JPS62172875A - Reproducing system for multiplex sample transmission signal - Google Patents

Abstract

PURPOSE:To eliminate unevenness in the motion of a picture by performing a picture reproduction as it is, when a multiplex subsample transmission signal is decoded. CONSTITUTION:An input multiplex subsample transmission signal not including a folded component in its low-pass area, the component below 4MHz of which is taken out at a low-pass filter 21, and also, a difference calculation is performed at a difference calculator 27, and a complementary high-pass component is taken out, and it is given to a decoder 22. Meanwhile, the low-pass component of an input transmission signal, after passing through a low-pass filter 21, a flip-flop 28, and a subsample shift switch S, passes a subsampling frequency conversion circuit 23, and its sampling frequency is converted to the same one as the output sampling frequency of the decoder 22. After it passes a delay line 24, it is added on the output of the decoder 22 at an adder 29.

Description

[Detailed Description of the Invention] [Industrial Field of Application] This invention relates to a method for reproducing (decoding) a multiplex sub-sampled transmission signal that does not include aliasing components between frames in the low frequency range (for example, 0 to 4 MHz). In particular, regarding the low-frequency components of the input signal, image reproduction using only the signal of the field causes (1) incomplete motion detection level, and (2) delay in motion of flat parts caused by the use of noise reduction. At the same time, it prevents a decrease in vertical resolution due to imperfections in the two-dimensional or three-dimensional interpolation filter.

In other words, the present invention relates to a receiving system for high-definition broadcasting, and is common regardless of whether it is a satellite transmission system, a package system, an AM transmission system, or a wired transmission system.

[Prior Art] One of the methods for band compression of television signals is a multiple subsample transmission method using interframe and interfield offset subsampling, such as MtlSE (MtlSE).
ultiple 5ub-Nyquisj Sample
There is a current high-definition television signal multiplex sub-sampled transmission method called ingEncoding, which effectively realizes band compression. The details are described in, for example, "Technical Research Report of the Institute of Electronics and Communication Engineers, Image Engineering IE84-72 J.

[Problems to be Solved by the Invention] In the current multiple sub-sampling transmission system, since sub-sampling goes around in two frames, when motion detection is performed on the receiver side, The difference signal cannot be used (there is no partner), and therefore the difference signal between two frames must be used, resulting in incomplete motion detection.

For this reason, a multiple subsample transmission method has been proposed that performs motion detection using a complete frame-to-frame difference signal, which simplifies the configuration of the receiver and significantly improves the image quality (patent application). (See No. 106132/1983).

That is, the multiplex subsample transmission method according to this proposal is a television signal subsample transmission method that compresses the transmission band using interfield and interframe offset subsamples, in which the television video signal is first subjected to interfield offset subsampling, This method is characterized in that the obtained signal is processed using a low-pass filter having a cutoff frequency lower than the sampling frequency of interframe offset subsampling, and then interframe offset subsampling is performed. The transmission signal obtained in this manner does not include an interframe aliasing component in its low frequency range. Therefore, on the receiver side, the low-frequency component is extracted from such a transmission signal, and the difference signal between adjacent frames (difference signal between one frame) is changed by the signal of the extracted low-frequency component. Obtained as a signal for detection. Therefore, accurate motion detection can be performed using such a one-frame difference signal.

FIG. 2 shows the configuration of a decoder on the receiver side for a transmission signal that does not include interframe aliasing components in the low frequency band as described above.

In FIG. 2, 1 is a signal detection synchronization detection circuit,
A transmission signal (interframe subsampling frequency: 16 MHz) that does not include interframe aliasing components in the low frequency range inputted from the signal input terminal 12 as described above is inputted, a synchronization signal is generated, and the synchronization of this decoder is controlled. .

Sl is a switch for subsample shift,
The frame memory 2 receives the transmission signal input to the input terminal 12.
Then, a one-frame delayed signal motion-compensated by the motion compensation circuit 3 is inserted to obtain a signal interpolated between frames sampled at 32 MHz (obtained at A in the figure). Note that the switch SL is not a simple switch, but can be given noise reduction characteristics by taking in 16M11Z input at an appropriate mixing ratio.

The inter-frame interpolated signal of the sampling of 32M) Iz from the switch Sl is input as one input to the motion detection circuit 4 and also to the intra-field interpolation circuit 5 via the switch S2 cooperating with the switch S1. and is further input to the mix (MIX) l1fil path as the - side input.

