JPH0324882A - Signal converter - Google Patents

Abstract

PURPOSE:To set a picture easy to observe by heightening the sensitivity of motion detection when correcting a motion vector by providing a sensitivity switching circuit in a motion detection circuit, and performing a processing to expand a motion detecting signal in a time base direction by providing a temporal filter simultaneously in the motion detection circuit. CONSTITUTION:Subsampled phase data in an image signal and a control signal are supplied to an in-field interpolation circuit 6 after delaying by one frame, respectively, and the in-field interpolation of a moving image system processing is processed with a signal of one preceding frame. Furthermore, the sensitivity of the motion detection when the motion vector being corrected is heightened by providing the sensitivity switching circuit 28 in the motion detection circuit 27, and simultaneously, the processing to expand the motion detecting signal in the time base direction is performed by providing the temporal filter 29 in the motion detection circuit 27. Thereby, the picture can be set easily to observe.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a signal converter, and particularly to a signal converter I4 device that converts a MUSE signal into a signal that can be received by a current television receiver. .

(Prior Art) A MUSE method has been proposed that compresses the bandwidth of high-definition television broadcasts and enables transmission by satellite broadcasting, and experimental broadcasting is currently underway.

MUSE isllUltiple sub-41yQui
This is an abbreviation for st Sallprirl (l encoding) and is a method developed by Nl{K (Bon Broadcasting Association).

The MUSE method is described in various documents (for example, rNHK Technical Research, 1988, Vol. 32, No. 2, p. 8-53 rMUs "Method Information" and Nikkei McGraw-Hill's "Nikkei Electronics"). l-D Nix J1
November 2, 1987 issue, p189-p212 ``Transmission system for high-definition broadcasting using satellites, MUSEJ, etc.'' A detailed explanation will be omitted here.

The MUSE method luminance signal (Y signal) is transmitted on the transmitting side by converting the original signal of a high-definition television signal (luminance signal @) having a band of approximately 22 MHI to 48.6 MHz as shown in FIG. 6 (A).
AD conversion is performed at a sampling frequency of 16.2 MHz, and the sampling frequency is increased to 16.2 MH. , compress the data, and D
It converts the signal back into an analog signal and transmits it.

This signal is shown in FIG. 6(A) as shown in FIG. 6(B).
8 of ■～■ shown in . IMH. The Takagi component above is 8.1M
H. Folds back within the band and increases the baseband bandwidth to 8.1MH
．． It is compressed into .

Receive this band-compressed MUSE signal. It is the MUSE-j' feeder (receiver) that is used.

However, as is well known, the MUSE decoder requires a very large-scale circuit and a special cathode ray tube with an aspect ratio of 1620, and is very expensive.

Therefore, a signal converter (down converter) is being considered that converts the MtJSE signal with 11125 scanning lines into a signal having 525 scanning lines that can be received by the current television receiver.

FIG. 7 is a block diagram showing a conventional signal conversion device.

In FIG. 7, the MUSE signal coming into input terminal 1 is approximately 8.1 MH. It is supplied to the AD converter 3 via a low-pass filter 2 that passes frequencies below 1
6.2MH. The clock signal is resampled and becomes a digital signal.

The output signal of the AD converter 3 is supplied to a de-emphasis circuit 4, where it is subjected to de-emphasis processing, and
It is also supplied to the control signal separation circuit 5, and the control signal is supplied to the control signal separation circuit 5.
signal is separated.

As is well known, the control signal includes motion vector data, sub-sampled phase data, and the like.

The output signal of the digital seven-system circuit 4 is supplied to the field interpolation circuit 6, and the video part. Regardless of the still image portion, all video processing is performed by interpolating from sample points within the current field.

The pixels of the MUS signal are in a state where the pixels of the current field and the pixels of the previous frame are offset between frames, and the intra-field interpolation t8 method uses data at this non-sampled point (interpolation point). It is created from the data of surrounding zumbling sample points) and interpolated.

Incidentally, the sampling point and the interpolation point are identified by the sub-sampled phase data in the control signal supplied from the control signal q separation circuit 5.

The output signal of the de-emphasis circuit 4 is also supplied to a one-frame delay device 7 and an interframe interpolation circuit 8 at the same time.

