JPH05236431A - Noise reducer and muse decoder provided with the noise reducer - Google Patents

Abstract

PURPOSE:To reduce a transmission noise component effect in the case of a moving image processing by subtracting a MUSE signal from the two-dimensional interpolation processing signal of the MUSE signal, extracting a transmission noise component by a frequency separating filter, multiplying a coefficient and subtracting that component from the processing signal. CONSTITUTION:Concerning the MUSE input signal, the two-dimensional interpolation processing signal is outputted through a two-dimensional interpolation processing circuit 1 for moving image. On the other hand, the noise signal component guided from a frequency separating filter 3 is guided to a subtracter 5 and subtracted from the two-dimensional interpolation processing signal from the circuit 1 after a suitable coefficient is multiplied by a coefficient equipment 4, and video signals reducing noise are outputted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a MUSE signal decoding apparatus which decodes a MUSE signal, which is a high-definition transmission system, into an original high-definition signal, or converts the MUSE signal into another signal other than the high-definition signal, such as an NTSC signal. The present invention relates to a noise reducer used in a MUSE signal converter and the like.

[0002]

2. Description of the Related Art Since a high-definition signal has a large amount of information as compared with a conventional television signal, the band-compressed M
It is converted into a USE signal and transmitted mainly using broadcasting satellites. This MUSE signal is processed differently in the still image area and the moving image area, and by using motion detection on the receiving side, the still image processing and the moving image processed signal are adaptively mixed and returned to the high-definition signal. ..

When the CN ratio at the time of receiving the MUSE signal is low, in the still image processing, the noise component of the video output is not visually noticeable due to the widening of the transmission noise by the inter-frame interpolation and the noise reduction by the frame memory. is doing.

On the other hand, in the moving image processing, as shown in FIG. 8, a signal is reproduced by applying a two-dimensional filter having 7 taps in the horizontal direction and 5 taps in the vertical direction to sample points (marked with a circle in the figure). For this reason, the noise at the sample point represented by the black circle in FIG. 8 does not stop at that position but affects the pixel at the center of the two-dimensional filter. That is, the noise component at a certain sample point affects the 7 × 5 pixel range.

Therefore, the noise component of the moving image processing system signal is 2
Since it spreads dimensionally, it increases slightly and is visually deteriorated from the MUSE input signal. Therefore, when the CN ratio is low, a difference occurs in the noise component and amount between the still image area and the moving image area.

Due to the nature of triangular noise during FM transmission, the peak of transmission noise in the MUSE signal is 8 MH.
By utilizing the fact that z is used, the conventional MUSE signal decoding apparatus has a configuration as shown in FIG.

In FIG. 10, a signal subjected to interpolation processing by the moving picture two-dimensional interpolation processing circuit 1 including a two-dimensional filter (hereinafter referred to as a two-dimensional interpolation processing signal) is an 8 MHz bandpass filter ( Hereinafter, this will be referred to as “BPF”) 8 and the transmission noise component near 8 MHz will be derived.

This transmission noise component is multiplied by an appropriate coefficient in the coefficient unit 9 at the next stage, and then subtracted from the two-dimensional interpolation processed signal by the subtractor 10. Thus, conventional MUS
In the E signal decoding device, in order to reduce the difference in noise between the still image region and the moving image region, the 8 MHz notch filter is provided only in the moving image processing system to reduce the noise.

[0009]

However, in the above-mentioned conventional apparatus, the notch filter also applies attenuation to the video signal component, so that there is a drawback in that the attenuation amount of the notch filter cannot be increased, and it is considered that the still image area is lost. There is a problem that it is difficult to sufficiently reduce the difference in noise in the moving image area. The present invention provides a noise reducer for a moving image processing system that prevents influence on a video signal by detecting transmission noise.

