JPS60130986A - Video signal processing unit - Google Patents

Abstract

PURPOSE:To decrease the noise in a video by improving the S/N of a VTR by means of a line noise canceller through the sampling at a frequency being less than two times of the upper limit frequency of an input video signal and at a frequency of fS=1/2fH(2n-1). CONSTITUTION:The input video signal is converted into a digital signal sampled by a sample frequency fS' by an AD converter 1. The digitized video signal is fed to an adder 8 and a subtractor 3 as it is and via a 1H memory 2. An output of the adder 8 is fed to a synthesizer 11 and an output of the subtractor 3 is fed to the synthesizer 11 via a low pass filter 9 and a slicer 10. An output of the synthesizer 11 is restored into an analog signal by the DA converter 7 and becomes an output video signal.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a video signal processing device that can be used in video tape recorders and the like, and more specifically to a circuit for reducing noise contained in a video signal by sample processing. It is about downsizing.

Conventional configuration and its problems In consumer video tape recorders (hereinafter referred to as VTRs) that are currently widely used, the luminance signal is FM modulated and placed on a magnetic tape, and the playback system reproduces the original signal. It is demodulated into a luminance signal.

However, the signal-to-noise ratio (S/N) of the tape head system
) [S of the luminance signal to be reproduced is not necessarily sufficient.
/N'i There are many cases where you want to improve.

Many attempts have been made to improve the S/N ratio of luminance signals, and one typical example is a method that utilizes the correlation of luminance signals and is called a line noise canceller.

This focuses on the line period nature of the luminance signal and uses a Y-shaped comb filter to remove unnecessary spectra to improve the S/N ratio. At this time, measures are taken to prevent the vertical resolution of the image from deteriorating. It is. Conventionally, this type of line noise canceler [1H delay circuit] used an ultrasonic glass delay line, but in that case, there were problems such as the size of the component itself being large, the frequency characteristics of the pass band being poor, and the insertion loss being large. However, it was not always satisfactory.

In order to solve these problems, recently a method of converting the 1H delay circuit into a semiconductor has been considered. FIG. 1 is a configuration diagram showing an example thereof.

The input video signal demodulated into the video signal reproduced from the VTR is converted into a frequency,
The signal is sampled at f3, converted into a digital signal, and applied to difference calculation circuits 3 and 6, as well as supplied to 1H memory 2. The output of the 1H memory 2 is given to a difference calculation circuit 3, and the output of the difference calculation circuit 3 is inputted to the difference calculation circuit 6 via an attenuator 4 (indicated in %) and a limiter 6, and its output is sent to a FiDA converter 7 (D/A ) and becomes the output video signal.

In the configuration shown in FIG. 1, the baseband input video signal is digitized and all of the line noise cancelers are executed in the form of digital signals. 1H memory 3 and difference calculation circuit 3 [
A C-shaped comb filter is formed, and its output is processed in an attenuator 4 so that the amplitude becomes %, processed in a limiter 5 so that signals with a predetermined amplitude or less are passed, and is added to a difference calculation circuit 6 and cD. The difference calculation circuit 6 outputs the difference between the input signal and the C-shaped comb filter to K when the output of the C-shaped comb filter, which is the output of the difference calculation circuit 3, is small (when the line correlation of the input luminance signal is strong). Therefore, this is equivalent to a Y-comb filter. When the line correlation of the input luminance signal is weak and the output of the C-shaped comb filter is large, the output of the t/iC-shaped filter does not pass through the limiter 6, so that the input signal appears as it is at the output of the difference calculation circuit 6. The output of the difference calculation circuit 6 is returned to an analog signal by the DA converter 7.
*The output video signal ends up passing through the LDY-type comb filter in the parts where the correlation of the input video signal is strong, so that unnecessary spectra are attenuated and the nS/N is improved, and in the parts where the correlation is weak, the input signal is passed through the LDY comb filter. Since only part 1 is output, there is no deterioration in vertical resolution. Here, the attenuator 4 is provided to attenuate the amplitude by 3 Avc to match the level with the input because the amplitude of the C-shaped comb filter which is the output of the difference calculation circuit is twice that of the input.

