JPH0389689A - Coding device for moving picture signal - Google Patents

Abstract

PURPOSE:To improve the quantity of a reproduced picture by comparing a threshold level and an orthogonal transform coefficient different from a block in the case of a luminance signal and a chrominance signal in response to the quantization characteristic in the unit of blocks and quantizing the orthogonal transform coefficient or zero depending on the quantity. CONSTITUTION:A significant/nonsignificant error block discrimination circuit 16 calculates a quantization step width (g) and a succeeding threshold level Th as a function Yc as to whether the block is a luminance signal or the color difference signal at first; Th=f(g, yc). In the case of Cij<=Th, where Cij is the orthogonal transform coefficient of the block, it is discriminated as a nonsignificant error block, and when the equation is not satisfied, it is discriminated as a significant error block. In the case of discrimination as the nonsignificant error block, a coefficient quantization circuit 7a quantizes zero. In the case of the discrimination of the significant error block, an orthogonal transform circuit 6 calculates the coefficient, which is quantized. Thus, coding of noise is reduced and the quality of the reproduced picture is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a moving image signal encoding device used for videophone calls, video conferences, and the like.

BACKGROUND OF THE INVENTION Conventional high-efficiency coding devices for video signals have either performed prediction that involves a mixture of motion-compensated inter-frame prediction and intra-frame prediction, or performed either one of the above-mentioned motion-compensated inter-frame prediction or intra-frame prediction. Alternatively, a prediction error value is calculated by performing prediction that selectively switches between motion-compensated inter-frame prediction and intra-frame prediction, and this prediction error value is calculated for each block that is a set of multiple pixels (for example, 8 x 8). It is known that the coefficients are orthogonally transformed and the coefficients are encoded.

FIG. 3 shows the configuration of a conventional high-efficiency encoding device for moving image signals.

In FIG. 3, when an analog moving image signal is converted into an upstream digital moving image signal by an A/D conversion converter, a motion vector detection circuit 2 uses this digital moving image signal and the previous frame read out from a frame memory 3. Using the reproduced pixel values, a motion vector is calculated for each block, which is a set of a plurality of pixels (for example, 8×8).

The prediction circuit 4 uses this motion vector and the frame memory 3.
Using the reproduced pixel values of the current frame and/or the previous frame read from the previous frame, prediction is performed using a mixture of motion compensated interframe prediction and intraframe prediction, or one of the above motion compensated interframe prediction and intraframe prediction is performed. A predicted value is calculated by performing prediction that selectively switches between the motion-compensated inter-frame prediction and intra-frame prediction.

The subtracter 5 subtracts the predicted value from the digital video signal from the A/D converter 1 to calculate a prediction error value, and the orthogonal transform circuit 6 converts this prediction error value into a set of a plurality of pixels. The coefficient quantization circuit 7 calculates the coefficients by performing orthogonal transformation for each block, and then the coefficient quantization circuit 7
The orthogonal transform coefficients are quantized with a quantization step width determined by 5, and the quantized coefficient encoding circuit 8 encodes the quantized orthogonal transform coefficients.

The code of this quantized orthogonal transform coefficient is determined by the line coding circuit 13 along with the quantization step width determined by the quantization step width control circuit 15 and the motion vector detected by the motion vector detection circuit 2. The data is then speed-smoothed by the transmission buffer 14 and sent to the line. The quantization step width control circuit 15 increases the quantization step width to suppress the amount of generated information when the amount of residual information in the transmission buffer 14 is large, and decreases the quantization step width when the amount of residual information is small. The image is made smaller and transmitted.

The sign of the quantized orthogonal transform coefficient is locally decoded by a quantized coefficient decoding circuit 9, and then a coefficient inverse quantization circuit 10 performs a quantization step determined by a quantization step width control circuit 15. The width is inversely quantized into orthogonal transform coefficients. Furthermore, the orthogonal inverse transform circuit 11 reproduces the prediction error value, and the adder 12 adds this prediction error value to the predicted value from the prediction circuit 4 to reproduce the pixel value and writes it into the frame memory 3.

Therefore, according to the above conventional example, by adaptively controlling the quantization step width according to the amount of residual information in the transmission buffer 14, for example, 64 Kb/s, 384 Kb/s, etc.
Moving images with little deterioration in image quality can be transmitted both at a constant rate such as S and at a variable rate.

Problems to be Solved by the Invention However, in the conventional video signal encoding device described above, when noise is superimposed on the input signal, this noise is encoded in the same way as the image signal, and therefore the effective There is a problem that the amount of information for transmitting the image signal decreases, and the reproduced image deteriorates.

In view of the above-mentioned conventional problems, the present invention provides a moving image that does not reduce the amount of information for transmitting an effective image signal when noise is superimposed on the input signal, thereby improving the reproduced image. An object of the present invention is to provide an image signal encoding device.