The transmission signal from the input terminal 12 is input to the other input of the motion detection circuit 4 . The motion detection circuit 4 generates a motion detection signal based on two input signals.

The switch S2 extracts the signal of only the sampling point corresponding to the current field from the signal from the switch Sl, interpolates it within the field by the field interpolation circuit 5 consisting of a low-pass filter, and inputs it to the mix circuit 6.

The mix circuit 6 uses the output signal of the motion detection circuit 4 to determine the amount of motion (from the field interpolation circuit 5).
A moving image signal and a still image signal (from switch Sl) are mixed.

The interfield interpolation circuit 9 interpolates the l-field delayed signal obtained by the field memory 7 and the motion compensation circuit 8 into the output signal from the mix circuit 6, thereby obtaining an interfield interpolation signal sampled at 48 MHz. This is input as one input to the frequency conversion/mixing circuit IO.

The 32 MHz sampling signal from the mix circuit 6 input to the other input of the frequency conversion/mix circuit 10 is converted to a sampling frequency of 48 M) Iz in this circuit IO, and the signal from the interfield interpolation circuit 9 is converted to a sampling frequency of 48 M)Iz. The TCI (Time-C
The signal is input to a decoder 11, where it is converted into a desired television signal, and output from the output terminal.

In the multi-subsample transmission method described above, which does not include aliasing components between frames in the low frequency range, a frame memory is required in the decoder, so this is used to apply noise reduction and improve the transmission path. It is possible to lower the required C/N.

However, since applying noise reduction generally degrades moving images, noise reduction is not applied to moving areas. However, in this case, due to imperfect motion/stillness judgment, flat parts of the image in particular may be judged to be stationary when they are actually moving, and as a result, these parts lag behind the movement. For this reason, when the camera moves, the movement across the screen will somehow not be uniform.

Exactly the same thing occurs in the decoding itself of the current multiple subsample transmission method described above or the multiple subsample transmission method that does not include aliasing components between frames in the low frequency band. The reason for this is that the decoding of these signals uses temporal interpolation in the static domain and spatial (intra-field) interpolation in the dynamic domain, and flat parts may move due to an error in determining motion/staticity. I will be late. Moreover, in the case of these decodings, since it is preferable in terms of overall image quality that the sensitivity of motion area detection is low, this tendency is becoming more and more common.

On the other hand, in FM transmission, VS[]-AM transmission, etc., since most noise is in the high frequency range, there is no need to reduce low frequency noise much.

In addition, when considered as a signal, the multiplex sub-sampled transmission signal that does not include aliasing components between frames in the low band is a low band (4
MllZ and below) does not include an aliasing component between frames, so it is thought that the above problem can be solved using this.

The problems to be solved by the present invention are (1) nonuniformity of motion due to noise reducer, and (2) nonuniformity of motion due to MUSE II decoding. Due to the design of the two-dimensional filter when performing interpolation or inter-field interpolation, ideal characteristics could not be obtained with a small number of taps, resulting in unnecessarily blurred images, but this problem was also resolved in the same way. do.

[Means for Solving the Problem] When decoding a multiplex sub-sampled transmission signal that does not include inter-frame aliasing components in the low frequency range, the present invention provides interframe interpolation, field It is characterized in that images are reproduced as they are without performing interpolation or intrafield interpolation operations.

[Embodiments of the invention] The present invention will be described in detail below with reference to the drawings.

FIG. 1 shows the most basic system of the system of the present invention. In FIG. 1, 21 is a low-pass filter, 22 is a decoder as shown in FIG. 2, 23 is a sampling frequency conversion circuit, 24 is a delay line, 27 is a subtracter, and 28 is a D (delay) type flip-flop ( 32 for clock input
29 is an adder, and S is a sub-sample shift switch.

The input multiple sub-sampled transmission signal that does not contain aliasing components in the low frequency range is filtered by a low-pass filter 21 to extract the degree component below 4 M) Iz, and a subtracter 27 performs subtraction.
Complementary high frequency components are extracted and fed to the decoder 22. On the other hand, the low-frequency components of the input transmission signal are processed by the low-pass filter 21 and the flip-flop 2.
8 and a sub-sample shift switch S, then passes through a sampling frequency conversion circuit 23 and a decoder 22.
The signal is converted to the same sampling frequency as the output sampling frequency of , is passed through a delay line 24 , and then added to the output of the decoder 22 in an adder 29 . Note that the delay line 24 is for matching the delay of the decoder 22.