In Fig. 1, since the 1H memory 2 can use a shift register or random access memory (RAM), it does not in principle cause amplitude attenuation, frequency characteristic fluctuations, or time errors, and can be used as a conventional glass delay line. It has the feature of being an ideal delay element.

However, the reason why the line noise canceler shown in FIG. 1 has not been put into practical use is the fundamental problem of increasing the circuit scale. In other words, since it is necessary to select a sampling frequency that is at least twice the maximum frequency of the input video signal, the AD converter 1, DA converter 7, and 1H memory 2 are required to operate at high speed, and the current consumption of the two circuits is inevitably large. Become something and change it.

FIG. 2 shows a spectrum distribution diagram for explaining the conditions of the sampling frequency fs. If the frequency band of the input video signal is, for example, the maximum frequency of 31M as shown by fB in Figure 2, the sampling frequency must be selected to be 16Vh or higher in order to prevent aliasing distortion, and in this example, the sampling frequency must be selected to be 7Vh or higher. It is set. In this case, the upper side wave fU1 of the lower measurement wave fL of the sample frequency fs is generated, but since the lowest frequency of fL does not overlap with fB, aliasing distortion does not occur.If only fB is taken out with a low-pass filter at zero shift, no distortion will occur. A complete video signal without any distortion can be restored. In this way, in order to prevent aliasing distortion, the sampling frequency must be selected high. For example, in the example shown in Figure 2, if the resolution of the AD converter L is required to be 7 bits, the transmission bit rate will be Is 49 MB/sec. There is a problem in that the circuit scale is very large because it depends on the transmission bit rate.

Purpose of the Invention An object of the present invention is to reduce noise in a video signal without deteriorating the vertical resolution, and in this case, the sampling frequency should be set to less than twice the highest frequency of the video signal. It is an object of the present invention to provide a video signal processing device that can easily be made into a semiconductor by reducing the circuit scale of a 1H delay circuit and the depletion current.

Structure of the Invention The video signal processing device of the present invention includes a sampling means for sampling at a frequency of fB-j4fH (2n-1), which is less than twice the upper limit frequency of the input video signal, and delaying the sampled information by one horizontal scanning period. A comb filter means for forming a comb characteristic using a delay circuit.The comb filter means is equipped with a processing means and a signal synthesis means for manipulating the amplitude and frequency characteristics of the output of the comb filter means, and a Y-shaped comb is formed in a part of the band of the input video signal. Y only when line correlation is strong in some bands.
The 1H delay circuit uses semiconductor memory IJk, so its characteristics can be dramatically improved compared to conventional glass delay lines, and the sampling frequency can be lowered. Since the transmitter can be made smaller, it can be easily integrated into an IC, making it possible to reduce the size and cost of the VTR circuit.

Description of examples Examples of the present invention will be described below.

FIG. 3 is a block diagram of the main parts of a line noise canceler for video signals showing an embodiment of the present invention. In the figure, blocks having the same functions as in FIG. 1 are indicated by the same numbers.

The input video signal is sampled by the AD converter 1 at a sampling frequency f'S't1. converted into a digital signal
Ru. The digitized video signal is shuffled and passed through the 1H memory 2 before being applied to an adder 8 and a subtracter 3. The output of the adder 8 is applied to a synthesizer 11, and the output of the subtracter 3 is applied to a synthesizer 11 via a low-pass filter 9 (indicated by LPF) and a slicer 10.The output of the synthesizer 11 is converted into an analog signal by the FiDA converter 7. restored to n output j
It becomes a video signal.

