JPH0332952B2 - - Google Patents

Description

[Detailed Description of the Invention] (Technical Field) The present invention relates to a method for transmitting sub-samples of high-definition television image signals.
Low-frequency component signals for simple TV receivers can be easily obtained from sub-sampled signals, and high-definition television signals can be easily converted to standard format signals, making them suitable for both high-definition and standard format receivers. However, the present invention aims to provide a sub-sampled signal transmission method that can ensure compatibility.

(Prior art) The sub-sampling transmission method for high-definition television image signals can transmit high-definition television signals with a band equivalent to 20 MHz at approximately 8 MHz, making it an extremely efficient transmission method for future high-definition television image signals. It is a powerful means of broadcasting.

However, the multiple subsample transmission methods that have been proposed in the past also have the following drawbacks.

(1) Converting a transmitted high-definition television signal to a standard format television signal requires frame memory as it is necessary to restore the original image signal on the transmitting side, which can increase the cost of the conversion device. The sex was hot.

(2) The subsample transmission method essentially requires processing of moving parts of the image on the receiver side as well.
Processing this moving part requires advanced technology,
In addition, it is difficult to completely interpolate missing signals on the receiving side by processing moving parts of the image, so the cost of the receiver cannot be avoided.

The causes of the above two items will be explained below.

FIG. 1 shows a subsample pattern of a multiple subsample encoder. In the figure, f _S is the sampling frequency, ○ mark, △ mark, × mark, and + mark are the nth, respectively.
Corresponding to the sample points of the (n+1)th, (n+2), and (n+3)th fields,
The field constitutes the k-th frame, and the (n+2)-th and (n+3)-th fields constitute the (k+1)-th frame.

When the multiplex subsample pattern of two frames as shown in Figure 1 is sampled at a sampling frequency of f _S , the spectrum of the signal in the case of a still image is as shown in Figures 2a and 2a for the kth frame and (k+1)th frame, respectively. It will look like Figure 2b. Here, f _P is the highest frequency component of the image signal. As can be seen from Figures 2a and 2b, the sample point of the high-frequency component folded back into the low-frequency component area due to subsampling is shifted by 1/2f _S between the k-th frame and the (k+1)-th frame. Because of this, the phase appears reversed, and in the multiple sub-sampling pattern shown in Figure 1, the received image is offset for each frame, and the spatial high-frequency components on the image are turned into temporal flicker components, causing the received image to become It becomes very unsightly.

In addition, in an attempt to compress the low frequency components and the high frequency components folded back into the low frequency component region in FIGS. 2a and 2b and transmit them, and to separate both components on the receiving side, the second
Using a frame comb filter with the configuration shown in Figure d,
Even if the characteristics of diagram c are used, the separation will be insufficient. Therefore, the conversion of a high definition television signal for multi-subsample band compressed transmission to a standard television signal requires a plurality of separate frame memories to complete the conversion.

Figures 2a and 2b deal with the spectrum of the image signal in the case of a still image, but in the case of a moving image, the spectrum becomes even more complex, and it goes without saying that it becomes more difficult to separate the two components. .

Furthermore, in the sub-sample band compression transmission method, it is necessary to detect a moving part of an image as a means before interpolating a missing signal part using various interpolation methods on the receiving side. There is also a method of creating an image motion information signal on the transmitting side, multiplexing this signal with the transmission signal, and transmitting it to the receiving side, but this places a burden on the transmitting side and affects the transmission band. Therefore, it is necessary to create a moving part detection signal from the subsample signals transmitted individually on the receiving side, and in the multiple subsample pattern method shown in Fig. 1, since the frame is offset, the structure shown in Fig. 3 is A 2-frame difference signal is created from a 2-frame delayed signal by a 2-frame memory such as 2-frame memory, and a motion information signal is obtained.

As described above, with the conventional technology, it is difficult to drive a high-definition system receiver and a standard system receiver efficiently and compatiblely with the transmitted high-definition television signal without simplifying the configuration of the receiving side. It was hot.

(Summary of the Invention) It is an object of the present invention to eliminate the above-mentioned conventional drawbacks, to simplify the configuration of the receiving side, to be compatible with both the high-definition system and the standard system, and to compress the subsample band. It is an object of the present invention to provide a subsample transmission method capable of removing subsample aliasing signals, which can reproduce an original image more efficiently than a high-definition television transmission signal.

