JPS6040234B2 - video signal processing equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a noise reduction scheme that is particularly useful for reducing noise in highly redundant and repetitive electrical signals such as video signals, and to a signal processing apparatus for such a scheme.

The quality of the video signal can be improved by averaging two or more fields. This has the effect of reinforcing the desired signal content against unwanted noise that varies randomly from field to field. However, this averaging process inevitably introduces some changes in the signal, especially those related to the reduction in detail caused by unsharp areas when there is movement in the image.
An object of the present invention is to provide a method that does not have this drawback. According to the invention, there is provided a device for generating a plurality of replicated signals obtained from an input signal and having different signal delays, and a variable coupling device for variably combining these replicated signals to generate an output signal. in response to the difference or group of differences between these replicated signals such that the output signal consists of a combination of all of these replicated signals when the difference or each difference is small, and when any of the differences is not small. and a signal processing device which sometimes comprises a combination of one or more of the replicated signals selected to reduce the difference or each difference with respect to the selected group of replicated signals. provided. This signal processing device can operate in either encoding mode or decoding mode.

In decoding mode, the combining action is additive, whereby an averaging effect is usually achieved. This effect disappears or decreases when either of the differences is not small. Thus, undesired averaging, which has the effect of smearing the changing signal, is eliminated. In the encoding mode, the combining operation subtracts one or more of the replicated signals from the remaining replicated signals to perform a processing operation complementary to that provided by the decoding mode. The mode of the circuit is determined by the coefficients and relative polarities of the coupled replicated signals. Although it is possible to use the decoding processor alone to provide a noise reduction effect that makes changes in the signal itself relatively small, it is also possible to use both the encoding and decoding circuits to form a completely complementary noise reduction scheme. It is preferable to use

In the complementary scheme, the signal is first processed by an encoding circuit. The encoded signal is supplied via a transmission channel to a decoding circuit in real time or via a recording and reproducing device. The term "information channel" is used generally to refer to the means for conveying encoded signals to decoding circuitry. When a recording/reproducing device is used,
One circuit can be used for both encoding and decoding, in which case suitable switching means are used to switch between encoding and decoding modes. One of the different signal delays can be zero delay, in other words, one of the replicated signals consists of the direct, undelayed signal derived from the input signal.

In addition to being economical, this leads to a circuit geometry that provides completely complementary encoding and decoding characteristics. The distinction between when the signal difference is small and when it is not small can be arbitrarily determined to suit a given application of the invention, but typically this distinction is made within 1% of the peak signal level. % to 10%.

In the simplest case just two duplicate signals are used, and when the difference between them is "small" they are both combined.

When this difference is "not small", only one of these signals (which can always be the same predetermined one of the two signals) contributes to the output signal.
The change between these two states can be steep, in which case the two states are determined by a single threshold value applied to the difference between the signals. For example, if the threshold is about 1% of the peak signal level, a difference of up to 1% is considered "small" and a difference of more than 1% is considered "not small." However, the change can also be gradual, in which case it is considered "small" as long as the difference does not exceed a low threshold of 1%, and both duplicate signals contribute to the output signal. If the difference exceeds 1%, the contribution of the second signal is progressively reduced, and for example 10
When the difference is considered "not small" at a relatively high threshold of %, the second signal no longer contributes to the output signal. Between 1% and 10%, the difference is "intermediate"
, that is, between "small" and "not small."

The second signal therefore makes an intermediate contribution to the output signal. When more than two replicate signals are used, various possibilities exist for selecting one or more of these signals when any of the differences becomes small.

The simplest possibility is to select only one predetermined replica signal, eg the direct signal, whenever any of the differences are no longer small. However, from a more general point of view, when there are not small differences, the replicated signals can be divided into two classes, called "valid signals" and "invalid signals". The differences between valid signals are all small, but the differences between invalid and valid signals are not. The decision as to which is a valid signal and which is an invalid signal is that the valid signal is always one of the replicated signals.
This can be done, for example, by specifying that the signal contains a direct signal, or by specifying that the valid signal is a signal of relatively large numerical value. This separation of the replicated signals may consist of any combination of one or more of the output signals. An example is given below where this is the case, in which a predetermined one of these signals, the direct signal, is always the valid signal.

