JPS6318888A - Processing device for high definition television signal - Google Patents

Abstract

PURPOSE:To obtain a still picture signal which is exactly motion-corrected by executing previously a motion-correction through controlling a memory address, and simultaneonsly setting a correction vector to zero, when the still picture is obtained by repeating the specified field of a high definition television signal with the motioncorrecting vector. CONSTITUTION:The memories 141-144 of a complement field part for storing the high grade television signal in which a sampling phase is variable to every field, and simultaneously, a sub-nyquist sampling taking one round in plural fields is executed, and which is trasmitted while containing the correcting vector to show the motion of a whole picture area, are provided. By means of an address controller 19 controlled in response to the motion-correcting vector detected by the detecting part 21 of the motion-correcting vector, the signal of the complement field part is repeatingly read-out from the memories 141-l44, and the still picture signal is obtained. At such a time, the address controller 19 is controlled to make the signal of every field fit the picture position of the specified field among the plural fields read-out repeatingly, and simultaneously, the motioncorrecting vector is set to come to zero.

Description

[Detailed description of the invention] The invention will be explained in the following order.

A. Industrial application field - B. Essentials of the invention C. Prior art D. Problem to be solved by the invention E. Means for solving the problem (Fig. 1) F. Effect G. Example G1 Description of processing device ( Fig. 1) G2 operation explanation (Figs. 2 to 6) Effects of the invention A Industrial application field This invention is applicable to, for example, the MUSE system (Multiple S
ub-Nyquist Sampling Enco
The present invention relates to a processing device for high-definition television signals transmitted using the dkng) method.

B. Summary of the Invention This invention produces a still image by changing the sampling phase for each field and repeating the same signal in four fields of a high-definition television signal that is sub-Nyquist sampled and transmitted, for example, in four fields. The decoder for high-definition television signals has been devised so that it can obtain correctly motion-compensated still images even if it operates in the same way as it would normally do to obtain a normal image along the flow of time. be.

C. Conventional technology ■ The MIISE method has been proposed as a transmission method for high-definition television consisting of 1125 horizontal lines per frame (preliminary paper for the founding anniversary concert, ``New transmission method for high-definition television (MtlSE) J, Yu Ninomiya -1 June 1982
(Refer to the announcement by the NHK Research Institute for Broadcasting Science and Technology on April 6th).

In this method, on the transmitting side, for example, a luminance signal Y such as neo as shown in FIG. The time axis is compressed to /4 and inserted into the horizontal blanking period at the front of the luminance signal Y in each corresponding horizontal section line-by-line to form a TCI signal as shown in the figure. This T
The CI signal is sampled at 64.8 MIIz as shown in Figure E of l1, and one sample is extracted every four samples of this sampling data, and sub-Nyquist sampling is performed in one cycle in 4 fields as shown in Figure F. . Therefore, the frequency of sub-Nyquist sampling becomes 16.2MIIz, which makes the original 20M
IIz baseband signal bandwidth is 8.1M)l
It can be compressed up to z. This signal is frequency modulated, for example, and transmitted via a broadcasting satellite or the like.

In this case, if an FM modulation method with constant carrier power is used for the transmission signal, the transmission band will be three times the baseband. By the way, the band per channel of satellite broadcasting is 27 MIIz. Therefore, baseband 20
If Ml1z's high-definition TV No. 14 were to be converted to FM and transmitted as is, two or more channels would be required, but by changing the signal to the MUSE system as described above, one channel transmission becomes possible.

On the other hand, on the receiving side, the received signal is demodulated and the above-mentioned 8.1M
Extract the Hz sub-Nyquist sampled signal. This signal is decoded as follows. That is, first, sample at 16.2 MHz to restore the sampling data as shown in Figure 7F, write this sampling data to the memory as shown in Figure 7 above, and then use these data and blanks to restore the sampling data. The data at the non-sampling points shown are converted and interpolated (between Jili) to restore the TCI signal as shown in Figure 7 above. Then, the luminance signal Y and color signal CN+CW as shown in FIG. 7A to C are extracted from this signal.

In this way, it becomes possible to transmit a high-definition television signal using the MtlSE method, for example, by one channel of satellite broadcasting with a bandwidth of 27 MIIz.

Figure 8 shows an example of a decoder device for MIISE television signals, in which (1) is an antenna that receives 12 GIIZ satellite broadcast signals, and (2) is an FM receiver. In this case, the satellite broadcast signal is FM demodulated to obtain a MUSE television signal MV.

