case the signal l in FIG. 13 is used as the signal e in FIG. 3(b).

In the above explanation, encoded data in which a prediction error is encoded is used as data for selecting the optimal prediction method, but it is of course possible to use the prediction error data itself.

<Effects of the Invention> As described above, according to the present invention, an optimal prediction method can be adaptively selected from among intra-field prediction, inter-field prediction, intra-frame prediction, and inter-frame prediction. Furthermore, since frame data is created for each field time, it is possible to switch from intra-field prediction and inter-field prediction to intra-frame prediction and inter-frame prediction when data for all fields is completed. That is, since switching to any prediction method is possible for each arbitrary field, it is possible to select a prediction method suitable for the video signal in short time units.

Furthermore, in the present invention, prediction error data and data in which the prediction error is encoded are created and stored when field data is completed, and prediction error data or encoded data based on the optimal prediction method is selected from the stored data. Since the converted data is selected and output, there is no time delay.

In addition, since the prediction error is actually calculated and the optimal one is selected from the prediction error data that is the f-co result or the encoded data that encoded the prediction error data, the optimal one that matches the actual amount of code and encoding distortion is selected. A prediction method can be selected adaptively.

As described above, according to the encoding device of the present invention, it is possible to adaptively select an optimal prediction method from among intra-field prediction, inter-field prediction, intra-frame prediction, and inter-frame prediction, and It is possible to provide an encoding device that does not cause a time delay when encoding errors and can switch to each prediction method for each arbitrary field.

FIG. 1 is a block diagram representing an embodiment of an adaptive predictive encoder;
Figure 2 is a diagram showing the relationship between fields and frames;
The figure is a block diagram of a video signal encoding section and decoding section using the present invention, FIG. 4 is a time chart of main signals, and FIG.
6 is a block diagram of the intra-field predictor, FIG. 7 is a block diagram of the inter-field predictor, and FIG. 8 is a block diagram of the intra-frame predictor.
Figure 9 is a block diagram of the interframe predictor, Figure 1O is a block diagram of the adaptive switch, Figure 11 is a block diagram of the adaptive decoder, Figure 12 is a time chart of the main signals, and Figure 13 is a block diagram of the adaptive switch. Block diagram of another embodiment of the adaptive decoder, first
Figure 4 is a time chart of the main signals.

l... A/D converter, 2... adaptive predictive encoder, 3
... Decoder, 4... D/A converter, 5... Framing circuit, 6... Intra-field predictor, 7... Inter-field predictor, 8... Intra-frame predictor, 9... Interframe predictor, lO... Adaptive switch, 11...
・Adaptive decoder, 26...Decoder, 27...Adaptive restorer, 28.29...Field delay circuit, 30...
- Framing circuit, 31... field delay circuit.

Agent: Patent attorney Masaru Umeda (2 others) Figure 3 4 appetizers (- Figure 4 Figure 5 Figure 6 Figure 7 Figure 08 Paddy? Figure 9 Time closed (74-le F' wound 0 Figure 12

Claims (1)

[Claims] 1. Current field data and 1 every field time
a framing circuit that configures current frame data from field data from the previous field time; an intra-field predictor that performs intra-field prediction from the current field data and outputs an intra-field prediction error; an inter-field predictor that performs inter-field prediction from field data and the current field data and outputs an inter-field prediction error; and an inter-field predictor that performs intra-frame prediction from the current frame data output from the framing circuit to predict the current field. an intra-frame predictor that outputs an intra-frame prediction error of a portion corresponding to the data; and an intra-frame predictor that outputs an intra-frame prediction error of a portion corresponding to the data; and an intra-frame predictor that performs inter-frame prediction from the frame data of one frame time ago and the current frame data output from the framing circuit to generate the current field. an interframe predictor that outputs an interframe prediction error of a portion corresponding to data; and encoded data by encoding the intrafield prediction error, the interfield prediction error, the intraframe prediction error, and the interframe prediction error, respectively. and an adaptive switch that appropriately switches and outputs the optimal encoded data for each field from among the four types of stored encoded data based on the encoded data; The field data corresponding to the encoded data is restored from the encoded data, and is output as the field data of one field time ago, and the field data is stored, and the two fields of the already stored two field time ago are output. and an adaptive decoder that configures the frame data of the previous frame based on the data. 2. A decoder that decodes encoded data in which a prediction error is encoded into prediction error data; and the prediction error data, field data two field times earlier, and frame data one frame time earlier delayed by one field time. an adaptive restorer for restoring field data from one field time before, a field delay circuit for delaying field data from one field time before and outputting field data from two field times before; 1. An adaptive decoder comprising: a frame forming means for outputting frame data of one frame time before field data and frame data of one frame time before the field data delayed by one field time. 3. A field delay circuit in which the framing means delays the field data from two field times before and outputs the field data from three field times before, and the field data from two field times before and the field from three field times before. a framing circuit that outputs frame data of one frame time before the data; and a second frame circuit that delays the frame data of one frame time before and outputs frame data of one frame time before delayed by the one field time. 3. The adaptive decoder according to claim 2, further comprising a field delay circuit. 4. A field delay circuit in which the framing means delays the field data from two field times before and outputs the field data from three field times before and the field data from four field times before; a first framing circuit that outputs frame data from one frame time ago from the field data and the field data from three field times before; and from the field data from the three field times before and the field data from four field times before. 3. The adaptive decoder according to claim 2, further comprising a second framing circuit that outputs the frame data of one frame time before the one field time delay.