JPH0685576B2 - TV signal processing circuit - Google Patents

Description

Description: FIELD OF THE INVENTION The present invention relates to a television receiver for subjecting a transmitted television signal to signal processing in the time axis direction by digital technology,
In particular, the present invention relates to a signal processing circuit capable of obtaining an optimum image quality even when moving images and when moving from still images to still images.

[Background of the Invention]

As a television receiver requiring signal processing in the time axis direction, there is, for example, a high-definition television receiver. As a high-definition television broadcast received by this high-definition television receiver by using a broadcasting satellite, there is a technique in which a high-definition television signal is band-compressed and transmitted. As an example of this band compression method and high-definition television receiver, NHK (NHK)
Giken Monthly Report, July 1984 by Ninomiya, entitled "New Transmission System for High-Definition Television (MUSE)," Muse (MUSE, Multiple Sub-Ny-quist Sampli)
ng Encoding) method.

As described in the paper, this method applies a 4: 1 sub-Nyquist sampling that makes a cycle of four fields over a wideband high-definition television signal, and
It is a method of band compression and transmission.

This compressed high-definition television signal (hereinafter muse signal)
FIG. 8 is a block diagram showing the principle of the system described in the paper as a decoder for converting the signal into the original high-definition television signal. The features of this system are: a frame memory 23 that stores a muse signal for four consecutive fields; a circuit 26 that detects a movement between the signal of the current field and a signal from the frame memory 23 one frame or two frames before;
A circuit 24 for performing still region interpolation using the signal of the current field and the signal of 1, 2, 3 fields before from the frame memory 23,
Circuit for motion area interpolation using only the current field signal
25 and a circuit 18 that mixes the still area-interpolated signal and the moving area-interpolated signal by changing the ratio in accordance with the amount of motion from the motion detection circuit 26, thereby performing motion adaptation processing. is there. The problem here is the deterioration of the image quality due to the still region interpolation for the moving image due to an error in the mix ratio of the still region interpolation and the moving region interpolation due to the malfunction of the motion detection. This will be described below.

FIG. 9 shows a block diagram of the motion detection circuit 26. The motion can be detected in principle by detecting the difference between the frames of the video signal. At the input terminal 1 in FIG. 9, a signal including the video signals of one frame before and two frames before,
The video signal of the current frame and that of the previous frame is input to the input terminal 2. A difference signal between the two frames is obtained by introducing the difference between the two signals to the subtractor 4 and taking the difference. Further, since the signal of the current frame is input to the input terminal 3, the pseudo 1-frame difference signal is obtained by taking the difference between this and the input signal of the input terminal 2 to the subtractor 5.

After the absolute values of the difference signals are calculated in the absolute value circuits 6 and 7, the nonlinear conversion is performed in the nonlinear conversion circuits 8 and 9. The characteristic of this non-linear transformation determines the sensitivity of motion detection. For convenience, the outputs of the non-linear conversion circuits 8 and 9 will be referred to as motion amounts. FIG. 10 shows an embodiment of the characteristic of this non-linear conversion, where the absolute value of the input difference signal is d and the output motion amount is m. The maximum value of the absolute value d of the difference signal for which the motion amount output m = 0 is d ₀ , and the minimum d for m = m _max is d ₁ . The reason that d ₀ > 0 is set is to prevent the amount of motion from being detected due to a slight motion of the image or fluctuation due to noise or the like.