That is, the sub-sampling transmission method of the present invention transmits a high-definition television image signal in one cycle of two frames, and at a sampling frequency of 2 for both lines and frames.
In the multiple subsampling transmission method, in which samples are sequentially offset by corresponding distances on the screen sampled at twice the frequency, on the transmitting side, the original image signal is separated into low-frequency components and high-frequency components, and the high-frequency components are The component is characterized in that the same signal is transmitted on two lines, and on the receiving side, the sum of the two lines of the transmitted signal is calculated, and the aliasing component of the sub-sampled transmission signal is removed.

(Example) The present invention will be described in detail below with reference to the drawings.

FIG. 4 shows the basic configuration of the method of the present invention. That is, the input original image signal is separated into low-frequency components and high-frequency components as shown in Figure 5, and the low-frequency components remain the same, while the high-frequency components are the same for the two lines as shown in Figure 6. A signal is created using a sub-sample encoder.
Here, FIG. 5 shows the spectrum of the input signal of k frames, and FIG. 6 shows the spectrum (k frames) of the combined output signal of low-frequency and high-frequency components.

In this way, the frequency spectrum of the output signal from the sub-sample encoder becomes as shown in Figure 7b, but what should be noted here is the aliasing component.Since the sampling pattern is line offset, there is no aliasing. Ingredient is the 8th
As shown in the figure, the lines are reversed.

Here, Figure 7a shows the spectrum of the original image signal,
FIG. 7b shows the spectrum of the sub-sampled signal, and FIGS. 8a and 8b show the spectra of the sub-sampled signal of the i-th and (i+1)-th lines of the k-th frame, respectively.

Next, when this signal is decoded by an aliasing removal decoder having the configuration shown in FIG. 9, the resulting output becomes a signal without aliasing. In other words, in decoding, as shown in Fig. 10, if the sum of two lines having the same high-frequency component is taken and passed through the coefficient 1/2, the aliased component is removed as shown in Fig. 11. The signal consists only of area components. In this case the 12th
As shown in the figure, the same signal is output on the two lines. S ₂ in Fig. 9 switches to (d) and (c) in accordance with the switching to (a) and (b) of S ₁ in Fig. 4, and i, (i +
1), (i+2), (i+3) line S ₁ is (a),
(B), (A), (B) and S ₂ switch to (D), (C), (D), and (C). In addition, FIG. 10 shows the spectrum (k frame) of the input to the decoder adder in FIG. 9 on the (i+1)th line,
Figure 11 shows the spectrum (k frame) of the Figure 9 decoder multiplier output of the (i+1)th line, the first
FIG. 2 shows the spectrum (k frame) of the output of the decoder shown in FIG. 9.

When using this method, the configuration of a converter that halves the number of scanning lines, that is, the configuration for converting the number of scanning lines from a high-definition television signal to a standard television signal, for example, from 1125 lines to 525 lines, is as follows: 13th
The result will be as shown in the figure. _S3 in Fig. 13 opens and closes for each line, and the write clock frequency to the line memory for time axis conversion is (1125 x 30) x nHz, and the read clock frequency is (525 x 30) x mHz (m,
When converting from 1125 lines to 525 lines (n is a positive integer), considering that the difference between 1125/2 = 562.5 and 525 is achieved by cutting the screen, the number of scanning lines should be halved. become. As is clear from the configuration shown in Figure 13, there is no need for a frame memory or the like in this case, and the line delay 2 and adder are capable of handling the vertical shift that occurs when the scanning line is halved in any case. Since this is a necessary configuration to prevent loopback, adding these does not add any special functionality.

On the other hand, in order to obtain a one-frame difference signal for detecting a moving part, it is more convenient to have blur in the vertical direction, so the output of the configuration shown in FIG. 9 may be directly used.

Note that the sampling is line offset as shown in the first figure, so from the perspective of the configuration, the fourth
Rather, the adder shown in FIGS. 9 and 9 corresponds to using a switch to change the sampling points of two adjacent lines in a vertical line.

As is clear from the above explanation, this method corresponds to vertical subsampling of horizontal high frequency components. Therefore, if switches S ₁ (Figure 4), S ₂ (Figure 9), and S ₃ (Figure 13) lock the operating phase to the line number and do not cause frame offset, the vertical resolution of the image will decrease. I will do it. However, since the occurrence of aliasing is exactly the same for each frame, it is especially advantageous for detecting moving parts. This is called the normal type.