Now, as is well known, digitized video signals and their
(The sum with the H1H delayed signal exhibits a Y-shaped comb characteristic as shown in Figure 4 (-), and the difference is C as shown in Figure 4 (b).
Exhibits comb-shaped characteristics. The output of the adder 8, which is the output of the Y-shaped comb filter, is directly applied to a synthesizer 11 for synthesizing signals. On the other hand, the output of the subtracter 3, which is the output of the C-shaped comb filter, passes through a low-pass filter 9 having a cutoff frequency within the band of the input video signal, and then passes through a low-pass filter 9, which passes only signals having a predetermined amplitude or more. It passes through the slicer 10 having the input/output characteristics as shown and is added to the synthesizer 11.
Ru. With this configuration, the output video signal will have the following characteristics depending on the amplitude and frequency components of the input video signal. First, for the high-frequency portion of the input video signal, the output of the C-shaped comb filter is blocked by the low-pass filter 9, so that it is not applied to the synthesizer 11, and only the output of the IdY-type filter is output. In other words, the high frequency portion of the video signal always passes through the UY-shaped comb filter. Second
is within the passband (low frequency) of the low-pass filter and the C-shaped comb filter output is below a predetermined level. In this case, the C-shaped filter output is cut off by the slicer 1ovc, and the output is The output is also a Y-shaped comb filter. In other words, the correlation between the input video signal is strong in the low frequency range (
Even when the C-shaped comb filter output is small, the signal passes through the Y-shaped comb filter. FIG. 3 shows a case where the C-shaped comb output is at a predetermined level or higher in the low frequency band.

Since the output of the Y-shaped comb filter is also added to the synthesizer 11, the Y-shaped and C-shaped comb filters are added, resulting in an all-pass characteristic without a comb characteristic. In other words, when the correlation of the input video signal is weak in the low frequency range (the C-shaped comb filter output is a dog), the comb characteristic does not exist and the input video signal passes the entire range. A Y-shaped comb filter is applied to the high-frequency and low-frequency areas of the video signal where the line correlation is strong to reduce noise, and a comb filter is applied to the low-frequency areas where the line correlation is weak. This prevents vertical resolution from deteriorating.

The difference from the conventional line noise canceller is that the Y-shaped comb filter is always effective in the high frequency range, but there is no practical problem because the high frequency range is not so effective in visualizing vertical resolution degradation.

Next, the effect of such a configuration, that is, the ability to reduce the circuit scale, will be explained.

In the conventional example described in FIGS. 1 and 2, it is necessary to select the sampling frequency fs to be at least twice the upper limit frequency of the input video signal so that no aliasing distortion spectrum due to the sample exists within the input video signal band. If f3 is selected to be twice or less, the original signal cannot be reproduced due to distortion due to folded spectrum. In the example, the upper limit frequency is 3 ridges and fsk
In this case, if the resolution of AD converter 1 is 7 bits, the pit rate is 49 MB.
/sec or more is required.

In this case, if the configuration explained in FIGS. 3 and 4 is used, the bit rate i can be significantly reduced.

The sampling frequency fs' is selected to be less than twice the upper limit frequency of the input video signal, and as a result, the aliasing distortion that occurs is removed using the Y-shaped comb characteristic, using the so-called Sapner uneven sampling technique. Yes Figure 6 is a spectrum diagram to explain this point.
The band of the input video signal JB is assumed to be 3 times as in Fig. 2, and the 0 sample frequency fs''k is less than twice the upper limit frequency of the input video signal, f3' =%fH(2n-1)... ...(1)f
H: Horizontal scanning frequency n: If a frequency is selected that satisfies the relationship of integer, the lower side wave fL of fS'
', upper side wave fU', and sideband wave 2f that is an integer multiple of fs'
L' appears as shown in (a) 0 where fs/
Components of 6 or more stones including
The component remaining within and on the high frequency side thereof is fL', which is the aliasing distortion component. To restore the original video signal, this aliasing distortion component, that is, the sampling frequency fs
'lower wave component L' k must be removed. However, the video signal fB and the folded spectrum fL/'
When zoomed in, the video signal of the base node is distributed like a solid line on integral multiples of fH as shown in the figure (in the same figure), whereas in 9', the sample frequency fs' satisfies equation (1). As long as you are satisfied! It is distributed on odd multiples of 113AfH as shown by the broken line. In this way, the baseband video signal and the aliased spectrum due to the sample have a relationship of % line offset, and in this case, aliased spectra exist at 2M or more.3 Therefore, by removing unnecessary aliasing spectra, only the necessary video signals can be generated. 2 to take out
It is sufficient to pass one or more bands through a Y-shaped comb filter.