Since the one-frame delay device 7 operates to delay the input signal by one frame, its output signal becomes a signal obtained by delaying the input signal by one frame.

The output signal of the de-emphasis circuit 4 and the output signal of the one-frame delay device 7 are respectively supplied to an interframe interpolation circuit 8.

The interframe interpolation circuit 8 interpolates the current field signal and 1
Performs interframe interpolation processing between signals before the frame,
Still image processing is performed by creating data at unsampled points (interpolation points) from data at sampled points one frame before (sample points).

Then, aliasing components due to interframe offsets are removed, improving image quality.

However, as is well known, the MUSE signal is sampled offset between fields, and aliasing components for this still remain.

A part of this review is from the above-mentioned material rNHK Technical Research, 1988, Vol. 32. No. 2 p18-p53rMUsE
12 to 24 MH. The band components are folded back.

Then, this component can be removed by a vertical low-pass filter that blocks 1125 TV lines and the vicinity thereof, which are expressed by regarding the vertical frequency as 1125 TV frames without considering interlacing.

The output signal of the interframe interpolation circuit 8 is supplied to a vertical low-pass filter 9.

The vertical low-pass filter 9 removes a certain portion of 112 STV vertical frequencies and removes aliasing components due to inter-field offsets.

In the interframe interpolation process, the sample point of the current field is determined based on the subsample phase data detected by the control signal separation circuit 5.

Next, interframe interpolation processing when a stationary image is translated in parallel due to banning or the like operates as follows.

That is, in this case, the images between one frame are displaced due to parallel movement, and interpolation between frames cannot be performed as is.

Therefore, in the control signal separation circuit 5, the motion vector data transmitted from the transmitting side is extracted, and the delay amount of the 1-frame delayer 7 is tlJtil according to the amount of motion vector 1.times. After eliminating positional deviations, interframe interpolation processing is performed.

The two signals, the output signal of the intra-field interpolation circuit 6 which has been subjected to moving image processing, and the output signal of the vertical low-pass filter 9 which has undergone still image processing, are respectively supplied to a mixing circuit 10 and mixed. being processed.

The output signal of the mixing circuit 10 is supplied to a scanning line conversion circuit 11.

Further, the output signal of the de-emphasis circuit 4 and the output signal of the one-frame delay device 7 are respectively supplied to a subtracter 13 in the motion detection circuit 12.

The motion detection circuit 12 includes a subtracter 13 and a low-pass filter 1.
4 and an absolute value circuit 15, and its operation will be explained below.

The subtracter 13 extracts a difference signal between frames.

This inter-frame difference signal is supplied to a low-pass filter 14.

The low-pass filter 14 is a low-pass filter that passes some frequencies below 4 MHz.

That is, as is clear from Figure 6 (B), the folded precept is 4MH. Since the following components do not exist, the low-pass filter 14 can prevent motion detection from malfunctioning due to aliasing components.

The output signal of the low-pass filter 14 is an absolute filter 1
5, the absolute value of the signal is obtained, and the amount of motion is detected.

The amount of motion, which is the output signal of the absolute value circuit 15, is supplied to the mixing circuit 10, which operates to control the mixing ratio of the two signals in the mixing circuit 10 in accordance with this motion.

That is, when the amount of motion is below a predetermined level, the mixture is performed in accordance with the amount of motion, and when it is above a predetermined level, only signals for video processing are output.

The scanning line conversion circuit 11 converts signals of 1125 scanning lines per frame by, for example, controlling writing to and reading from memory, while keeping the frame frequency (30 horses) unchanged.
It is converted into a signal of 525 scanning lines per frame.

Note that the read clock frequency from the memory is selected at a frequency that does not distort the image during conversion.

The output signal of the scanning line conversion circuit 11 is supplied to a TCI decoding circuit 16.

The color signal (C signal) in the MUSE signal is the luminance signal (
T C T (Time coa + pressure
It is transmitted in a signal format called rntegration.

The TC1 dehood circuit 16 time-expands the C signal by four times, converts it into two color difference signals, and outputs the two color difference signals together with the Y signal.

The output signal of the TCI decode circuit 16 is supplied to a DA converter 17 and becomes an analog signal.

The output signal of the OA converter 17 is supplied to an NTSC encoder circuit 18.

