JPS63146677A - Predictive coding device for video signal - Google Patents

Abstract

PURPOSE:To attain the use of a low speed circuit element by dividing a picture pattern into plural regions, storing a video signal into a buffer memory corresponding to each region and applying predictive coding to the signal stored in each region independently in parallel thereby giving a timewise margin to the loop processing speed of a predictive coder. CONSTITUTION:A video signal entering an input terminal 1 is digitized at a sampling speed by an A/D converter 2. When the sampled picture element exists in a region A, a digital video signal is written in an input side buffer memory 10A, in case of the region C, the signal is written in input buffer memories 10A, 10B, and a boundary region buffer memory 13, in case of the region B, the signal is written in the input buffer memory 10B. The digital video signal stored in the memory 10A, 10B is subject to predictive coding respective by predictive coder 12A, 12B. The coded signal subject to predictive coding by the predictive encoders 12A, 12B is written in output buffer memories 11A, 11B. The data written in the output buffer memories 11A, 11B and the boundary region buffer memory 13 is read sequentially and outputted from an output terminal 6.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a predictive coding device for video signals.

2. Description of the Related Art In recent years, with advances in digital technology, video signals have come to be digitized for processing and transmission.

Hereinafter, digitization of an NTSC video signal will be explained.

Since the video signal band of the NTSC system is approximately 4 segments, taking Nyquist's theorem into consideration, it is necessary to sample at a rate of 8 or more. This sampled video signal is subjected to analog-to-digital conversion (A-D conversion).

Assuming that the sampling frequency is 10M and the quantization per sample is 8 bits, the amount of information of the digitized video signal is 80 Mbit/s. In this way, when transmitting a digitized signal, approximately 2 times the analog video signal
It requires 0 times the bandwidth.

Therefore, when transmitting digitized video signals, it is important to reduce the amount of information and compress the transmission band.

A DPCM device that predictively encodes a signal is well known as a device that encodes a digitized video signal with high efficiency. Hereinafter, conventional techniques will be explained using examples.

FIG. 3 shows a block diagram of a conventional DPCM device, in which 1 is a video signal input terminal, 2 is an analog-to-digital converter (hereinafter referred to as A, 3 is a subtracter that performs subtraction between the digitized video signal 5 is an encoder that inputs the quantized signal Y and converts it into a transmission line code, 6 is an output terminal that outputs the encoded signal from this device, and 7 is a quantizer that outputs a quantized signal Y. 8 is an adder that adds the encoded signal Y and the predicted signal X and outputs the locally decoded signal 2; and 8 is a predictor that outputs the predicted signal X using the locally decoded signal 2.

Next, the operation will be explained.

A video signal input to an input terminal 1 is sampled at a sampling frequency by an A/D converter 2 to become a digital signal X. The subtracter 3 subtracts the prediction signal X from the digital video signal X to obtain a prediction error signal E. The prediction error signal E is quantized by a quantizer 4 to produce a quantized signal Y.
become. The quantized signal Y is converted into a transmission line code by the encoder 6 and outputted from the output terminal 6. Further, the quantized signal Y and the predicted signal X are added by an adder 7 to obtain a locally decoded signal Z, and a predictor 8 generates a predicted signal X using the locally decoded signal Z. The prediction signal X newly obtained here repeats the above loop operation with the digital video signal X obtained at the next sampling time.

FIG. 4 shows an example of the positional relationship of pixels and a prediction function used for prediction. I is the pixel to be predicted. Pixel an (n=1, 2.3) is a signal from the same line as I, one sampling ago, and pixel bn (n=-1, 0,
1, 2, 3) are the signals of the previous line.

A prediction function using pixels around X can be expressed as shown in the figure, and the predictor 8 obtains a prediction signal X by calculating a prediction function using pixels of the previous line and the current line.

As described above, the DPCM device predicts the current pixel using the video signal that appeared in the past, and quantizes and encodes the prediction error, which is the difference between the predicted value and the current pixel value. It compresses the bandwidth.

Problems to be Solved by the Invention However, the above configuration has the following problems.

