JPS6026356B2 - Color television signal sampling processing method - Google Patents

Description

[Detailed Description of the Invention] ○} Field of Application of the Invention The present invention samples a composite color television signal using a sampling processing method, particularly the NTSC method, at a sampling frequency lower than twice the signal band. The present invention relates to a signal processing method for reproducing color television signals, particularly to a color television signal sampling processing method when used in a still image transmission device.

[21 Conventional technology In a composite color television signal such as the NTSC system, the color signal is modulated at a subcarrier frequency (hereinafter referred to as fsc), becomes a carrier color signal (hereinafter referred to as C signal), and becomes a luminance signal (Y signal). are superimposed and frequency multiplexed.

In order to efficiently encode this composite color television signal into a digital signal, it is first necessary to perform sampling. When sampling a composite color television signal, the sampling frequency fs needs to be at least twice the composite color television signal band fc (so-called Nyquist frequency), and is usually three times fsc (approximately 11 MHz). frequency was selected. For this reason, the transmission bit rate of encoded signals has become extremely high. In order to eliminate this problem, a so-called sub-Nyquist sampling method is known in which the sampling frequency fs is lowered to less than twice the signal band width fc. The frequency spectrum of a composite color television signal, as shown in FIG. 1, consists of a Y signal and a C signal centered at the subcarrier frequency fsc.

This is shown in Figure b in detail, where the Y signal has a power peak at an integer position of the horizontal scanning frequency fH.
In addition, there is a signal peak at fsc±ifH, centered on the C signal 'MafSC nire' (Mahikyou fH and Tetsuro Koru), and Y,
The C signals are arranged so that the peaks and troughs of each signal correspond to the peaks and troughs of the other signal. Further, if we look at these in more detail, it is known that both the Y signal and the C signal consist of rope spectra that are integral multiples of the vertical scanning frequency.

If this is sampled at a frequency fs that is less than twice the signal band 0 to fc, its harmonic components (shown by broken lines in FIG. 1a) overlap between the frequency band (fs-fc to fc).

One of the conventionally known subnykist sampling methods is
Limit the band width of the Y signal to 0~(fS~fC), and set fS to n
fH is chosen to be the same as fH, so that the peaks and valleys of each fH of the C signal and its high frequency signal do not overlap with each other. In this case, in order to remove the harmonic components and reproduce the original television signal, the Y signal is limited to fs-fc, and the C signal is limited to the range around iL to f. C
For signals, the vertical band of the image is 1 of the original signal band.
This corresponds to limiting the value to /2. For this reason, there is a problem in that the high frequency component of the Y signal of the reproduced image is particularly removed, resulting in an image with low resolution. In the second conventional example, f
S is selected as fH for MH, and the peak of the harmonic component is ifH
1/4 fH.

In this case, in order to separate the Y signal, the C signal, and their harmonics, the bands around iL and (i+k)fM are limited to fH for both the Y and C signals. This corresponds to limiting the vertical band width of the image to the original signal band width, and the frequency band (
Since the signal band is narrowly limited in the vertical direction between fS-fC) and fC, there is a problem in that the deterioration of the color edge portion in the horizontal direction is particularly noticeable. In this second example, interpolation is performed using a signal two horizontal scanning periods before (or after, hereinafter referred to as before and after) when the phase of the subcarrier signal becomes the same as the signal of the current scanning line, and the interpolation distance is Since they are separated (assistance between 4 horizontal scanning lines considering interlacing), the deterioration becomes noticeable at vertical changes in the image (ie, at horizontal edges). The third conventional example is one field before (or after) where the phase of the subcarrier signal matches the current scanning line.
Specifically, at the signal before 26th group or after 26th group,
The sampling phases are varied by 1800, and by switching and using them, a signal with a sampling frequency of phantom s is created.

This method focuses on the spectrum of the color television signal at each vertical scanning frequency fH/525,
Field frequency (
To be precise, this corresponds to processing with a rectangular filter having a pass band and a blocking band for each fH/262).
Therefore, 30 days 2 (fH
This means that approximately 1/2 of the spectrum for every /525) has been removed, and in particular, the removal of the blind line spectrum is also performed for the bands from 0 to (fs to fc) where no harmonic components are present due to sampling. As a result, the resolution of the horizontal stripes deteriorates, and a false signal of frequency fs appears as described later at the edge portion of the horizontal stripes where the direct current changes greatly between scanning lines, resulting in unsightly deterioration of image quality. Figure 2 shows this situation, where a indicates the i-th scanning line (luminance signal Y, chrominance signal C,), and (i-262) scanning line (scanning line immediately below the i-th scanning line). , the brightness signal Y, -Crab2, the color signal C, -Dark2) and their sample points are indicated by circles. If there is a large change in the luminance level between the two scanning lines, the composite signal of the i-th scanning line (c in the same figure) will be
) The sample value of the scanning line is interpolated between the sample values of the i-th scanning line (indicated by the signal marked c in the figure), resulting in a signal that is significantly different from the original signal (a in the figure). Put it away. [31
OBJECTS OF THE INVENTION It is an object of the present invention to realize a sampling frequency and a signal processing method that minimize deterioration in image quality when a composite color television signal is subjected to Subnyquist sampling and reproduction.