The NTSC encoder circuit 18 creates an NTSC composite video signal from the two color difference signals and the Y signal, and outputs it from the output terminal 19.

FIG. 8 is a diagram showing the internal configuration of the vertical low-pass filter in FIG. 7, and FIG. 9 is a diagram showing a pixel arrangement for explaining the operation of FIG. 8, and will be explained together.

In FIG. 8, the current line signal @to, which is the output signal of the interframe interpolation circuit 14 in FIG. 7, is supplied to the delay device 21 via the input terminal 20.

The delay device 21 delays the input signal by 562 lines and outputs 5 lines.
A signal j5s2 from 62 lines earlier is obtained.

The signal 1562 from 562 lines earlier is supplied to the delay device 22.

The delay device 22 delays the input signal by one line and outputs 00
A signal j563 563 lines before is obtained.

These signals 1562 and 1563 are each supplied to an adder 23, added together, and further halved and output.

The output signal of this adder 23 and the signal to of the current line are:
They are each supplied to the adder 24 and added, and further 1/
2 and becomes a signal l, which is output from the output terminal 25.

Therefore, this signal l is expressed by the following equation.

1=1 0/2+1 5 6 2/4+j 5 8 3
/4 these jo, i562. The relationship on the screen of l563 is as shown in FIG.

By performing processing as in the above formula, 1125 TV books'i
The vertical frequency component near i is removed, and the filter operates as a vertical low-pass filter.

(Problem to be Solved by the Invention) However, the conventional signal conversion device described above uses motion adaptive processing that performs processing separately for still images and moving images in order to remove aliasing. In processing, aliasing is removed by an interframe interpolation circuit 8 and an interfield vertical low-pass filter 9.

However, in the conventional motion detection circuit 12, one frame detection is performed based on one frame difference, so four
M.H. There is a problem in that only the following bands can be used, and motion signals of JMl& and above are not detected.

Furthermore, when motion vector correction is performed, the sending side (encoder side) processes it as a still image system, so if a mistake is made in motion detection, the position of the moving part of the image will shift, making it look discontinuous. Severe interference appears.

As a countermeasure for this, the motion detection sensitivity is adjusted when motion vector correction is performed.
A method was used to raise it.

However, even in this case, the following problems occurred.

FIG. 10 is a diagram showing positional deviation during motion vector correction. FIG. 10 (A> represents the time to enter lJ/71·le correction (fFf! start time), and FIG. 10 (B) represents the time to exit from motion vector correction (end time).

In this figure, inter-frame interpolation will be explained.

The waveform in the figure shows the position of the image signal in each frame m1 to m3, and the presence or absence of motion vector correction means the presence or absence of a motion vector signal sent from the transmitting side.
This may be considered as the presence or absence of the output signal of the absolute value circuit 15 in FIG.

Further, the ratio of moving image signals is the ratio of hatching in the mouth, and indicates the mixing ratio of moving image signals in the mixer 10.

If it is 100% hatched, it means that 100% moving image processing will be performed.

In FIG. 10 (Δ), there is no motion vector correction in frames ml and m2, and there is correction in frame m3.

It can be seen that motion vector correction is entered (started) from the current frame m3.

If a motion detection error occurs in frames m1 and m2, the moving image signal is interpolated as shown by the waveform shown by the broken line, resulting in a double image.

At this time, in frame m3, a motion vector signal enters and the mixer 10 strongly υ1 causes 100% video processing, or when increasing li4 degrees and similarly results in nearly 100% video processing. As shown in the figure, a positional shift occurs in this frame m3.

This positional shift corresponds to the distance between two frames, and the moving parts of the image appear discontinuous.

In FIG. 10(B), the frames ml and m2r have motion vector correction (yes), and the frame m3 has (no) motion vector correction, which is the case where motion vector correction is exited (finished). No major positional shifts occur and there are no visual problems.

These problems in motion vector correction are the first problem to be solved.
This is an issue.

Furthermore, there are cases where the motion detection circuit 12 makes a motion detection error (or insufficient detection) on the same screen and processes the video signal using the still image system, and where the video signal is completely detected and is processed using the video system. If the two points occur at the same time, the video signal to be processed will be
In particular, when moving in the same direction and the moving speed fluctuates, a speed fluctuation phase difference occurs in both processing systems.