Means for Solving the Problems In order to achieve the above object, the present invention has the following objectives: When noise is superimposed on an input signal, the influence of the noise increases near the boundary of the quantization characteristic, and the noise pattern becomes different from the luminance signal. Considering that it differs depending on the color difference signal, in addition to depending on the quantization characteristics, the values of the threshold value and orthogonal transformation coefficient, which are different when the block is a luminance signal and when the block is a chrominance signal, are compared for each block. The orthogonal transform coefficient is quantized when the value of the orthogonal transform coefficient at a predetermined position is larger than the threshold value, and "0" is quantized when the value of the orthogonal transform coefficient at a predetermined position of the block is smaller than the threshold value. This is how it was done.

According to the present invention, with the above configuration, it is possible to reduce the amount of noise to be encoded, so that the amount of information for transmitting an effective image signal does not decrease, and the reproduced image can be improved.

EXAMPLES Hereinafter, the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a moving picture signal encoding device according to the present invention, and FIG. 2 is an explanatory diagram showing characteristics of the encoding device of FIG. 1.

In FIG. 1, 1 is an A/D converter that converts an analog moving image signal into a digital moving image signal, and 2 is a digital moving image signal from the A/D converter 1 and the digital moving image signal before being read from the frame memory 3. Using the reproduced pixel values of the frame,
This is a motion vector detection circuit that calculates a motion vector for each block, which is a set of a plurality of pixels (for example, 8×8).

4 uses this motion vector and the reproduced pixel values of the current frame and/or previous frame read from the frame memory 3 to perform prediction in which motion-compensated inter-frame prediction and intra-frame prediction are mixed; This is a prediction circuit that calculates a predicted value by performing either inter-frame prediction or intra-frame prediction, or by performing prediction that selectively switches between the motion-compensated inter-frame prediction and intra-frame prediction.

5 is a subtracter that calculates a prediction error value by subtracting the predicted value from the digital video signal from the A/D converter 1;
6 is an orthogonal transform circuit that orthogonally transforms this prediction error value for each block, which is a set of a plurality of pixels, and calculates its coefficients.

16 indicates whether the block orthogonally transformed by the orthogonal transform circuit 6 is a significant error block or not based on the orthogonal transform coefficient calculated by the orthogonal transform circuit 6 and the quantization step width determined by the quantization step width control circuit 15. This is a significant/non-significant error block determination circuit that determines whether the block is a significant error block and switches the switch 17 based on the determination result.

7a is a coefficient quantization circuit that quantizes the orthogonal transform coefficient or "0" with the quantization step width determined by the quantization step width control circuit 15; 8 is a coefficient quantization circuit that encodes the quantized orthogonal transform coefficient; This is a quantization coefficient encoding circuit.

9 is a quantization coefficient decoding circuit that locally decodes the code of the quantized orthogonal transform coefficient from the quantization coefficient encoding circuit 8; 10 is a quantization step determined by the quantization step width control circuit 15; 11 is an orthogonal inverse transform circuit that reproduces a prediction error value using this orthogonal transform coefficient; 12 is a coefficient dequantization circuit that inversely quantizes into orthogonal transform coefficients based on the width; 12 is an orthogonal inverse transform circuit that reproduces a prediction error value using this orthogonal transform coefficient; This is an adder 12 for reproducing pixel values by adding values and writing them into the frame memory 3.

13 is the code of the quantized orthogonal transform coefficient from the quantization coefficient encoding circuit 8, the quantization step width determined by the quantization step width control circuit 15, and the motion detected by the motion vector detection circuit 2. A line encoding circuit 14 for line encoding a vector is a transmission buffer for smoothing the speed of the code from the line encoding circuit 13 and sending it out to the line, and a quantization step width control circuit 15
When the amount of residual information is large, the quantization step width is increased to suppress the amount of generated information, and when the amount of residual information is small, the quantization step width is decreased to make the image finer and transmitted.

Next, referring to FIG. 2, we will explain the operation of the above embodiment, especially the significant/
The operation of the non-significant error block determination circuit 16 will be explained.

First, when noise is superimposed on the input signal, a characteristic pattern appears in the orthogonal transform coefficients of the prediction error calculated by the orthogonal transform circuit 6. That is, when white noise is superimposed, noise is superimposed on all orthogonal transform coefficients, and when noise due to flicker is superimposed, noise is superimposed on a specific orthogonal transform coefficient.

Also, depending on the input signal processing method, the way noise is superimposed on the luminance signal and color difference signal differs; for example, NT
When separating the luminance signal and color difference signal for the SC signal,
When a simple bandpass filter is used, the high frequency components of the luminance signal are included in the color signal as noise, but the comb filter can remove the above noise.

Next, considering the increase in the amount of transmitted information due to quantization characteristics and noise, generally speaking, when the amount of transmitted information increases due to noise for a certain quantization characteristic, the value input to the coefficient quantization circuit 7a is This is the case near the boundary of the region determined by its quantization characteristics.