Other replicate signal groups are individually compared to this. As long as the difference between any one of said other replicated signals and said one signal is small, this other signal is a valid signal and is included in the combination forming the output signal. If the difference is not small, the other signal is invalid and excluded from the combination forming the output signal. Again, abrupt or gradual transitions are used.

In the latter case, the combination consists of all duplicate signals when all differences are small and all signals are valid, and only valid signals when some of the differences are not small and some signals are invalid. Consisting of When there is an intermediate difference, there are several signals that are intermediately valid and invalid. Such a signal makes an intermediate contribution to the output signal. The total number of replicated signals can be expressed as an integer N greater than one.

Preferably, the variable combining device is configured to always combine N signals regardless of whether any signal is invalid. If any signal is invalid, the variable combiner replaces it with a duplicate of one valid signal. Since M signals are always combined, the level of the output signal is unaffected by changes in the number of signals that are valid. This avoids problems that would otherwise occur if the invalid signal were simply separated from the combination. a main signal path including a first coupling device;
an auxiliary signal path to the first coupling device in response to a delayed signal at the input or output of the first coupling device and an undelayed signal at the input or output of the first coupling device; an auxiliary signal path including a second variable coupling device for providing a signal, the auxiliary signal path signal being combined with the signal of the main signal path, the auxiliary signal path signal in parentheses being the delayed signal; It is preferable that when the difference between the signal and the undelayed signal is small, the signal is substantially the delayed signal, and when the difference between the parentheses is not small, the signal is substantially the signal that is not delayed.

In the decoding mode, the auxiliary signal increases the level of the main signal path signal, and in the encoding mode, the auxiliary signal path signal decreases the level of the main signal path signal.

In operation, as long as the difference is small, the delayed signal is combined with the main signal path signal to provide an averaging, noise reduction effect in the case of a decoder. If the difference is not small, the direct signal is used, which eliminates the blurring.
The reason for this is that the signals combined by the first combining device are now one and essentially the same signal. The output signal will be the same as the direct signal, which is the same as the input signal except that a level difference has been created.
In such a case, the noise reduction effect is removed, but in the case of video signals, for example, this is not a disadvantage, since the noise is substantially affected by movement in the image from which the signal is directly utilized. This is because it will be masked. The action of the variable coupling circuit occurs with a time constant that is short enough to remove the noise reduction effect only in the part of the image where movement is occurring, but maintain the noise reduction effect in the remainder of the string image. is preferred.

Preferably, the delay time t experienced by the delayed signal is equal to the period of the input signal when it is a repetitive redundant signal, such as a video signal.

For example, this time t may be equal to one or more scan line periods or one or more field or image periods in the case of video signals. However, the time t can also be short compared to the repetition period. In this case, the decoding circuit can eliminate noise at frequencies higher than the repetition frequency, and even noise at the highest frequency component of interest, by averaging. In order to improve the averaging effect of the decoding circuit, t
, public... Obtain a plurality of delayed signals having a delay of Nt (N' or an integer greater than 1), and these delayed signals cooperate to increase the level of the signal in the main signal path. It is preferable to do this.

Each delayed signal may be obtained by any suitable delay device, such as a delay line of a suitable delay time.

When the present circuit is used in conjunction with a recorder, whether the recorder records video signals magnetically, optically, or mechanically (i.e., by grooves, etc.), it may It is possible to obtain multiple signals with different delay times from the head at the same time. A video straight tube can also be used to provide the necessary delay. The transition from using a delayed signal to using a direct signal and vice versa may be abrupt (switching) or gradual.