Furthermore, it is an object of the present invention to eliminate the image quality deterioration in the third conventional example described above and to realize a sub-Nyquist sampling processing method that utilizes high-quality inter-field correlation.

[41 General Description of the Invention] In order to achieve the above object, the present invention provides sampling of the current scanning line and the scanning line 26 before or after 26 which is closest to it on the screen and has the same subcarrier phase. The phases of the signals are made to differ from each other by 1800 degrees, and the current scan line signal is converted into a direct current or (fs-fc) signal without the influence of subnykist sampling, i.e., without harmonic signals due to sampling. The ~fc band is used to create a standard signal during interpolation by averaging the current scanning line and signals separated by 262 days from it, thereby creating a publicly sampled signal.

According to the signal processing method of the present invention, for the reasons described below, even when the brightness signal and color signal of an image change suddenly in the vertical direction, there is little deterioration in the image quality based on the Tachibana signal, and the image quality of the reproduced image is lower than that of the conventional one. This is significantly improved compared to the proposed method.

In other words, since the low frequency component of DC ~ (fs-fc) where there is no harmonic component due to sub-Nyquist sampling is interpolated from the signal of the current scanning line, the i-th scanning line and the (i-2
62) In the case of a and 0 in Fig. 2 where the DC values of the luminance signals differ greatly between scanning lines, the DC level of the signal interpolated by the present invention is the DC level of the i-th scanning line (L, in d of the same figure) becomes.

On the other hand, in bands from (fs-fc) to fc, where harmonic components due to Sapnaikist sampling exist, the high-frequency components of the i-th scanning line (day d in Figure 2) and the (i-262
) When we calculate the average value of the high frequency components of the scanning lines (day e, -282 in Figure 2), we find that the harmonic components due to sampling have a phase difference of 180o between the i-th scanning line and the (i-262)th scanning line. Therefore, they cancel each other out, and only the components of the original signal can be extracted. Therefore, the interpolated signal (FIG. 2f) can accurately reproduce the original signal (FIG. 2a). '51 Examples Hereinafter, the present invention will be explained in detail with reference to examples.

FIG. 3 shows a block diagram of one embodiment of the sampling processing method according to the present invention.

In the same figure, the composite color television signal in the frequency band fc inputted in mode 1 is sampled at the sampling frequency fs (however, fc x fs x c) in the analog-to-digital conversion circuit 2, and converted into a digital code. Ru. This sampling frequency fs is 26 (horizontal scanning period) before (
Alternatively, the phase is adjusted to be 180° different from the sample phase of (after). This can be realized by creating two types of sampling signals fs, fs2 having mutually different phases by 180° from a signal of frequency as synchronized with the horizontal synchronizing signal, and by switching and selecting them. The signal sampled at fs and digitized is transmitted through the transmission line 3. Note that in order to reduce the transmission capacity, it is possible to install compression code circuits and decoding circuits (not shown) before and after the transmission line 3.
Further, the transmission line 3 can also be replaced with a storage medium.
The transmitted digital signal of frequency fs is input to the field memory 4, and a signal delayed by exactly 262 days with respect to the input signal is obtained at its output. The received signal xi and the output signal (xi-262) of the field memory 4 are switched between code "0" and "0" at a speed of s by switching circuits 5 and 6, respectively, and "0" is inserted between the sample value and the next sample value. The result is a signal with a repetition frequency of 23. The power of these signals is 1'2 of the original signal. The output of the switching circuit 5 is set to 0 by a digital low-pass filter (LPF) 7 and a double coefficient circuit 34.
The band signal xr of ~(fs~fc) is extracted, and the signal xr of band width (f
s-fc) to fc signals xr/2 are extracted. The output of the switching circuit 6 is output by BPF9 having the same configuration as BPF8.
Frequency band (fs-fc) − for 262 signals in x
Every 262/2 signals x of fc are extracted. BPF8 and 9
Since the harmonic components resulting from sampling contained in the output signals xr/2 and xL2/2 have opposite phases to each other, they are canceled out by the adder circuit 10. Therefore, when the output of the adder circuit 10 and the output of the coefficient circuit 34 are added by the adder circuit 11, a repeating digital signal containing no harmonic components is obtained as the output, and this is converted into an analog signal by the digital-to-analog conversion circuit 12. By converting, a desired composite color television signal that does not contain harmonic components due to sampling can be obtained from the signal subjected to Subnyquist sampling at fs at the terminal 13. In Fig. 3, the signal yi with repetition rate s obtained by the adder circuit 11 can be expressed as yFXt 10 × 10 × 262, which corresponds to yF + X ÷ + -262, it can be separated into a received signal and an interpolated signal, and can be configured as shown in FIG.