[0010]

In order to solve the above-mentioned problems, the present invention solves the above problems by using a moving image two-dimensional interpolation processing means for deriving a two-dimensional filter output for reproducing a moving image of a MUSE signal, and the moving image two-dimensional interpolation processing means. First subtraction means for taking a difference between the two-dimensional filter output derived from the insertion processing means and the MUSE input signal,
At least 8 for the difference output from the first subtraction means.
Filter means for detecting a transmission noise component by passing through a filter that passes around MHz, coefficient multiplication means for multiplying the transmission noise component derived by the filter means by a coefficient, and coefficient multiplied by the coefficient multiplication means And a second subtraction means for reducing the transmission noise by subtracting the transmission noise component from the output of the two-dimensional filter.

Further, the coefficient multiplying means applies a coefficient only to a transmission noise component having a small amplitude among the transmission noise components that have passed through the filter means, thereby erroneously detecting a transmission noise component due to a folding signal included in the MUSE input signal. To prevent this.

Further, the coefficient multiplying means has a function of changing the coefficient value according to the reception CN ratio, and it is possible to set different coefficient values in the color difference signal period and the luminance signal period of the MUSE signal. The noise reduction amount in both periods is changed.

Furthermore, the above-mentioned filter means is provided with a pass band near DC as well as around 8 MHz so as to have both functions of improving vertical resolution by low-frequency replacement and reducing noise. Further, the noise reducer described above is applied to a MUSE decoder that omits the still image processing and simply processes the MUSE signal only by moving image processing.

[0014]

According to the above construction, the difference between the output of the two-dimensional filter for reproducing the moving image of the MUSE signal and the MUSE signal is calculated,
Transmission noise is detected by a filter that passes at least around 8 MHz of this differential output, and this transmission noise is multiplied by a coefficient and subtracted from the output of the two-dimensional filter, so only the transmission noise component in moving image processing is effectively reduced. Therefore, it is possible to reduce the difference in noise between the still image region and the moving image region without attenuating the moving image signal component.

Further, the coefficient multiplying means can also apply the coefficient only to the transmission noise component having a small amplitude, so that erroneous detection of the transmission noise component due to the aliasing signal included in the MUSE signal can be prevented. You can

Further, the coefficient multiplying means can change the coefficient value by which the transmission noise is multiplied by the reception CN ratio, and can also multiply different coefficient values in the color difference signal period and the luminance signal period, so that the color difference and the luminance signal period can be changed. The noise reduction amount can be set independently and optimally.

Further, the filter means has a pass band of 8 MHz.
Since it can be provided near the direct current as well as near the direct current, it is possible to improve the vertical resolution by low-frequency replacement and simultaneously have both functions of noise reduction. Furthermore,
When used in a simple MUSE decoder, the noise reduction effect is particularly remarkable because this type of decoder always performs moving image processing.

[0018]

Embodiments of the present invention will be described below with reference to the drawings. In the present invention, since the noise component is detected, M
The difference between the USE input signal and the two-dimensional interpolation processed signal by the two-dimensional filter is calculated. This state is shown in FIG. In this figure, the direction of the transmission noise is shown as a positive direction with respect to the signal, but the same can be considered in the negative direction.

When the MUSE input signal does not include the aliasing signal, the difference between the MUSE input signal and the two-dimensional interpolation processed signal is 8 MHz as shown in FIG.
Considering the vicinity of z, this can be considered as an increase in noise due to the interpolation processing. By slightly amplifying this difference and subtracting it from the two-dimensional interpolation processing signal, noise can be reduced without affecting the video signal.

Here, since the reproduction band of the moving image system in the MUSE signal is substantially about 12 MHz, the signal is not folded back to the direct current. That is, since the fold-back is not included in the image portion where the noise is visually noticeable, the component near 8 MHz of the difference signal can be determined and detected as the transmission noise, and the noise can be effectively reduced.

FIG. 2 shows the case where the MUSE input signal includes the folding component of the video signal. In this case, the component near 8 MHz of the differential signal cannot be generally determined to be a transmission noise component or a folding component. However, the amplitude of the increased noise due to the interpolation process is usually small because it is due to the interpolation error. Therefore, when the amplitude value of the difference signal is small, it is determined that the noise is transmission noise and noise is reduced, and when the amplitude value is large, the noise reduction is given as a folding component.