In the complementary scheme, the level-increasing effect performed by the combination of the delayed signal and the input signal in the decoding circuit performs the above-mentioned averaging, but this is because each delayed signal is a code that performs a level reduction of the input signal. is complementary to the level reduction action in the encoding circuit, and therefore the overall effect of the encoding and decoding circuits is to restore the input signal to its original form (while the decoding circuit is reduce the noise introduced into the information channel by averaging the noise).

It is clear that the level reduction action within the encoding circuit must not completely cancel the input signal. The present invention will be described below with reference to the drawings.

For convenience, FIGS. 1-4 show a delay circuit 12, designated by the symbol Δ, for generating a delayed signal. From the foregoing it can be seen that in some cases it is not necessary to provide a separate group of delay devices, ie the delayed signal can be taken directly from the recording medium. To simplify the description of the invention, throughout the drawings, when signals are combined they are additively combined (i.e., mixed) (by blocks marked "10") and subtractively combined (by blocks marked "10"). A convention has been adopted that when necessary, the inverter 13 (block marked with "") is shown.

It will be appreciated that the same overall result may be achieved in various ways other than in the illustrated locations by using inverters or by using combination circuits such as differential amplifiers that subtract one signal from the other. All you just need is
Closed loops illustrated as inverting must remain inverting as a whole, non-inverting loops must remain non-inverting as a whole, and the result of signal combination may be additive or subtractive, as the case may be. The only thing that must be maintained is that In each of Figures 1-4, the encoder is shown at a and the decoder is shown at b.

The encoder supplies a signal to the decoder via an information channel represented by a disconnected connection line with the symbol N denoting the noise introduced into the information channel and averaged in the decoder. The apparatus shown in FIGS. 1 to 4 will now be explained.

Each decoder can itself be used independently to process the uncoded signal to perform noise reduction simply by averaging. However, a complete scheme using an encoder and decoder has the advantage that the decoder output is the same as the encoder input. A completely complementary operation is achieved in each of FIGS. 1-4, the decoder being of the same type as the encoder. In some cases it may be convenient to use different types of encoders and decoders, for example the encoder of FIG. 2a may be used with the decoder of FIG. 3b. Each encoder and decoder has a main signal path 10 and a coupling device 11 inside.

The auxiliary signal path has a direct input OR and a delayed input DL and provides the direct and delayed signals of the variable combiner 20, the output 21 of which is connected to the main signal path in the encoder. It is connected to a coupling device 11 via an inverter 3 to reduce the level of the signal, and in the decoder to a coupling device to increase the level of the main signal path signal. The difference between the four types shown in these four figures is that a direct signal and a delayed signal are obtained. If the symbols 1 and ◯ represent the input and output sides of the coupling device 11, respectively, these points are shown in the table below. In each case the decoder is complementary to the encoder. The variable coupling circuit 20 normally supplies the signal from the delayed input DL to the coupling device 11, but when the direct signal and the delayed signal differ by more than some small amount, the variable coupling circuit 20 supplies the signal from the direct input DR to the coupling device 11. It has such characteristics or is controlled to supply a signal to the coupling device 11 .

An example will be explained later. Certain limitations are placed on the gains of some of the feedforward and feedback loops of FIGS. 1-4.

First, the gain of the feedforward loop in the encoder is 1
Must be smaller. Otherwise, the output of the auxiliary signal path would cancel or even invert the signal of the main signal path. Second, the gain of the positive feedback loop must be less than 1 to avoid instability. Using a gain less than 1 in the decoder is disadvantageous because it does not provide the best averaging. Although it is possible to provide a gain of unity in the auxiliary signal path for the decoder of FIG. 3b, it is not possible to use such a gain in a complementary device. This is because the gain of the encoder of FIG. . Therefore,
A signal path gain of less than 1, for example a gain of no more than 0.9, preferably a gain of 0.8, is used to ensure stability, and a delay to improve averaging and therefore noise reduction. It is preferable to use two or more signals.