As already explained with reference to the operation in FIG. 3, in the band above the cutoff frequency of the low-pass filter 9, only the optical power of the adder 8 appears in the output video signal, so the signal is passed through a Y-shaped comb filter. In this case, if the cutoff frequency '1211 of the low-pass filter 9 is selected, the aliasing spectrum existing on the higher frequency side will be removed.

As is clear from the above explanation, in the embodiment shown in FIG. 3, it is possible to set the sampling frequency Js' to less than twice the upper limit frequency of the input video signal, for example, fB' =
If the resolution of the AD converter is set to 7 bits, the bit rate can be set to 36 MB/sec, which is approximately 70% lower than the conventional example shown in Figure 1. It shows.

To reduce the bit rate, it is possible to reduce the number of quantization bits in the AD converter by applying predictive coding, but in order to reduce the bit rate to 70%, it is necessary to reduce the bit rate from 7 bits to 5 bits. Considering this fact, it can be understood how effective this configuration is in reducing the bit rate. Since circuit scale, power consumption, etc. are generally proportional to bit rate, it can be said that the configuration according to the present invention can significantly reduce the circuit scale.

Nao, in the embodiment shown in Figure 3, the video signal is digitized to perform the line noise canceller operation, but it is of course possible to process it using sampled analog amounts without digitizing it, and in that case, the AD converter 1, DA No need for 7 converters, 1cccD instead of 2 1H memories
The adder a, the subtractor 53, the low-pass filter 9, the slicer 1o, and the synthesizer 11 may be changed to elements that can process sample value data using semiconductor delay elements such as (charge transfer devices). In this case as well, the sampling frequency can be set to twice the upper limit frequency of the input video signal or less, which is effective in reducing the number of CCD stages and power consumption.

FIG. 6 is a configuration diagram showing another embodiment of the present invention, and FIG.
Blocks that operate in the same way as in the illustrated embodiment are designated by the same numbers. The difference from FIG. 3 is that the operation of the synthesizer 12 is the bypass filter 1 of the output of the subtracter 3 from the output of the AD converter 1.
3, low pass filter 9, limiter 1
4 is configured to subtract the signal that has passed through. No-i pass filter 13 and low pass filter 9
The cutoff frequency of the input video signal is set to a frequency near the boundary where aliasing distortion due to sub-Nyquist sampling exists within the band of the input video signal. When , the amplitude is limited to a constant value.

The operation of the configuration shown in FIG. 6 is as follows. The high-frequency part of the input video signal is output from the C-shaped comb filter, which is the output of the subtracter 3, and is converted from the input signal by the synthesizer 12 through the bypass filter 13. Since it is subtracted, the output id is the one that has passed through the Y-type comb filter. In other words, the high-frequency part of the video signal is always the low-frequency part of the 0-input video signal that removes the aliasing spectrum caused by the Y-shaped comb filter resubsampling, and when the line correlation is strong, the output of the low-pass filter 9 is small ( Since the correlation is strong, the C-shaped comb output is small), it passes through the limiter 14 as it is and is subtracted from the input signal by the synthesizer 12, so it still exhibits a Y-shaped comb characteristic. In other words, when the line correlation is strong in the low frequency region of the input video signal, the input video signal passes through the Y-shaped comb filter, improving the S/N ratio and performing a line noise canceling operation. When the correlation between the input video signals is weak, the output of the low-pass filter 9 becomes a dog, so the output of the C-shaped comb filter is a limiter and the amplitude limiter does not reach the synthesizer 12, so the input video signal is output. It appears as is and does not exhibit comb-shaped characteristics. Therefore, almost no vertical resolution deterioration occurs.