Since the edge component of the image contains the aliasing component, it has little visual impact without noise reduction, and since the image signal is output as it is, there is no problem in sharpness. Here, the threshold value of the differential signal amplitude value, which determines whether or not noise reduction is performed, is determined while observing the influence of the folding back of the edge portion.

FIG. 3 shows a basic circuit configuration for performing this processing. In FIG. 3, reference numeral 1 is a moving image two-dimensional interpolation processing circuit for performing interpolation processing on a MUSE input signal by a two-dimensional filter, and reference numeral 2 is a two-dimensional interpolation processing derived from the moving image two-dimensional interpolation processing circuit 1. A subtractor that subtracts the MUSE input signal from the processed signal to derive a difference signal between the two signals, and the difference signal derived from the subtractor 2 derives a noise signal component near 8 MHz through a frequency separation filter 3. To do.

The noise signal component derived from the frequency separation filter 3 is multiplied by an appropriate coefficient in the coefficient unit 4 and then guided to the subtracter 5 to be output from the moving image two-dimensional interpolation processing circuit 1. A noise-reduced video signal is derived by subtraction from the two-dimensional interpolation processing signal.

In this case, when the value of the coefficient unit 4 is increased, the noise reduction effect is enhanced, and when it is 0, no noise reduction is performed. Noise can be effectively reduced by changing this value according to the reception CN ratio. Further, since the color difference signal (C) period is extended in time, noise is also extended in the same manner,
Since it is visually noticeable, it is more effective to make the coefficient value larger in the color difference signal period than in the luminance signal (Y) period also in the CN ratio.

The characteristic of the frequency separation filter 3 in FIG. 3 is that shown in FIG. 4 because the transmission noise becomes a triangular noise having a peak at 8 MHz.
It may be a characteristic as shown in. In the filter having the characteristics shown in FIG. 5, the reduction unit of the two-dimensional interpolation processed signal can be replaced with the MUSE input signal, so that the vertical resolution is improved.

In the MUSE signal decoding apparatus, this low-frequency replacement processing is usually provided after adaptively mixing the signals of the still image processing and the moving image processing. But MUSE
This is an effective means in a device that converts a signal mainly into two-dimensional interpolation processing into another signal because noise reduction and low-frequency replacement can be performed at the same time.

FIG. 6 shows a characteristic example of the coefficient unit 4. As described above, the coefficient is multiplied and output only when the amplitude of the coefficient input is small. Such a coefficient unit can be easily realized by using a ROM. FIG. 7 shows a ROM in the coefficient unit 4
An example of the circuit when using is shown.

In FIG. 7, the parts corresponding to those in FIG. 3 are designated by the same reference numerals and the description thereof will be omitted. In FIG. 7, 6 is an MU
A time adjustment register for compensating the time delay of the signal generated by the SE input signal being subjected to the two-dimensional filter processing in the two-dimensional interpolation processing circuit 1 for moving images, and 7 is a register near 8 MHz extracted by the frequency separation filter 3. The coefficient multiplier 7 is a coefficient multiplier for multiplying transmission noise by a coefficient. The coefficient multiplier 7 is composed of a memory such as a ROM for setting an address according to an input signal level, CN information, a Y / C discrimination signal, etc., and appropriately selecting the coefficient value. To do.

Reference numeral 8 is a time adjusting register for delaying the two-dimensional interpolation processing signal from the moving picture two-dimensional interpolation processing circuit 1 by the time delay of the transmission noise by the frequency separation filter 3 and the coefficient unit 7.

Therefore, the MU supplied from the input terminal T1
The SE input signal is supplied to the subtracter 2 and the time adjustment register 8 as a two-dimensional interpolation processing signal that has been subjected to two-dimensional interpolation processing by the moving picture two-dimensional interpolation processing circuit 1.

