JPH07112276B2 - Decode signal processing method - Google Patents

Description

The present invention relates to a high-definition television signal transmission system, and more particularly to a decoding signal processing system on the receiving side of a multiple sub-sampling band compression transmission system.

MUSE (Mu ltipl to one of the (prior art) the inventors have a high definition television signal bandwidth compression transmission technology already proposed (compressed to about 1/4)
there is e S ub-Nyquist Sampling E ncoding ) method.

FIG. 3 is a block diagram of the principle configuration of the sampling pattern (a), the transmission side encoder (b) and the reception side decoder (c) of this MUSE transmission system.

The sampling pattern shown in FIG. 3 (a) is a sub-sampling pattern in the order of 2 frames, and is marked with ○, □,
The ●, ■, and × marks are the 4n + 1st and 4n + 2, respectively.
The 4th, 4n + 3rd, 4th (n + 1) th (n = 0,1,2
......) Indicates the sub-sampling position of the field and the sampling point that does not transmit. Further, d in the drawing is a sampling interval, and at the sampling frequency, 1 / d corresponds to, for example, 64.8 MHz.

On the other hand, in the transmission side encoder shown in FIG. 3 (b), first, the RGB input signal of the high definition television signal is the TCI encoder 31.
After being made into a TCI (Time Axis Compression Multiplex) composite signal, the encoding process required for the MUSE method has been performed. Therefore, Y signal (luminance signal), C signal (color signal)
It is not necessary to make a processing circuit separately. In the encoding process required for the MUSE method, the still area prefilter 32 or the mixer 37 is used for the signal processing of the subsequent two-frame one-round fields and the inter-frame offset sub-sampling 38, and the decoder on the receiving side of the MUSE method ( Pre-processing of the image signal suitable for the processing of FIG. Regarding these details, Ninomiya: New transmission system for high-definition television ~ MUSE ~, NHK Giken monthly report, Vol.27, No.7, pp.
19-30 (1984). This MUSE method has achieved satisfactory results as band compression transmission of high-definition television signals, but since it is a multi-subsample transmission of one cycle of two frames, the motion information required by the receiving side decoder is 2 Only the interframe difference signal can be obtained accurately. There is also a MUSE system in which the completeness of this motion information is improved and the decoder side can also obtain a one-frame difference signal. For details of this, Japanese Patent Application No. 60-10
See 6132, "Multiple Subsample Transmission Scheme".

(Problems to be solved by the invention) Since the MUSE method adopts the sub-sampling method,
There is aliasing of high frequency components into the transmission baseband due to sampling. Of course, the two components are interleaved so that the low frequency component and the folded high frequency component do not overlap (see Japanese Patent Application No. 59-30145).
However, if the signal processing in the middle becomes complicated or if transmission distortion is superimposed, it causes deterioration of both qualities. These distortions are called folding distortions. Considering this point also in the conventional MUSE system, for example, by using the static region prefilter 32 or the moving region prefilter 33 as a signal prior to sub-sampling, it is possible to prevent unnecessary unnecessary high frequency components from being brought to the low frequency side, The high frequency components of the signal are designed so that the folds do not fall into the baseband at all. However, this was very limited, and removal of the aliasing distortion component was insufficient.

In addition, since video / still image switching by this transmission method is essentially a non-linear signal processing, new distortion is introduced by the filter processing and switching, and moving area detection for video / still image switching is performed. However, there is a drawback in that distortion is introduced due to the fact that switching is not possible with sufficient accuracy, especially with a component having a small image amplitude.

Therefore, the object of the present invention is to eliminate the above-mentioned drawbacks, and
By providing a temporal filter processing system with a simple structure next to the SE-type receiving side decoder, the above-mentioned various distortions remaining in this transmission method are deleted, and a decoder signal processing method that can maintain high image quality is provided. To do.

(Means for Solving the Problems) In order to achieve this object, the decoder signal processing method of the present invention is such that a plurality of filters on the receiving side are actively switched in a multiplex subsampling band compression transmission method of a television signal. In decoding the transmitted television signal, the high-frequency component of the image of the decoded signal is subjected to temporal filter processing in the time axis direction, and the degree of the temporal filter processing is transmitted to the television. It is characterized in that the aliasing distortion component contained in the transmitted television signal is removed by being controlled by the image information signal contained in the television signal.