It is clear that the embodiment shown in FIG. 6 can also achieve the same effect as shown in FIG. 3, and the scale of the bit rate drop circuit can be reduced by lowering the sampling frequency.

Effects of the Invention As is clear from the above explanation, the present invention improves the SIN by the VTR line noise canceller at fs = "AfH (2n-
1) It is configured to be able to sample at different frequencies, and it overcomes many of the drawbacks of line noise cancelers that use conventional glass delay lines, and also significantly reduces the scale of the 1H delay circuit and peripheral processing circuits. Can be reduced. It satisfies the essential conditions for IC conversion, such as low power consumption, and is suitable for VTRs.
This is highly effective in reducing the size and cost of signal processing systems.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a conventional example, FIG. 2 is a spectrum distribution diagram thereof, FIG. 3 is a block diagram showing an implementation of the present invention, and FIGS. 4(a), (b), ( C), Figure 6 (m
, (b) are a characteristic diagram and a spectrum distribution diagram for explaining its operation, and FIG. 6 is a diagram showing the main part configuration of another embodiment of the present invention. 1...AD-ff converter, 2...1
H memo 1, 3...subtractor, 7...D
A Konno Kuichita, B... Ka I'l Calculator, 9...
・・・・Roono Kusufili, 1O・・・・Slicer, 11.12・・・・Synthesizer, 13・・・・Nohei no Kusufili, 14・・・・・・Limiter. Name of agent: Patent attorney Toshio Nakao +1 11 people Figure 2 f. r/jJ Wave Comparison rPHz'J Figure 3 Figure 4

Claims (1)

[Claims] 0) Less than twice the upper limit frequency of the input video signal. A sampling means for sampling at a frequency of f5 = '3 (fH (2n 1)), a comb filter means for forming a comb characteristic using a delay circuit that delays the sampled information by one horizontal scanning period, and a comb filter means for forming a comb characteristic using a delay circuit that delays the sampled information by one horizontal scanning period. Processing means for manipulating the amplitude and frequency characteristics of the output.
□1, ika, □3° multi-band il'1l-Y-shaped comb characteristic, and a video characterized in that it is configured to have Y-shaped comb filter characteristic only when line correlation is strong in some bands. Signal processing device. (2) fs at less than twice the upper limit frequency of the input video signal
AD(f
H: horizontal scanning frequency) (・: integer) AD conversion means to convert, memo IJ' (i- comb filter means to form a comb characteristic using this) to delay the AD converted digital signal by one horizontal scanning period. It is equipped with a processing means for limiting the band and amplitude of the output of the comb filter means, a synthesizing means for synthesizing a plurality of signals, and a DA converting means for converting the output of the synthesizing means into an analog signal. A video signal processing device characterized in that a part of a signal band has a Y-shaped comb characteristic, and a part of the band has a Y-shaped comb filter characteristic only when line correlation is strong. (3) <L The Y type filter means obtains a Y signal by adding the input video signal and a signal obtained by delaying the input video signal by one horizontal scanning period.
The apparatus is characterized in that it has a Y-shaped comb filter and a C-shaped comb filter obtained by subtraction, and is configured such that the output of the Y-shaped comb filter and the C-shaped comb filter that has passed through the processing means are synthesized by a signal synthesizing means. A video signal processing device according to claim 1 or 2. (4) The processing means is configured to include a low-pass filter that allows only the low frequency portion of the input video signal to pass through with respect to the output of the C-shaped comb filter, and a level slicer that allows the signal having a predetermined amplitude or more to pass through the output of the low-pass filter. A video signal processing device (5) according to claim 1 or 2, characterized in that the <, t, type filter means:
A C-shaped comb filter is obtained by subtracting the input video signal and a signal delayed by one horizontal scanning period from the input video signal, and the output of the C-shaped comb filter is passed through a bypass filter, and a signal having a predetermined amplitude or less is passed through a low-pass filter. The video signal processing device according to claim 1 or 2, characterized in that the synthesizing means is configured to subtract the output of the C-shaped comb filter that has passed through a limiter that passes from the input video signal. .