An example of an auxiliary signal path having a direct input DR and a delayed input DL is shown in FIG. 5, the outputs 21 of the auxiliary signal path being numbered correspondingly in each of FIGS. 1 to 4; Figure 5 can be replaced in any of these figures.

Three delays 16, 17 and 18 provide t, Z and milk delays.

There is no meaning in specifying the number of delay devices1, in other words,
Two, three, four or more delay devices can be provided. Each delay device has its own variable coupling device 20
The outputs of these combiners are summed by combiner 19 to produce an auxiliary signal path output at 21. Using the symbols appearing in FIGS. 3 and 5, it can be seen that in the encoder:

y = x - (y, 10 y2 10 y3) In the decoder, z = y + (y, 10 y2 10 y3) = x - (y, 10 y2 10 y3) 10 y, 10 y2 10 y3 d , it can be seen that the decoder and encoder are completely complementary and the decoder output z is the same as the encoder input x.

We have ignored the noise N introduced into the information channel; however, this is reduced by the averaging action of the decoder. This is because it is non-coherent in the four signals that are combined in the decoder, whereas y, y, , y2 and y3 are coherent and mutually reinforcing. Although not analyzed, it is easy to see that if the gains of the auxiliary signal paths in the encoder and decoder of FIG. 1, FIG. 2, or FIG. 3 are equal, then the encoder and decoder are still complementary. Let's be known.

The signal y in the information channel is at a lower level than the input signal x (and the output signal z).

This is known for example from FIGS. 3 and 5, where y,, y2 and y3 are all substantially the same (x is a video signal and the delay t is an integer multiple of the field period of this video signal, (preferably equal to an even multiple) and the gain of each loop is 0.8, then the equation in the encoder becomes: y=x-(y, 10 y2 10 y3) Also, y,=y2=y3=0. (for exact approximation) Therefore, y=x-3×0. Therefore, y=x/3.4 or approximately y=0. In the sphere decoder, z:y+(y, y2y3) =x/3.43x0. &/3.4 Nini X FIG. 6 shows a variable coupling device 20 in which a delayed signal and a direct signal are applied to the two ends of the potentiometer 22.

Auxiliary constant signal path output 21 is taken from the sliding arm of the potentiometer. A comparator 23 compares the signals at the two ends of the potentiometer to generate a difference or error signal, which drives the sliding arm via the servo represented by dashed line 23a. As long as the error signal (which may be subject to smoothing with a suitable time constant) remains below the predetermined threshold value, the sliding arm of the potentiometer remains at the left-hand end in the drawing and the auxiliary signal path output are exclusively delayed signals. Above the threshold, as the error signal increases, it moves the sliding arm further and further to the right, making the contribution of the direct signal larger and larger, and this direct signal eventually becomes completely overtaken by the delayed signal. replace. In drawing FIG. 6, electromechanical servo control of the potentiometer sliding arm has been previously assumed.

In practice, the above-mentioned effects can be obtained by purely electrical means involving known techniques, such as by using transistors or field-effect transistors as controlled resistors. This is illustrated in FIG. 7, where the error signal increases the conductivity of the field effect transistor when the resistance of the field effect transistor 28 in series with the resistor 24 drops substantially below the resistance of the resistor 24. Increasingly, the direct signal is replaced by a delayed signal. This change is gradual. It is not necessary to use a controlled variable coupling device.

FIG. 8 shows a simple passive circuit in which the delayed signal is normally coupled to output 21 via resistor 24. FIG. When the difference between the direct signal and the delayed signal exceeds the threshold of the two back-to-back diodes 25, they conduct and send the direct signal to output 21. This is because their resistance values are now much smaller than the resistance value of resistor 24. The transition from the delayed signal to the direct signal depends in this case on the characteristics of the diode. FIG. 9 shows a modification of FIG. 8 in which resistor 24 is replaced by a high-pass filter consisting of capacitor 26 and resistor 27. FIG.