A preferred embodiment of the method of the present invention is characterized in that the image information signal is a signal representing an absolute value of a high frequency component of an image.

(Example) Hereinafter, the method of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 (a) shows a block diagram of the principle configuration of the method of the present invention, and FIG. 1 (b) shows a configuration block diagram of another embodiment in FIG. 2 of one embodiment.

The MUSE signal transmitted in FIG. 1 (a) is input to the input terminal thereof, and the MUSE decoder 1 restores a high quality color television signal. This is the conventional example,
The signal processing circuit added from this point onward is the configuration related to the method of the present invention. More detailed description of FIG. 3 (c) shows that the signal restored by the conventional MUSE decoder is TCI-decoded at the end, so that it is a set of R, G, B color signals, or Y (luminance signal). , Cw (wideband color signal), Cn (narrowband color signal). The signal processing relating to the signal processing method of the present invention is applied to the output signals of this set. In this case, this may be applied to all the signals of the above-mentioned set of three signals, or the signal component of particular interest. This may be applied to only.

The output signal decoded by the MUSE decoder 1 in FIG. 1A is filtered by the complementary high pass / low pass (HP / LP) filter 2 before being processed by the temporal filter 3. This filter may be a two-dimensional filter process or a one-dimensional filter process. (HP / LP) The high-frequency component HPS of the signal processed by the filter 2 is filtered by the temporal filter 3 in the time axis direction, and added by the low-frequency component LPS output from the filter 2 and the adder 4. It is used as an output signal. In the figure, another output signal from the MUSE decoder 1 is input to the temporal filter 3. This is for controlling the degree of filter processing by the temporal filter 3 by the image information signal extracted by the decoder 1 described later. It is a thing.

FIG. 1 (b) shows a block diagram of one configuration in the case where the decoder 1 does not control the temporal filter 3 in the block diagram of the basic configuration of FIG. 1 (a).

In this figure, vertical LPF5 and horizontal LP
Vertical and horizontal low pass filter processing is performed via F6, and the difference between this output signal and the input signal via the delay line 7 is obtained by the subtractor 8 to obtain the high frequency component of the image. obtain. The high-frequency component is filtered by the temporal filter 9 in the time axis direction, for example, in the frame direction. Next, by using the adder 10 and the 1/2 coefficient unit 11, the signal passing through the temporal filter 9, the signal not passing through the temporal filter 9 and the signal obtained by adding the two signals are multiplied by the coefficient 1/2. Signal, that is, an average signal of both signals, is generated, and the absolute value minimum value selection circuit 12 selects the one having the minimum absolute value from these three signals, and outputs the horizontal direction LPF6. That is, the adder 13 adds the low frequency component of the input signal to obtain the output signal.
The meaning of selecting the signal having the smallest absolute value out of the three signals is to insert the moving image of the input image signal in order to eliminate multi-line blurring due to the temporal filter processing and achieve stable image signal processing. However, the details will be apparent from the specification of the patent application (1) "Image signal processing method" filed on the same day as this application which is the present applicant.

In the configuration of FIG. 1 (b), the LPF is applied both horizontally and vertically, but this does not mean that both are always operated, and it is possible to use only one of them. It is also possible to control the gain of LPF according to the characteristics of the signal.

Next, the configuration block diagram of another embodiment relating to the system of the present invention in which the degree of the filter processing by the temporal filter 3 is controlled by the image information signal extracted by the MUSE decoder 1 shown in FIG. It is shown in FIG. FIG. 2A shows an embodiment in which the motion amount level P, the edge amount level Q, and the high frequency component absolute value amount R are all used as the image information signal.