The motion amount thus detected is introduced from the motion detection circuit in FIG. 8 to the mix circuit 18, and the image signal Y _s subjected to the static region interpolation and the motion region interpolation are performed according to the magnitude m of the motion amount. The mixing ratio r with the image signal Y _m is continuously changed. The image signal Y after mixing is Y = (1-r) · Y _s + r
・ If Y _m , the mix ratio r is r =
It becomes m / m _max . For example, if the amount of motion is 4 bits and the value of the amount of motion is 5, the mix ratio r = 5 / (2 ⁴ −1) = 1/3
Becomes

However, since this motion detection method has the following problems, the motion amount temporal filter 15 is provided. Since it is difficult to detect motion between fields in motion detection, interpolation errors are likely to occur. In addition, when the moving image processing is suddenly switched to the still image processing, the change in resolution becomes conspicuous and the image quality becomes unnatural. For this reason, the motion amount temporal filter 15 extends the detected motion amount in the time axis direction to reduce detection errors and eases switching from moving image processing to still image processing. Specifically, in the maximum value circuit 12, the amount of motion from the isolated point removal circuit 11 and the output of the coefficient circuit 14 are compared,
The larger value is used as the output of the motion amount temporal filter 15, and this output is delayed by one field in the field memory 13 and then multiplied by α in the coefficient circuit 14 (0 ≦ α ≦ 1).
As a result, the amount of movement is expanded in the time axis direction and attenuated over several fields after detection. There are two types of decay time constants, long and short, and they change depending on the screen content.

FIG. 12 shows a conventional example of the time constant of the motion amount temporal filter 15 and shows the input / output relationship of the motion amount in the coefficient circuit 14. The amount of motion is 4 bits. For example, when a long time constant is used, the output value for the motion amount input value 12 is 10.
FIG. 13 and FIG. 14 are graphs of the attenuation curve when the two long and short time constants shown in FIG. 12 are used, with the horizontal axis representing the number of stretched fields and the vertical axis representing the amount of movement. Shown in the figure.
For example, when a long time constant is used, when the motion amount 15 is detected in a certain field, the motion amount stretched 1 field later becomes 13, and 11 or 9, 11, ... .

If there is a large amount of motion in the image, a large amount of motion will be detected and the mode will be set to full motion picture mode, but if the above-mentioned motion detection error occurs and no amount of motion is detected, the amount of motion temporal filter will be used.
The mix ratio of the static region interpolation and the dynamic region interpolation is determined by the past motion amount stretched in the time axis direction by 15.
However, in the conventional method, as shown in FIG. 10, the motion amount reaches a peak at m = 15 for a large value where the absolute value d of the difference signal exceeds d ₁ in the figure, and the detected difference is detected. No matter how large the signal is, the amount of movement m is attenuated from m = 15 every time one field is extended. If an error occurs in the motion detection, this attenuated past motion amount is used, so that the mix ratio of the still region interpolation in the mix circuit 18 increases, and the image cannot be reproduced in the complete moving image mode. Therefore, image quality deterioration such as multi-line blurring occurs due to an interpolation error especially in an image with a large amount of movement.

Further, in the example of the short time constant shown in FIG. 12, the output for the motion amount input m is about m / 2, and the damping curve is convex downward as shown in FIG. In this example, since a large amount of motion detected at the time of screen switching is sharply attenuated between 1 and 2 fields after detection, image blur caused by interpolating the field before switching to the reproduction of the image after switching. Is likely to occur, and image quality deterioration is particularly noticeable in images in which the screen is frequently switched.

[Object of the Invention]

An object of the present invention is to perform optimum signal processing on a still image and a moving image, and particularly to deteriorate image quality by erroneously performing processing of a still region during moving images, and image quality when transitioning from moving region processing to still region processing. Another object of the present invention is to provide a signal processing circuit suitable for suppressing the deterioration of the signal.

[Outline of Invention]

In the present invention, in order to achieve the above object, a clip circuit is provided in the subsequent stage of the motion amount temporal filter for extending the motion amount in the time axis direction, and a motion amount of a certain value or more is clipped. As a result, when the detected amount of motion is large, image reproduction is performed using only moving region interpolation during the several fields after detection.

Example of Invention

An embodiment of the present invention will be described below with reference to FIG. The feature of this embodiment is that a clipping circuit is provided in the subsequent stage of the maximum value circuit 16.
17 is provided, and other than this clip circuit 17,
This is the same as the above-mentioned conventional example in the figure. The clip circuit 17 clips the motion amount m from the maximum value circuit 16 and leads it to the mix circuit 18 when the motion amount m is equal to or larger than a certain value m ₀ . The characteristics of the mix circuit 18 are such that when the input motion amount m = m ₀ , the mix ratio r = 1, that is, the complete moving image mode is set. Therefore, the amount of movement m from the output of the maximum value circuit 16 and the mix ratio r have the relationship shown in FIG.