The impedance of this filter is lower at high frequencies than at low frequencies, so the temptation to replace the direct signal with a delayed signal is less at low frequencies than at high frequencies. big. This has the advantage that the difference between these two signals is typically more significant at low frequencies than at high frequencies. A relatively high threshold value that acts at high frequencies is such that at such frequencies even when the direct signal replaces the delayed signal at relatively low frequencies, Allow noise reduction to occur. Capacitor 26 can be supplemented or replaced with a tuned circuit if an RF signal is present. Different bands of the input spectrum can generally be processed separately by separate combiners, each of which has a suitable frequency characteristic, and which are parallel to a given signal source with different delays. connected to. The outputs of these separate coupling devices are then coupled together. Although it is preferred to combine signals in the manner shown in FIGS. 1-4 using parallel arrangements or main and auxiliary signal paths, series mode circuits may also be used.

This is shown in FIG. 10 for the decoder only. The voltage divider circuit is composed of a resistor 30, a transistor 31, a resistor 32, a transistor 33, and a resistor 34. Transistor 33 acts as a current source controlled by the input signal, and resistor 32
contributes to the power to the output taken from the collector of the
The voltage across transistor 31 is combined with the voltage across resistor 32. Transistor 31 is used as a variable impedance controlled by an auxiliary signal path 35, which may have a simple form, for example as in FIG. 3b, or a relatively complex form as in FIG. and obtains both a direct input DR and a delayed input DL from the collector of transistor 31. Transistor 31 generates a signal that is related to the current flowing through the circuit and therefore to the input signal. In FIG. 5, the direct signal is always treated as the main signal which replaces any delayed signal when it differs from the direct signal.

However, one of the delayed signals can also be processed as the main signal as shown in the decoder of FIG. The four delays are respectively t and 2 in response to the input signal SI.
, 3, and 4 (where t can be a scan line period or a field period) also result in 2 to S5.

S3 is processed as the main signal and provided as an input to the combining device 11. The second input is normally the coupling device 1
SI + S2 + S4 + S5 power supplied by 9). However, the variable coupling device 20 has SI, S2, S4
Comparing each of S5 and S3 with S3, the difference with respect to S3 is a certain amount, and as the value exceeds SI, S2 gradually increases.
, S4 and S5 are replaced with S3. The invention can be used to reduce noise introduced by standard converters, in which case the encoder preceding the converter has delay lines and other parameters that match the standard of the input signal. and the decoder following the converter has delay lines and other parameters that match the standards of the output signal. Various other features which are also related to noise reduction of video signals and similar signals and which may be incorporated into the present invention, such as processing of signals including carrier components, between encoding and decoding processing units, etc. Samurai Kensho 48-37947 by the present applicant, which describes the switching technology, etc.
Please also refer to No.

[Brief explanation of the drawing]

1-4 illustrate four generalized embodiments of the present invention, and FIG. 5 shows an auxiliary example that may be used in any of FIGS. 1-4 and utilizes more than one delay. 6 shows the principle of the variable coupling device; FIGS. 7 to 9 show two practical circuits for the variable coupling device; FIG. 10 shows the implementation of the invention. FIG. 11 is a diagram showing another decoder implementing the present invention, in which 10 is a main signal path, 11 is a coupling device, 12 is a delay circuit, and 13 is a diagram showing a serial mode decoder. Inverter, 16
, 17, 18 are delay devices, 20 is a variable coupling device, D
R indicates a direct input, DL a delayed input, 23 a comparator, 28 a field effect transistor, and 35 an auxiliary signal path. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11

Claims (1)

[Claims] 1. A video signal processing device for noise reduction in an input video signal, including (a) a delay device 12 that delays a signal obtained from the input video signal, and (b) an adder circuit 11 that adds the delayed signal and the input video signal.
, and C. another circuit 20 that decreases the ratio of the delayed signal to the input video signal as the difference between the input video signal and the delayed signal increases;
The input of the delay device 12 is connected to the adder circuit 11.
A video signal processing device, characterized in that the video signal processing device is connected to an output of the video signal processing device.