That is, when the output xr of the LPF 14 of the received signal xi and the output x of the BPF 15 of the output signal x'-262 of the field memory 4 are added by the adder circuit 16, the inter-tub signal x
r+x262 is obtained. This and the received signal are coded “
Each time a ``0'' is inserted, the switching circuits 17a and 1 generate a repeating signal of s, and operate in opposite directions at an operating speed of 2 s.
By switching at 7b and adding at adder circuit 18, a signal yi of repetition of s is obtained. FIG. 5 shows a block diagram of an embodiment in which the present invention is applied to a still image transmission device.

In other words, terminal 19 is connected to a color camera (
A composite color television signal captured by an analog-to-digital conversion circuit 20 (not shown) is inputted, subjected to sub-Nyquist sampling at the above-mentioned frequency fs, and stored in a frame memory 21. The contents of the frame memory 21 are read out at a slow speed, transmitted through the narrow transmission line 22, and stored in the frame memory 23 on the receiving side. Note that it is possible to install a compression encoding circuit, a decoding circuit, a modulation/demodulation circuit, etc. in order to shorten transmission time before and after the narrowband transmission line. The frame memory 23 on the receiving side consists of two field memories 24 and 25, which each have a 26-frame memory.
Scanning lines separated by a period of time are laid out so that they are output at the same time.

Switching circuits 26a and 26b whose outputs are reversely connected to each other
If the signal is switched every field period, xl can be obtained as the output of 26a, and a signal xi-282 in which 26 is before or after xi can be obtained as the output of 26b. Therefore, an interpolated signal is obtained from these signals by the LPF 27, the BPF 28, and the adder circuit 29, and when this and X are switched by the switching circuit 30 at the common s, a signal with the repetition frequency as is obtained. (It is clear that the operation of the switching circuits 17a and 17b addition circuits in FIG. 4 can be replaced by the switching circuit 301.) Since this signal is the same color subcarrier signal every frame, the color subcarrier phase is inverted every frame. After passing through the color phase adjustment circuit 31, the digital
By converting the signal back to an analog signal using the analog conversion circuit 32, a correct composite color television signal can be obtained. In this embodiment, the frame memory 23 is essential as a still image transmission device, and the present invention has the advantage that it can be implemented without providing a special field memory. Note that the color phase adjustment circuit 31 is realized by an analog circuit,
It is also possible to install it after the digital/analog conversion circuit 32.

Furthermore, ``In each embodiment, before subnykist sampling at frequency fs on the transmitting side, a field listening filter having circuit characteristics similar to those shown on the receiving side is installed,
It is possible to reduce aliasing noise caused by sampling by removing components that enter the band in advance.

This can be easily achieved by operating the analog/digital conversion circuit at normal s, providing a circuit similar to that on the receiving side for this signal, and resampling its output at fs. '6} Summary As explained above, according to the present invention, noise caused by sampling a composite color television signal at a frequency lower than twice its band fc is eliminated by using inter-field correlation.
It can be removed using field memory, LPF, BPF, etc., and a high quality image can be regenerated.

Note that each block shown in the example can be easily realized using general-purpose logic circuit elements and memory elements, and the pulses etc. that control them can be easily realized as a timing circuit (not shown). Therefore, detailed explanation of the specific logical configuration will be omitted.

[Brief explanation of the drawing]

FIG. 1 is a frequency spectrum diagram of a composite color television signal, FIG. 2 is an operational waveform diagram of the conventional method and the method of the present invention for subnyquist sampling processing of a composite color television signal, and FIGS. FIG. 5 shows a block diagram of an embodiment of the invention. 2, 20...analog-to-digital conversion circuit, 4,
24, 25...Field memory, 7, 14, 27...
・...Digital LPF, 8,9,1 5,28...Digital BPF, 1 0,1 1,1 6,18,29・
...Digital addition circuit, 12, 32...
・Digital/analog conversion circuit. Multi J General Tsutomu no Zu 3 Box Multi 4 Captive Leopard Legs

Claims (1)

[Claims] 1 In a color television signal sampling processing method in which a composite color television signal in frequency band f_c is sampled and transmitted at a sampling frequency f_c lower than 2f_c, and the original composite color television signal is reproduced from the transmitted signal. , the sampling phase of two samples separated by a time equal to 262 times the horizontal scanning period H with respect to the sampling frequency f_s is 180
° set differently and store odd field signals and even field signals of the transmitted composite color television signal in first and second field memories, respectively; A first signal and a second signal separated by a time of 262H from each other are extracted, a low frequency component of the first signal and a high frequency component of the second signal are added to obtain a third signal, and a third signal is obtained by adding the low frequency component of the first signal and the high frequency component of the second signal. and the third signal above at a frequency 2 which is twice the sampling frequency.
A color television signal sampling processing method characterized in that a composite color television signal sampled at a sampling frequency of 2f_s is obtained by alternately switching at a sampling frequency of 2f_s, and the composite color television signal is reproduced from the obtained composite color television signal.