On the other hand, if the operating phase of switches S ₁ and S ₂ is set by frame offset, that is, the switching of switches S ₁ and S ₂ is fixed between frames and changed for each frame.
As shown in FIG. 1, the subsample pattern has already been frame offset, and since it will be synchronized with this, the vertical resolution of the high frequency component cannot be increased. Therefore, in this case, if the switch is offset by two frames, that is, if the switching of S ₁ and S ₂ is reversed every two frames, (k
+2) The frame output signal is as shown in Figure 15, and if you overlap it with the composite output spectrum diagram in Figure 6, each line will be a complete signal with all the low to high frequency components in two frames, and it will be stationary. In the case of images, complete images can be transmitted. Using this method, it is possible to restore a complete high-definition television image from low to high frequencies for each line in one cycle of two frames. This is called an offset type.

Considering the advantages and disadvantages of both types described above, it is expected that the offset type will naturally be more complex in terms of structure, and since it is offset, it will be inconvenient when performing motion detection. That is, if the frame difference is taken, there will be a case where there is no other line, and in the end, the difference between two frames will be used. On the other hand, the normal type is ideal in these respects, but has a drop in vertical resolution of high frequency components.

However, in any case, since subsample aliasing is a problem only in the case of still images, the decrease in vertical resolution of high frequency components can be improved by interpolation between fields, and the decrease in resolution of high frequency components The normal type is considered to be advantageous from a practical point of view, since it does not have as much of an adverse effect on image quality as on the left. Moreover, in the normal type, the decoding method of a high-quality receiver can basically be performed in the same way as the conventional sub-sampling method.

(Effects) According to the method of the present invention, a conversion device for receiving a broadcast of a high-definition television image signal and displaying an image on a display of a standard receiver is greatly simplified. It is also compatible with high-definition receivers.
Furthermore, if the normal type is used, a stable one-frame difference signal can be obtained, so the necessary moving part detection in the receiver in the sub-sampling system can be easily and completely performed, reducing the cost of the receiver and improving the image quality. It will be done. In particular, since detection of moving parts can be performed using a one-frame difference signal, the required capacity of the frame memory is halved.

[Brief explanation of the drawing]

FIG. 1 is a pattern showing a sampling mode of multiplexed subsamples in one round of two frames, FIG. 2 is a diagram for sequentially explaining the subsample signal and the removal of aliasing components, and FIGS. 2a and 2b are A diagram showing the spectra of sub-sampled signals of k frames and (k+1) frames, respectively, FIG. 2c,
Fig. 2d is a diagram showing the characteristics of a frame comb filter that removes aliasing components and the configuration of the frame comb filter, and Fig. 3 is a circuit for obtaining a two-frame difference signal from a frame offset subsample signal. 4 is a simple block diagram of the basic configuration of the method of the present invention, FIG. 5 is a spectrum (k frame) of the input signal of the method of the present invention, and FIG. 6 is a diagram of the method of the present invention. low range,
Figure 7 shows the spectrum of the subsampled signal, and Figure 8 shows the spectrum (k frame) of the subsampled signal of the method of the present invention. The figure shows the i-th and (i+
1) Spectrum of line sub-sampled signal,
FIG. 9 is a block diagram showing the aliasing removal decoder according to the present invention; FIG. 10 is the spectrum of the decoder adder input of the (i+1)th line in the k-th frame using the present invention; FIG. Figure 12 is the spectrum of the output of the k-frame decoder according to the method of the present invention; Figure 13 is a block diagram simply showing the configuration of the converter to 1:2 lines; 14,15,
Figures 16 and 17 are spectrum diagrams of the input signal of the 2-frame offset type (k+2)th frame, the output signal combining both low and high frequency components, the subsample signal, and the decoder output signal, respectively, according to the method of the present invention.
Furthermore, FIG. 17a is a sum of two lines, and FIG. 17b is a spectral diagram of the decoder output. 1...Low pass filter, 2...Line delay, 3...Line memory for time axis conversion, 4...Frame memory, 5...Absolute value unit.

Claims (1)

[Claims] 1. In a multiplex sub-sampling transmission method in which a high-definition television image signal is sampled in one cycle of two frames and sequentially offset by corresponding distances on the screen where both lines and frames are sampled at a frequency twice the sampling frequency, On the receiving side, the original image signal is separated into low frequency components and high frequency components, and the high frequency component is transmitted as the same signal on two lines, and on the receiving side, the sum of the two lines of the transmitted signal is calculated, and the high frequency component is transmitted as the same signal on two lines. A subsample transmission method characterized by removing folded components of a subsample transmission signal.