Although it is possible to manufacture hardware according to this principle system, there are some practical problems. The first method requires low-frequency components in the decoder for motion area detection, so it is necessary to send the low-frequency components to the decoder separately from the main system, making the system complicated. Second, when monitoring signals from various parts of the decoder, it is difficult to tell at first glance because they do not have low-frequency components. Third, if a simple decoder without the improvements of the present invention is considered, a completely different decoder would have to be made.

The method shown in FIG. 3 improves this point, and is considered preferable in practical terms. In Figure 3, 2
2 is a decoder as shown in Fig. 2, 24 is a delay line, 25 is a low-pass filter, 26 is a delay line, 30 is a sub-sample shift circuit, 31 is a 16MH
4 for phasing the z sample and the 48MHz sample.
A circuit for shifting the 8MHz clock (GK) by l/2 synchronization, 32 is a D-type flip-flop (clock frequency is 16MHz), 33 is an adder, and 34 is a D-type flip-flop (clock frequency is 48M)! x), 35 is an adder, 36 is a subtracter, and 37 is an adder.

In this scheme, the decoder 22 is as described above and provides as its output a nearly perfect signal. On the other hand, by separately extracting the low frequency component of the input transmission signal and replacing it with the low frequency component of the output signal of the decoder 22, the same operation as in FIG. 1 is performed isometrically.

In the example of FIG. 3, the output signal of the decoder 22 is a 48MHz sample (strictly speaking, 48.6M1lZ
). A replacement circuit can be configured by the low-pass filter 25 and its phase matching delay line 26, and in the cut-off region of the low-pass filter 25, the output from the decoder 22 is output, and in the region where the low-pass filter 25 is a complete pass, the low-pass filter 25
The output from is printed directly.

Note that the configuration of the low-pass filter 25 as shown in FIG. 4 is sufficient since its characteristics are not particularly noisy. In Figure 4, 40.41, 42,
43. 44 and 45 are delay circuits, and the written numbers (Z-', Z-2.
are adders, 52.53', 54, 55.56 and 57
is a 172 multiplier.

The filter shown in FIG. 4 has the advantage that it does not require a multiplier or the like and can be made extremely easily. Note that, as is well known, the 172 multiplier can perform only bit shifting, so no hardware is required.

[Effect of the invention] According to the present invention, there are the following effects.

(1) 'iff Since the current input signal is used for the low components of the raw image signal, unevenness in image movement caused by decoding or noise reduction is eliminated, increasing the stability of the image.

(2) Therefore, noise reduction can be applied more strongly than when the present invention is not used, and the required C/N can be lowered accordingly.

(3) Unstable motion due to incomplete motion compensation (motion compensation is performed using an incomplete motion vector) can also be eliminated by the present invention.

(4) Since the low-frequency component uses a signal that is not filtered in the vertical direction, the vertical resolution is perfect, so even if the vertical resolution is degraded due to incomplete interpolation in the decoder. You can cover it even if you are. Regarding vertical resolution, as is well known from a visual point of view, the vertical resolution of horizontal low-frequency components is important.

[Brief explanation of drawings]

Figure 1 is a system diagram showing the basic structure of the youngest son style.
:' Fig. 3 is a system diagram showing an example of a practical system of the present invention, and Fig. 4 is a diagram showing a specific example of a low-pass filter used in the present invention. 21...Low pass filter, 22...Decoder, 23...Sampling frequency conversion circuit, 24...Delay line, 27...Difference device, 28...D type flip-flop, 29... Adder.

Claims (1)

[Scope of Claims] 1) When decoding a multiplex sub-sampled transmission signal that does not include interframe aliasing components in the low frequency range, interframe interpolation, interfield interpolation, and intrafield interpolation are performed for the low frequency components of the transmission signal. A method for reproducing multiple subsampled transmission signals, which is characterized in that it can be used as is for image reproduction without interpolation. 2) A method for reproducing a multiple sub-sampled transmission signal according to claim 1, characterized in that an inlet for the low-frequency component is provided before a noise reduction system. 3) In the method for reproducing a multiple sub-sampled transmission signal according to claim 1 or 2, only the low-frequency component of the signal resulting from the decoding is replaced with the low-frequency component of the input transmission signal. A method for reproducing multiple sub-sampled transmission signals characterized by: 4) In the method for reproducing a multiple sub-sampled transmission signal according to claim 3, the multiple sub-sampled transmission signal is characterized in that the signal to be replaced is not used as the final output, but is used as a signal before inter-field interpolation. playback method.