The effects of this embodiment will be described below. The characteristics of the nonlinear conversion circuits 8 and 9 in FIG. 1 are shown in FIG. 4, and the time constant of the motion amount temporal filter 15 is such that when the input value to the coefficient circuit 14 is m, the output value is (m-1 ) Will be described. FIG. 7 shows a decay curve of the amount of movement after clipping is performed by the clipping circuit 17. The motion amount m detected in a certain field is m ≦ m ₀ ,
That is, when the absolute value d of the difference signal is d ≦ d ₁ , the motion amount sent to the mix circuit 18 is the motion amount temporal filter 15
By stretching in the time axis direction at 1, the value is attenuated by 1 every 1 field. The detected motion amount m is m> m ₀ ,
That is, when the absolute value d of the difference signal is d> d ₁ , the output of the motion amount temporal filter 15 is attenuated by 1 each time one field elapses after detection, but after the detection that the motion amount exceeds m _0. During a few fields, the amount of motion sent to the mix circuit 18 is clipped to the reference value m ₀ by the clip circuit 17.
At this time, the mix ratio r = 1 and the complete moving image mode is set. In the subsequent fields, the amount of movement sent to the mix circuit 18 is attenuated by 1 every time one field elapses, and shifts to the still image mode.

According to this embodiment, when the absolute value d of the difference signal is a large motion amount such that d> d ₁ , the image can be reproduced in the complete moving image mode for several fields after the detection of the motion amount.
It is possible to reduce the image quality deterioration due to a motion detection error in an image with a lot of motion. Also, when a short time constant is used as the characteristic of the coefficient circuit 14 of the motion amount temporal filter 15 at the time of screen switching, the image is reproduced at the screen switching by performing the image reproduction in the complete moving image mode for a few fields after detecting the motion. In this case, it is possible to reduce the image quality deterioration caused by interpolating the field before switching.

An example of the clip circuit 17 used in the embodiment according to the present invention shown in FIG. 1 is shown in FIG. The comparator 20 shown in FIG. 2 compares the movement amount m with the reference value m _0, and the switching circuit 19 selects the one with the smaller level to obtain the reference value.
A motion amount m of m ₀ or more can be clipped. Further, by switching the value of the reference value m _0, the number of fields for maintaining the complete moving image mode after detecting the motion can be arbitrarily set.

Another example of the clip circuit 17 used in the embodiment according to the present invention shown in FIG. 1 is shown in FIG. In this clipping circuit, the motion amount sent to the mix circuit 18 is (m _max −
Clip to 1) / 2. For example, if the detected amount of motion is 5 bits, the maximum value of the amount of motion m _max = 31, and the clip circuit 17 clips the amount of motion of 15 or more to 15. The clip circuit of FIG. 3 is equivalent to the clip circuit of FIG. 2 with the reference value m ₀ = (m _max −1) / 2, but the circuit configuration becomes very simple.

The above-described effect of the embodiment shown in FIG. 1 is obtained by clipping the motion amount extended in the time axis direction by the temporal filter 15. Therefore, in the circuit configuration of the embodiment shown in FIG. 1, the clip circuit 17 is provided at the subsequent stage of the maximum value circuit 16, but the position at which the clip circuit 17 is provided may be at a stage subsequent to the motion amount temporal filter 15.
FIG. 6 shows another embodiment according to the present invention. In this embodiment, the clip circuit 17 is provided in the preceding stage of the maximum value circuit 16. Similar to the embodiment of FIG. 1, the clipping circuit 17 clips the motion amount of the reference value m ₀ or more. Since it is necessary to prevent the motion amount due to the pseudo 1-frame detection from the output of the non-linear conversion circuit 9 guided to the maximum value circuit 16 from exceeding the reference value m ₀ , this is clipped by the clipping circuit 28 in this embodiment. To do. The other parts are the same as in the embodiment of FIG. According to this embodiment, the same effect as that of the embodiment shown in FIG. 1 is obtained.