In some cases, objects moving in the same direction may appear to flicker, as if they are moving in opposite directions, which is an unnatural phenomenon that visually appears to be inconsistent movement.

If the function of the motion detection circuit on the receiver (decoder) side is perfect, that is, if it matches the motion detection circuit on the transmission (encoder) side, the unnatural phenomenon due to the speed fluctuation phase difference will not occur.

However, since the signal on the receiver (decoder) side is offset by 1 between frames, it is not possible to perfectly detect motion between one frame as is done on the transmitter (encoder) side. It is extremely difficult to eliminate the adverse effects caused by phase difference.

FIG. 11 is a diagram showing the time axis phase difference between the still image system and the moving image system when a motion detection error occurs.

01 to n4 represent each field number, where n4 is the current field, n3 is one field, and msn2 is two fields (one frame) before.

In FIG. 11, as mentioned above, in the still image system, inter-frame and inter-field interpolation processing is performed for four fields 01 to n4 (
4v), and four images 01 to n4 are superimposed.

Therefore, the center phase of the e-image signal that has not been processed in the still image system is
It exists between fields n2 and 03.

On the other hand, in video systems, the current intra-field interpolation process is
The moving image phase is in the current field n4.

Therefore, as is clear from FIG. 11, the phase difference between the center phase of the moving image 4s @ processed by the still image system and the phase of the moving image system is 1.5 fields (1.5 V), and the maximum phase difference is There will be 3 fields (3v).

This phase difference is the origin of the speed fluctuation phase difference of the moving image signal, and is the second problem to be solved.

The present invention has been made with attention to the above points, and the sub-sampled phase data in the image signal and the control signal are each delayed by one frame and supplied to an intra-field interpolation circuit to process the moving image system. The intra-field interpolation is processed using the signal from one frame before, and a sensitivity switching circuit is provided in the motion detection circuit to increase the sensitivity of motion detection during motion vector correction.At the same time, a temporal filter is provided in the motion detection circuit. , by processing the motion detection signal in the time axis direction. It is an object of the present invention to provide a signal conversion device that solves the second problem.

(Means for Solving the Problems) In order to achieve the above objectives, a signal that receives and demodulates an MtJSE signal obtained by band-compressing a high-definition television signal and converts it into a signal that can be received by current television receivers. The conversion device includes an AD converter that resamples the MUSE signal using a clock signal, a de-emphasis circuit that performs de-emphasis processing on the output signal of the AD converter, and the de-emphasis? a first one-frame delay device that delays the output signal of the cis circuit by one frame; an intra-field interpolation circuit that interpolates the output signal of the first one-frame delay device within the field to perform moving image processing; Dayenf? an interframe interpolation circuit that performs still image processing by interpolating the output signal of the cis circuit between frames, and a control that separates control signals such as motion vector data and subsample phase data from the output signal of the AD converter. a signal separation circuit; a second one-frame delay device configured to delay the sub-sampled phase data by one frame and supply it to the intra-field interpolation circuit to control the intra-field interpolation phase; 1 from the output signal of the insertion circuit
1. A vertical low-pass filter that removes vertical frequency components in the vicinity of 25 TV lines, a mixing circuit that mixes two signals, an output signal of the intra-field interpolation circuit and an output signal of the vertical low-pass filter, and the mixing circuit. a scanning line conversion circuit that converts the 1125 scanning line signals outputted by
An inter-frame difference signal is obtained from the output signal of the 7cis circuit and the output signal of the first 17-REM delayer, and the amount of motion is detected by converting it into an absolute value through a low-pass filter. Detection! and a motion detection circuit that controls the mixing ratio of the two signals in the mixing circuit according to the signal obtained by switching i4 degrees using the motion vector data and stretching it in the time axis direction in field units. The present invention provides a signal conversion device characterized by the following configuration.

(Embodiment) Fig. 1 is a block diagram showing an embodiment of a signal conversion device of the present invention.The same parts as in Fig. 7 are designated by the same reference numerals.

The main difference in configuration from FIG. 7 is that the 1-frame delay unit 26
, sensitivity switching circuit 28. This is because a temporal filter 29 is provided.