The above-mentioned two-dimensional interpolation processing signal is subjected to a two-dimensional filter processing in the above-mentioned two-dimensional interpolation processing circuit 1 or the like, and MU.
There is a time delay with respect to the SE input signal. This time delay is adjusted by passing through the time adjusting register 6, and the two-dimensional interpolation processing signal and M input to the subtractor 2 are adjusted.
The absolute phase of the pixel is matched with the USE input signal.

The subtractor 2 takes the difference between the above two signals whose pixel positions are matched. This difference signal is at least 8 MH
The transmission noise is extracted by applying the frequency separation filter 3 that passes the vicinity of z, and the coefficient is multiplied by the coefficient in the coefficient multiplier 7. At this time, in addition to the change in the coefficient value depending on the input level, C
A change in the coefficient value depending on the N information, the Y period, and the C period can be performed by setting the address of the ROM.

The transmission noise component adjusted by the coefficient unit 7 to an appropriate level by multiplying the coefficient supplied from the ROM in accordance with the input level, CN information, Y / C discrimination signal, etc.
The subtractor 5 subtracts the two-dimensional interpolation signal. in this case,
Since the noise component is delayed by the amount of processing by the frequency separation filter 3 and the coefficient unit 7, the time adjustment register 8 is set to 2
It is necessary to delay the dimensionally interpolated signal so that the pixel positions are aligned before the subtraction. As a result, the output signal of the subtractor 5 becomes a two-dimensional interpolation processed signal in which only noise is reduced.

When the above noise reducer is used in a conversion device for converting another signal such as an NTSC signal mainly for two-dimensional interpolation, the characteristics of the frequency separation filter 3 are set as shown in FIG. 5, for example. As described above, the noise reduction device that doubles as the low-frequency replacement can be obtained. FIG. 9 shows a configuration example in which the noise reducer described above is adapted to simply decode a MUSE signal into a high-definition signal only by moving image processing. The MUSE signal is input from the terminal 11, passes through the low-pass filter 12 having a roll-off characteristic at 8.1 MHz so as to satisfy the resampling condition, and is converted into a digital signal by the A / D converter 13. After this signal is subjected to synchronization detection, de-emphasis, etc. by the input processing 14, the moving picture 2
Video reproduction is performed only by the dimension interpolation processing 1. The noise reducer 1 outputs the output of the two-dimensional interpolation processing 1 for moving images and the output of the input processing 14
Input to 5 to reduce transmission noise. The principle and operation are as described above. After reducing the transmission noise, the time axis of the luminance and chrominance signals is made coincident by the TCI decoding 16, and the MUSE signal whose video signal period has been compressed to 11/12 with respect to the original sample is 12 / 11 Expanded and returned to the original time axis. The digital video signal whose time axis has returned to the original is output to the terminal 19 as a high-definition video signal of an analog signal by the D / A conversion 18. This constitutes a MUSE decoder having a noise reducer.

[0036]

Since the present invention has the above-mentioned structure,
Two-dimensional filter output and MU for video playback of MUSE signal
At least 8M of this difference output is obtained by taking the difference of SE signal
Transmission noise is detected by a filter that passes around Hz, and the transmission noise is multiplied by a coefficient and subtracted from the output of the two-dimensional filter, so that only the transmission noise component in the moving image processing can be effectively reduced. It is possible to reduce the difference in noise between the still image area and the moving image area at the time of reception at a low CN ratio without attenuating the moving image signal component. Further, since the coefficient multiplying means can apply the coefficient only to the transmission noise component having a small amplitude, it is possible to prevent the transmission noise component from being erroneously detected by the folding signal included in the MUSE signal. Further, the coefficient multiplying unit can change the coefficient value by which the transmission noise is multiplied by the reception CN ratio, and can also multiply different coefficient values in the color difference signal period and the luminance signal period, so that the noise reduction amount in the color difference and luminance signal period can be reduced. Can be set independently and optimally. Furthermore, since the filter means can be provided with a pass band near DC as well as around 8 MHz, it is possible to improve both vertical resolution by low-frequency replacement and simultaneously have both noise reducing functions.
Furthermore, the still image processing is omitted, and only the video processing is performed to the MUS.
When used for a simple MUSE decoder that simply processes E signals, this type of decoder is particularly effective because it always performs moving image processing.