That is, since the digital video signal X passes through the subtracter 3, quantizer 4, adder 7, and predictor 8 and is the predicted value of the digital video signal It must be shorter than the period. For example, the sampling rate is 10! 1 when tllz
The loop processing must be completed within 00nsec. Within this loop, it is possible to perform complex processing such as adaptively changing the quantization characteristics of the quantizer 4 and the prediction function of the predictor 8 according to the characteristics of the video. There was a need for circuit elements that operated at high speed. However, high-speed elements such as ECL and TTL generate a lot of heat and perform complicated processing, so they are not suitable for IC implementation and are expensive. In addition, the processing within the loop may be so complex that real-time processing is difficult due to insufficient processing time.

In view of this, an object of the present invention is to provide a predictive coding device for video signals that is inexpensive, consumes little power, and is compact by using low-speed circuit elements.

Means for Solving the Problems The present invention provides means for dividing a video screen into a plurality of regions, storing a video signal in a buffer memory corresponding to each region, means for predictively encoding independently in each region, and The configuration includes means for transmitting pixels at the boundary without predictive coding.

Effect of the Invention With the above-described configuration, the present invention enables predictive coding of the signals stored in each region independently and in parallel, so that the loop processing speed of the predictive encoder does not need to be faster than the sampling speed of the video signal. , can be processed using low-speed circuit elements.

Embodiment FIG. 1 shows a block diagram of a predictive encoding device in an embodiment of the present invention. In FIG. 1, parts that are the same as those in the conventional configuration shown in FIG. 3 are given the same numbers and explanations are omitted.

1oA and 1oB are input side buffer memories, 13 is boundary area buffer memory! J, 11A, and 11B are output side buffer memories. 12A is surrounded by a dotted line.

12B is a predictive encoder, which, like the conventional configuration, includes a subtracter 3, a quantizer 4, an encoder 6, an adder 7, and a predictor 8.
It consists of

The operation of the predictive encoding device of this embodiment configured as described above will be described below.

A video signal input to an input terminal 1 is digitized by an A, D converter 2 at a sampling rate. FIG. 2 shows how the video screen is divided in this embodiment. 1
This is an example in which the field is divided into two regions perpendicular to the scanning direction. Area C is a boundary area between area A and area B,
The horizontal width is chosen according to the prediction function.

When the sampled pixel exists in area A, input buffer memo! For J10A, input side buffer memo when in area C! J10A, 10B, in the boundary area buffer memory 13, if it exists in area B, input side buffer memory 1
Write a digital video signal to 0B. If there is enough buffer memory, use the input side buffer memory IJ10A.
, 10B, it is possible to write the digital signals of the entire field into the boundary area buffer memory 13.

Input side buffer memo! The digital video signals stored in J 1 oA and 10B are sent to predictive encoders 12A and 1
2B are respectively predictively encoded. The predictive encoders 12A and 12B operate similarly to the conventional encoder shown in FIG. 3, and output encoded signals from the encoder 6.

The prediction function used in predictive encoding is shown in FIG. In the example shown in the figure, a pixel an (n=1, 2.3) on the same line as the predicted pixel I and a pixel bn (n
=-1, 1, 2, 3) to predict X.

That is, when predicting with the predictive encoders 12A and 12B, the above-mentioned pixel an + bn is required. Next, prediction of the boundary portion between area A and area B will be described.

Among the prediction functions shown in FIG. 4, functions A and C are
Do not use. In this case, the predictive encoder 12A
When performing prediction on pixels in the same area, predictive encoding can be performed using only pixels in the same area.

On the other hand, when predicting pixels in area B, when prediction is started at the left end of area B, pixels in area C are used as pixels a1+ bl necessary for predictive encoding. Pixel a before prediction point X. ,b1! Since it is sufficient that the pixel width equal to the maximum value of m among l (m≧1) exists in the region C, the pixel width of the region C only needs to be 1 when using the function A or C. By the predictive encoder 12B using the information of the pixels in the area C, the prediction error is smaller than when the left end of the area B is determined to be the edge of the video and prediction is started without the pixel of one sample before. The amount of information H can be compressed, and discontinuity between areas A and B is eliminated.

On the other hand, in the prediction functions B and D, prediction is performed using pixels . Since the maximum absolute value of n of pixel a n + bn is 1, when prediction functions B and D are used, the pixel width of region C of 1 is sufficient. Since the pixels of area C are in input side buffer memory I) 11 A, predictive encoder 1
Prediction can also be performed on the rightmost pixel of area A where prediction is performed in 2A.