Of the three inputs shown on the left side of FIG. 2 (a), the high frequency components are taken from the HPF passing parts of the HP / LP filter 2 shown in FIG. 1 (a), and the movement amount and edge amount are shown in FIG. 3 (c). Each is extracted from the motion part detection 45 of the MUSE decoder shown. Motion amount level detector 22, edge amount level detector based on these input signals
23 and a high frequency component absolute value detector 24 detect signals of a motion amount level P, an edge amount level Q and a high frequency component absolute value R, respectively. Of these output signals, the signals P and Q are, for example, the edge amount of the pixel P ₂ as shown in FIG. Also, the motion amount can be obtained for each pixel by dividing the frame difference of the image signal by the edge amount. Using these signals P, Q and R, the most effective temporal filter output signal is obtained in advance. The calibrated function generator 25 controls the coefficients β21 and α28 of the coefficient unit β21 and the coefficient unit α28 to obtain the temporal filter. Controls the degree of filter processing.

As an operation example of the function generator 25, when the signal P is small or large, the coefficient α (α ₁ ) is small, and when the signal P is medium, the coefficient α is large. When the signal P is small and the signal Q is large, it is made slightly larger than the coefficient α ₁ described above. As for the relationship between the signal R and the coefficient β, as shown in FIG. 2 (c), when the signal R becomes small, the coefficient β tends to be suppressed, and in the portion where the signal R becomes large, it is preferable to make the change of the coefficient β linear. To be The meaning of the above control is to apply less temporal filter to areas with little movement and large movement, and to apply large temporal filters to areas with moderate movement. This is because when the motion is small, that is, when the motion is almost stationary, the image is suddenly blurred when the temporal filter is applied, and when the motion is large, the motion is blurred by the temporal filter. The case where the present invention is applied to the MUSE transmission method has been described above, but the present invention is not limited to this.
Of course, it can be applied to other types of high-volume sub-sampling band compression transmission systems.

(Effects of the Invention) As described in detail above, the signal processing method of the present invention is applied to the multiple sub-sampling band compression MU for high definition television signals.
By applying it to the receiving side decoder of the SE transmission system, the effect of removing the aliasing distortion component and various distortion components, which had been a problem in the conventional configuration, was obtained.

[Brief description of drawings]

FIG. 1 shows a block diagram of a principle configuration (a) and a configuration of one embodiment (b) of the system according to the present invention, and FIG. 2 shows a configuration diagram (a) of another embodiment of the present invention and its configuration. FIG. 3B is a diagram for explaining the edge amount detection and FIG. 3C is a diagram showing a suitable relationship between the signal R and the coefficient β. FIG. 3 is a sub-sampling pattern (a) of the conventional MUSE transmission system and a transmitter encoder ( The block diagram of b) and a receiving side decoder (c) is shown. 1 …… MUSE decoder, 2 …… HP / LP filter 3,9 …… Temporal filter 4,10,13,26 …… Adder 5,6 …… Vertical and horizontal LPF 7 …… Delayer, 8 ・ ・ ・… Subtractor 11, 21, 28 …… Coefficient unit, 12 …… Absolute value minimum selection circuit 22 …… Motion amount level detector 23 …… Edge amount level detector 24 …… High frequency component absolute value detector 25 …… Function generator 27,36,42 …… Frame memory 31,47 …… TCI encoding / decoding 32 …… Still area prefilter 33 …… Motion area prefilter 34 …… Motion vector detection, 35,45 …… Motion detection 37 , 46 …… Mixed, 38 …… Subsampling 39 …… Multiplexed, 41 …… Separated 43 …… Motion area interpolation, 44 …… Still area interpolation

Front page continuation (72) Inventor Seiichi Koshi 1-10-11 Kinuta, Setagaya-ku, Tokyo Inside the Japan Broadcasting Corporation Broadcasting Technology Laboratory (72) Inventor Yuichi Iwadate 1-1-10 Kinuta, Setagaya-ku, Tokyo Japan (56) References Japanese Patent Laid-Open No. 59-178886 (JP, A)

Claims (2)

[Claims] 1. In a multiplex sub-sampling band compression transmission system of a television signal, when a television signal transmitted by actively switching a plurality of filters on a receiving side is decoded, The high frequency component of the image is subjected to temporal filtering in the time axis direction, the degree of the temporal filtering is controlled by the image information signal included in the transmitted television signal, and the transmitted television signal. A decoding signal processing method characterized in that the aliasing distortion component included in is removed. 2. The decoding signal processing system according to claim 1, wherein the image information signal is a signal representing an absolute value of a high frequency component of an image.