[Brief description of drawings]

FIG. 1 is a diagram for explaining the operation principle of the present invention.

FIG. 2 is a diagram for explaining the operation principle of the present invention.

FIG. 3 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 4 is a characteristic diagram of a filter used in the present invention.

FIG. 5 is a characteristic diagram of another filter used in the present invention.

FIG. 6 is a characteristic diagram of a coefficient unit used in the present invention.

FIG. 7 is a block diagram showing a circuit configuration of an embodiment of the present invention.

FIG. 8 is a diagram for explaining the influence of noise on another pixel position by the two-dimensional filter.

FIG. 9 is a block diagram showing the configuration of another embodiment of the present invention.

FIG. 10 is a configuration diagram of a conventional example.

[Explanation of symbols]

 1 Two-dimensional interpolation processing circuit for moving image 2, 5 Subtractor 3 Frequency separation filter 4, 7 Coefficient unit 8 8MHz BPF (bandpass filter) 9 Coefficient unit 10 Subtractor 11, 19 Terminal 12 Low pass filter 13 A / D conversion 14 Input processing 15 Noise reducer 16 TCI decoding 17 Time axis expansion 18 D / A conversion

Claims (5)

[Claims] 1. A moving image two-dimensional interpolation processing means for deriving a two-dimensional filter output for reproducing a moving image of a MUSE signal, a two-dimensional filter output derived from the moving image two-dimensional interpolation processing means, and a MUSE input signal. Of the difference output from the first subtraction means, a filter means for detecting a transmission noise component by passing a filter passing at least around 8 MHz to the difference output from the first subtraction means, and a derivation from the filter means. Coefficient multiplying means for multiplying the transmission noise component by a coefficient, and second subtraction means for reducing the transmission noise by subtracting the transmission noise component multiplied by the coefficient derived from the coefficient multiplication means from the output of the two-dimensional filter. A noise reducer characterized by being provided. 2. The noise reducer according to claim 1, wherein the coefficient multiplying means applies a coefficient only to a transmission noise component having a small amplitude among the transmission noise components that have passed through the filter means, and is included in the MUSE input signal. A noise reducer characterized by preventing erroneous detection of a transmission noise component due to a folded signal. 3. The noise reducer according to claim 1, wherein the coefficient multiplying means has a function of changing the coefficient value according to the reception CN ratio, and the MUSE signal is different in color difference signal period and luminance signal period. A noise reducer characterized in that a coefficient value can be set so that the noise reduction amount in both the above-mentioned periods is changed. 4. The noise reducer according to claim 1, wherein the filter means has a pass band near DC as well as near 8 MHz to improve both vertical resolution and noise reduction by low-frequency replacement. A noise reducer characterized by being provided with. 5. A low-pass filter having a characteristic of being rolled off at 8.1 MHz, an A / D conversion unit, an input processing unit having at least a function of de-emphasis and synchronization detection, and a MUSE signal moving image reproducing unit. At least with respect to the two-dimensional filter, the first subtraction means for taking the difference between the two-dimensional filter output derived from the moving picture two-dimensional interpolation processing means and the MUSE input signal, and the difference output from the first subtraction means. A filter means for detecting a transmission noise component by passing through a filter that passes around 8 MHz, a coefficient multiplication means for multiplying a transmission noise component derived by the filter means by a coefficient, and a coefficient derived by the coefficient multiplication means are multiplied. A noise reducer provided with second subtraction means for reducing transmission noise by subtracting a transmission noise component from the output of the two-dimensional filter; MUS with a code portion, and time-base decompression unit, and the D / A converter, a noise reducer, characterized in that it consists
E decoder.