Furthermore, when the prediction function E is used, pixels for three temples ahead of the current prediction point X and pixels for one sample behind the current prediction point X are required. In this case, if the pixel width of area C is 3 pixels, when predicting the rightmost pixel of area A, use the pixel of area C as b, and when predicting the leftmost pixel of area B, use a.
,, b, (e=1.2.3) can be used as pixels in area C.

Even if the width of area C is zero, that is, even if area A and area B are in contact with each other, predictive encoding can be performed independently of each other. becomes discontinuous.

As described above, by providing the area C during predictive encoding, the predictive encoders 12A and 12B can be used without the need for exchanging predicted values between the predictive encoders 12A and 12B. Each can proceed with parallel processing independently. Since one field or one frame is predicted using two predictive encoders, the processing speed is about half that of the conventional method. Pixel a used for prediction by prediction function
If the maximum absolute value of n in n+bn is N, then N is sufficient for the pixel width of area C. Further, the above processing can be performed if the buffer memories 1oA and 10B are sufficient for two lines. That is, one of the memories for two lines is used as a buffer memory for the line in which the current predicted pixel X exists, and the other is used as a buffer memory for the next line that is undergoing AD conversion.

The encoded signals predictively encoded by the predictive encoders 12A and 12B are stored in output side buffer memories 11A and 11B, respectively.
will be written to. Output side buffer memory 11A, 11B
, the data written in the boundary area buffer memory 13 is sequentially read out and output from the output terminal 6.

The predictive decoding device can independently decode each of the regions A and B using the pixels of the region C that have not been predictively encoded and read out from the boundary region buffer memory 13. Furthermore, even if a pixel in area A causes an error during transmission, the error caused by decoding and prediction will not spread to area B because area C exists.

As described above, according to this embodiment, by dividing the input video signal into area A and area B, predictive encoding can be performed in parallel for each area, and prediction processing when prediction is performed using one predictor. Predictive coding can be performed even with a predictor whose speed is half that of the loop. In addition, by providing area C, predictions for the two areas can be performed independently and with high accuracy, and by transmitting the information in area C without prediction, the accuracy of prediction on the decoding side is improved. This also makes it possible to prevent errors from spreading.

In this embodiment, one field or one frame is divided into two areas perpendicular to the scanning direction and one boundary area, but there is no problem in changing the number or direction of the areas.

It is also possible to determine the number of regions depending on the speed of the circuit element, or to change the width of the boundary region using a prediction function.

As described in detail, according to the present invention, even if the predictive encoding processing speed cannot be kept up with the conventional apparatus, it can be overcome by performing parallel processing. In addition, parallel processing frees up time, making it possible to implement even low-speed circuit elements, and making it possible to implement prediction functions that require more complex and advanced calculations, resulting in higher integration and lower costs. LS using C-MOS device that operates with low power consumption
suitable for i. By using C-MOS LSi, it is possible to realize a predictive coding device with low power consumption and miniaturization, which has great practical effects.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a predictive coding device for a video signal according to an embodiment of the present invention, FIG. 2 is a state diagram of a screen area in the same embodiment, and FIG. 3 is a diagram of a conventional predictive coding device for a video signal. The block diagram, FIG. 4, is a correspondence diagram showing the relationship between pixels and prediction functions. 1...Video input terminal, 2...A, D converter, 3...Subtractor, 4...Quantizer, 5... ... Encoder, 6... Output terminal,
7...Adder, 8...Predictor, 1oA
, 1oB... Input side buffer memory, 11A. 11B...Output side buffer memory, 12A, 1
2B... Predictive encoder, 13... Boundary area buffer memory. Name of agent: Patent attorney Toshio Nakao and 1 other person 2nd
Figure @Area C Figure 3

Claims (1)

[Claims] Means for dividing a field or frame of a video signal into a plurality of regions, storing the input video signal in a buffer memory corresponding to each region, means for predictively encoding independently in each region, and pixel signals existing at the boundaries of each region. What is claimed is: 1. A predictive coding device for a video signal, comprising means for transmitting the predictively coded signal of each region together with the predictively coded signal without performing predictive coding.