The operations of only the different parts will be explained below.

A de-emphasis circuit 4 is connected to the field interpolation circuit 6.
A signal obtained by delaying the output signal by one frame by one frame delay device 7 (first one frame delay device) is input.

Also, the interpolation phase in the field interpolation circuit 6 is ¥1,
II tl is Control 1-Roll signal separation. The sub-sampled phase data in the control signal supplied from the circuit 5 is delayed by one frame by the one-frame delay device 26 (second one-frame delay device).

That is, the sub-sampled phase data in the image signal and the control signal are each delayed by one frame and supplied to the intra-field interpolation circuit, and intra-field interpolation for video processing is processed using the signal of one frame before. Become.

As a result, as shown in FIG. 11, the moving image system is delayed by one frame (two fields), and the center phase of the moving image signal processed by the still image system has a phase difference of 0.5 field.
This is within a range that poses no visual problem, and the second problem can be solved.

Furthermore, the output signal @ of the absolute value circuit 15 in the motion detection circuit 27 is transferred to the one-time switching circuit 28 and the temporal filter 29.
and is supplied to the mixing circuit 10. As mentioned above, it is used to control the mixing ratio of still image signals and moving image signals.

FIG. 2 is a diagram for explaining the operation of the sensitivity switching circuit. This circuit has nonlinear characteristics.

The output signal of the absolute value circuit 15 is sent to the sensitivity switching circuit 2.
8, and during motion vector correction, the sensitivity is switched from normal characteristic A to characteristic B in FIG. 2, increasing the sensitivity. That is, during motion correction, the sensitivity is forcibly increased.

The output signal of the sensitivity switching circuit 28 is supplied to a temporal filter 29.

Temporal filters process input signals field by field.
It is stretching in the direction of the time axis.

FIG. 3 is a diagram showing an example of the configuration of a temporal filter, and FIG. 4 is a diagram for explaining the operation of the temporal filter, which will be explained together.

The input signal coming from the input terminal 30 is sent to the MAX processor 3.
1.

The MAX processor 31 selects and outputs a signal with a large value from among the two input signals.

The output signal of the MAX processor 31 is output from an output terminal 32 and is also supplied to a 1-field delay device 33.

The 1-field delay device 33 provides a delay of 562 lines to 563 lines, and obtains a delay of 1125 lines by going around twice.

The output signal of the 1-field delay device 33 is supplied to the arithmetic unit 34.

The arithmetic unit 34 operates to multiply by a constant α having a relationship of 0 × α × 1, or to subtract a value less than the maximum value of the input signal.

Therefore, the input signal to the temporal filter is as shown in Fig. 4 (
As shown in A), even if it exists only in field n1, the output signal to the temporal filter is as shown in Fig. 4 (B).
As shown, a signal is obtained that is stretched and attenuated in the time axis direction field by field over fields 01 to n5.

FIG. 5 is a diagram showing a positional deviation during motion vector correction.

In FIG. 5(A), motion vector correction is (no) in frames ml and m2, and (with) motion vector correction in frame m3.

It can be seen that motion vector correction starts from the current frame m3.

If a motion detection error occurs in frames m1 and m2, the video signal is interpolated as shown in the waveform shown by the broken line.
In this case, as shown in the figure, no large positional deviation occurs,
There are no visual problems.

In FIG. 5(B), motion vector correction is present in frames ml and m2, and is absent in frame m3, and motion vector correction is
- This is the case when exiting (ending) from the correction.

If a motion detection error occurs, if motion vector correction is performed as in the past, the video signal will be interpolated as shown in the waveform shown by the broken line, even if the video signal ratio is 0%. For example, it would normally be a double image of the distance between two frames.

However, in this case, even if motion vector correction is (absent) in frame m3, the motion detection sensitivity is increased during motion vector correction as shown in characteristic B in FIG. Since it is stretched and attenuated by the filter, the proportion of the moving image attenuates with a time constant as shown in FIG. 5(B), so interpolation due to motion detection becomes slight and there is no visual problem.

Therefore, the first problem is also solved.

In this embodiment, the young detection process is performed using a 1-frame difference signal, but it goes without saying that the same effect can be obtained using a 2-frame difference signal.