Although the present invention has been described by taking a decoder for receiving a high-definition television as an example, the present invention is not limited to the above-mentioned embodiment, and for example, even in an NTSC type television receiver, the amount of motion extended in the time axis direction can be used. It includes all that are characterized by clipping and determining the ratio of static interpolation processing and dynamic interpolation processing by the value. Characteristics of non-linear transformation of motion detection,
Since the number of bits of the motion amount, the attenuation characteristic of the motion amount temporal filter, the clip value, the clipping method, the characteristics of the mix, etc. are arbitrary, the motion detection method, the signal processing method for still images and moving images, the mixing method, etc. By these values,
The characteristics can be set arbitrarily.

〔The invention's effect〕

According to the present invention, when a motion in an image is detected, the image can be reproduced using only the moving region interpolation for an arbitrary period after the detection. It is possible to reduce deterioration of image quality due to field interpolation, and improve image quality.

[Brief description of drawings]

1 and 6 are block diagrams showing a circuit configuration of an embodiment according to the present invention, and FIGS. 2 and 3 are block diagrams showing a configuration example of a clip circuit in the embodiment according to the present invention.
FIG. 4 is a characteristic diagram of a non-linear conversion circuit in the embodiment according to the present invention, FIG. 5 is a characteristic diagram of a mix circuit in the embodiment according to the present invention, and FIG. 7 is a characteristic diagram of a motion amount attenuation curve in the embodiment according to the present invention. , FIG. 8 is a system block diagram of the muse decoder, FIG. 9 is a block diagram of the motion detection circuit in the muse decoder shown in FIG. 8, and FIG. 10 is a non-linear conversion circuit in the motion detection circuit shown in FIG. FIG. 11 is a characteristic diagram showing general characteristics, FIG. 11 is a characteristic diagram showing general characteristics of the mix circuit in the muse decoder shown in FIG. 8, and FIG. 12 is a typical time constant used for the motion amount temporal filter. An example comparison chart, FIG. 13 and FIG. 14 are characteristic diagrams of the motion amount attenuation curve when the long and short two time constants shown in FIG. 12 are used. Description of symbols 1 ... 2 ... 3 input terminals, 4 ... 5 subtractor, 6 ... 7 absolute value circuit, 8 ... 9 nonlinear conversion circuit, 10 ... minimum value circuit, 11 ... isolated point removal circuit, 12 ... 16 ... Maximum value circuit, 13 ... Field memory, 14 ... Coefficient circuit, 15 ... Motion amount temporal filter, 17/28 ... Clip circuit, 18 ... Mix circuit, 19 ... Switching circuit, 20 ... Comparator, 21 ... OR circuit, 22 ... Separation circuit, 23 ... Frame memory, 24 ... Still area interpolation circuit, 25 ... Moving area interpolation circuit, 26 ... Motion detection circuit, 27 ... TCI decoder.

Claims (1)

[Claims] Claim: What is claimed is: 1. A motion detection circuit for detecting a motion of an image between at least one frame or two frames is compared with a motion amount from the motion detection circuit and a signal derived from another input terminal to have a large level. A maximum value circuit that selects one of the two, a delay circuit that delays the output from the maximum value circuit by one field or more, and a coefficient circuit that attenuates the output from the delay circuit and guides it to another input terminal of the maximum value circuit. An apparatus including a mixing circuit for mixing a signal subjected to image processing suitable for a still image and a signal subjected to image processing suitable for a moving image according to a motion amount from an output of the maximum value circuit A television signal processing circuit, characterized in that a clipping circuit is provided after the maximum value circuit and after the signal led to the delay circuit to clip a motion amount of a certain value or more and then lead to the mixing circuit.