For example, in the case of a quantization characteristic as shown in Fig. 2, the output value may be "0" or r T depending on whether the input value is Th-δ or Th+δ (δ>0). h
+g/2J. Furthermore, in this quantization characteristic, Th, T
The influence is large when noise is superimposed on the input values near h+g and Th+2g.

Therefore, the significant/insignificant error block determination circuit 16 uses the step width g determined by the quantization step width control circuit 15, the orthogonal transform coefficient C6 calculated by the orthogonal transform circuit 6, and whether the block is a luminance signal or a color difference signal. Depending on the signal, use the following method to judge.

First, the next threshold Th is calculated as a function (yc) of the quantization step width g and whether the block is a luminance signal or a color difference signal.

Th=f (g, yc) ... (1) Here, this number f (g, yc) varies depending on the quantization characteristics of the coefficient quantization circuit 7a, and the quantization as shown in FIG. In the case of a luminance signal, f (g, yc) = α × g ... (2)
In the case of a color difference signal, f (g, yc)=1/2·α×g (3). α is a constant.

Next, when Cz≦Th (4) for the value C of the orthogonal transform coefficient of that block, that block is determined to be a non-significant error block, and when equation (4) is not satisfied, the block is It is determined that the block has a significant error.

If the block is determined to be a non-significant error block, the coefficient quantization circuit 7a controls the switch 17 to quantize "0", and if the block is determined to be a significant error block, the coefficient quantization circuit 7a performs orthogonal transformation. The switch 17 is controlled to quantize the coefficients calculated by the circuit 6.

Therefore, according to the above embodiment, depending on whether the block is a luminance signal or a color difference signal, the quantization step width g
Since the threshold value Th proportional to is calculated, if more noise is superimposed on the luminance signal than on the color difference signal, it is possible to suppress an increase in the amount of information transmitted due to the superimposed noise.

Also, since the threshold Th proportional to the quantization step width g is calculated, no matter what value the quantization step width g is,
It is possible to efficiently suppress an increase in the amount of information transmitted due to superimposed noise.

Furthermore, according to the above embodiment, since the threshold Th is compared with the values of all the orthogonal transform coefficients C11 of the block,
It is possible to suppress an increase in the amount of information transmitted due to white noise.

In the above embodiment, the threshold Th is set to a different value depending on whether the block is a luminance signal or a chrominance signal, but if noise is superimposed in the same way regardless of whether the block is a luminance signal or a chrominance signal. One threshold may be used for each.

In addition, instead of comparing the values of all the orthogonal transform coefficients C of the block with the threshold value Th, by comparing the orthogonal transform coefficients of a predetermined position with the threshold value Th, information transmitted by noise of a frequency component of a certain characteristic can be obtained. The increase in amount can be efficiently suppressed.

As described in detail, the present invention provides that when noise is superimposed on an input signal, the influence of the noise increases near the boundary of the quantization characteristic, and the noise pattern changes depending on the luminance signal and color difference signal. In consideration of the difference, in addition to depending on the quantization characteristics, the threshold values and orthogonal transformation coefficient values that are different when the block is a luminance signal and when the block is a chrominance signal are compared for each block, and the orthogonal transform coefficient at a predetermined position in the block is calculated. The orthogonal transform coefficient is quantized when the value of the transform coefficient is larger than the threshold value, and "0" is set when the value of the orthogonal transform coefficient at a predetermined position in the block is smaller than the threshold value.
Since the image data is quantized, it is possible to reduce the amount of noise to be encoded, so that the amount of information for transmitting an effective image signal does not decrease, and the reproduced image can be improved.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of a video signal encoding device according to the present invention, FIG. 2 is an explanatory diagram showing characteristics of the encoding device of FIG. 1, and FIG. 3 is a conventional FIG. 2 is a block diagram showing a video signal encoding device of FIG. 1... A/D converter, 2... Motion vector detection circuit, 3... Frame memory, 4... Prediction circuit, 5...
- Subtractor, 6... Orthogonal transform circuit, 7a... Coefficient quantization circuit, 16... Significant/non-significant error block determination circuit, 17... Switch.

Claims (1)

[Scope of Claims] Means for calculating a predicted value for an input pixel based on frame correlation of an input image signal, and calculating a prediction error between the input pixel value and the predicted value, and blocking orthogonal transformation coefficients using the prediction error. means for calculating each unit; means for quantizing the orthogonal transform coefficient; and a threshold value that is determined to be a different value depending on the characteristics of the quantization means, depending on whether the block is a luminance signal or a color signal. and the value of the orthogonal transform coefficient on a block-by-block basis, and when the value of the orthogonal transform coefficient at a predetermined position in the block is larger than the threshold value, the quantizing means quantizes the orthogonal transform coefficient. and means for controlling the quantizing means to quantize "0" when the value of the orthogonal transform coefficient at a predetermined position in the block is smaller than a threshold value.