(Effects of the Invention) The signal conversion device of the present invention has the above-described configuration, and performs intra-field interpolation by delaying each of the sub-sample phase data in the image signal and control 1/roll signal by one frame. The signal is supplied to the circuit, and the field interpolation for video processing is processed using the signal from one frame before.Furthermore, a sensitivity switching circuit is provided in the motion detection circuit to increase the sensitivity of motion detection during motion vector correction, and simultaneously perform motion detection. By providing a temporal filter in the circuit and processing the motion detection signal in the time axis direction, the first. This has excellent practical effects, such as being able to solve the second problem and provide high-quality images.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the signal conversion device of the present invention, FIG. 2 is a diagram for explaining the operation of the sensitivity switching circuit, FIG. 3 is a diagram showing an example of the configuration of a temporal filter, and FIG. is a diagram to explain the operation of a temporal filter,
Figure 5. Fig. 10 is a diagram showing the positional deviation during motion vector correction, Fig. 6 is a diagram showing the frequency spectrum characteristics of the signal of the MtJSE transmission system, Fig. 7 is a block diagram showing a conventional signal conversion device, and Fig. 8 is a diagram showing the frequency spectrum characteristics of the signal of the MtJSE transmission method. FIG. 9 is a diagram showing the internal configuration of the vertical low-pass filter in FIG. 7, FIG. 9 is a diagram showing the pixel arrangement for explaining the operation in FIG. 8, and FIG. 11 is a diagram showing the time axis phase during motion vector correction. It is a diagram. 1, 20.30... Input terminal, 2.14... Low pass filter, 3... AD converter, 4... De-emphasis circuit, 5... Control signal separation circuit, 6...
...Intra-field interpolation circuit, 7.26...1-frame delay device, 8...Inter-frame interpolation processing circuit, 9...Vertical low-pass filter, 10...Mixing circuit, 11...・
Scanning line conversion circuit, 12.27...Motion detection circuit, 13
...Subtractor, 15...Absolute value circuit, 16...TC
I decoding circuit, 17...OA converter, 18...N
TSC encoder circuit, 19.25.32... Output terminal, 21.22... Delay device, 23.24... Adder, 28... Sensitivity switching circuit, 29... Temporal filter, 31. ... MAX processor, 33 ... 1 field delay device, 3 4 ... arithmetic unit, l. lo. I5
62. j5e: +- signal, m1 to mco... frame number, n1 to n4... field number, A. B...Characteristics, α...Constant.

Claims (1)

[Scope of Claims] In a signal conversion device that receives and demodulates a MUSE signal obtained by band-compressing a high-definition television signal and converts it into a signal that can be received by a current television receiver, the MUSE signal is resampled using a clock signal. a de-emphasis circuit that de-emphasizes the output signal of the AD converter; a first one-frame delay device that delays the output signal of the de-emphasis circuit by one frame; an intra-field interpolation circuit that performs video-related processing by interpolating the output signal of the frame delay device within the field; an inter-frame interpolation circuit that performs still-image processing by interpolating the output signal of the de-emphasis circuit between frames; a control signal separation circuit that separates control signals such as motion vector data and sub-sample phase data from the output signal of the AD converter; and a control signal separation circuit that delays the sub-sample phase data by one frame and supplies it to the intra-field interpolation circuit; a second one-frame delay device configured to control the intra-field interpolation phase; a vertical low-pass filter for removing vertical frequency components in the vicinity of 1125 TV lines from the output signal of the inter-frame interpolation circuit; A mixing circuit that mixes and processes two signals, the output signal of the interpolation circuit and the output signal of the vertical low-pass filter, and a signal of 1125 scanning lines, which is the output signal of the mixing circuit, as a signal of the scanning line of the current television system. a scanning line conversion circuit for converting the signal into a scanning line conversion circuit, and obtaining an inter-frame difference signal from the output signal of the de-emphasis circuit and the output signal of the first 1-frame delay device, and converting it into an absolute value through a low-pass filter. detecting the amount of motion, switching the detection sensitivity of this amount of motion using the motion vector data, and controlling the mixing ratio of the two signals in the mixing circuit in accordance with the signal obtained by stretching in the time axis direction in units of fields. 1. A signal conversion device comprising a motion detection circuit.