This signal MV is 16.2M by A/D converter (3)
l1yt is sampled and converted into digital data DM. This digital data DM is supplied to a separation circuit (4) where synchronization signals, control signals, etc. are separated, and then supplied to a static area interpolation circuit (5) and also supplied to a frame memo (January 6) for writing. .

From frame memo January 6), the signal D'M of one frame before
is output, and this is supplied to the static area interpolation circuit (5).

Then, in the still area interpolation circuit (5), the signal DM of two frames, that is, the past three fields, and the signal DM of the current field are combined, and ll! ! A still image signal M V 1 with a sampling rate of 64°8 M Ilz is obtained through elemental interpolation.

The signal DM obtained through the separation circuit (4) is also supplied to a dynamic area interpolation circuit (7). In this moving region interpolation circuit (7), only the latest field signal DM is used to interpolate all pixels, and a moving image signal M V 2 with a sampling rate of 64.8 MIIz is obtained.

The still image signal MV+ from the still area interpolation circuit (5) and the moving image signal MV2 from the moving area interpolation circuit (7) are added by a W adder (8).

Further, (9) is a moving part detection circuit which detects the movement of the image pixel by pixel from the equal number DM from the separation circuit (4) and the signal of one frame before from the frame memo (January 6). The detection signal SD from this moving part detection circuit (9) is supplied to an adder (8), and the signals M V 1 and M V 2 are added at a ratio according to this detection signal SD. For example, in a still part of an image, the proportion of the signal M V 1 is large, while in a moving part of the image, the proportion of the signal M V 2 is large.

The addition signal M V o from this adder (8) is supplied to a TCI decoder (10);
), the primary color signals R, G, and B of red, green, and blue are obtained.
For example, it is supplied to a monitor receiver.

In this way, by performing different interpolation processes on the still image portion and the moving image portion, sufficient resolution can be obtained in the still image portion, and an image without multi-line blur due to time differences can be obtained in the moving image portion.

By the way, if the image is of something that the camera is banging, even if it is a still image, it can be combined with a video and an inverted adder (
The output signal M V o of 8) has a large proportion of the moving image signal M V 2 with poor resolution. In general, human visual perception is insensitive to motion (bokeh), but when the camera is banging, the entire screen moves smoothly, and not only does the eye easily follow it, but the viewer's consciousness also makes it difficult to stop. The blur is noticeable because there is a sense of being there.

In order to prevent this, it is necessary to perform motion compensation, that is, to shift the previous frame by the amount that the entire screen has moved.
). As a result, the still image portion is correctly treated as a still image, and sufficient resolution can be obtained.

The MIISE television signal takes this point into consideration, and a motion correction vector indicating the direction and magnitude of the movement of the entire screen at the time of banging, etc. is preset on the transmitting side after each of the four fields in Figure 7F. On the receiving side, this motion correction vector SB is separated in the separation path (4),
This controls the readout address of the frame memo (January 6), so that the signal from this memory (6) one frame before matches the positional relationship on the screen with the signal DM of the current frame from the separation circuit (4). The still image portion can be determined to be a still image.

D. Problems to be Solved by the Invention In FIG. It is conceivable to record the signal MV from 2), reproduce it, and supply the reproduced signal to the decoder section via the A/D converter (3).

In this case, during normal playback in which the signals recorded by the recording/playback device are played back in the original order, if the decoder section performs motion correction using the motion correction vector in exactly the same manner as described above, a reproduced image with less deterioration in resolution can be obtained. It will be done. However, in a recording/reproducing device, various variable speed reproduction can be performed, and if the same signal of 4 fields in which the subsampling phase of the MLISE signal goes around is repeatedly obtained as a recurrent signal, and this is decoded. Still images are obtained on both sides.

During this still image reproduction, the motion correction vector possessed by the first field at the beginning of the four-field signal that is repeatedly reproduced is for motion correction between the signal of the first field and the previous frame.

However, during still image playback in which these four fields are repeatedly played back, the decoder section uses the motion correction vector of the first field for the frames after the frames of the first and second fields (the third and fourth fields). It is determined that the vector is a motion correction vector, and erroneous motion correction is performed. In this case, the signal will be detected as a moving part, and the resolution of the reproduced still image will deteriorate.

It is an object of the present invention to prevent deterioration in the resolution of reproduced still images by making it possible to correctly perform motion correction during reproduction of still images.

In addition, a memory for 4 fields is provided in place of the recording/reproducing device, or a memory for 4 fields is provided at the output stage of the recording/reproducing device, and the signals of 4 fields for one specific round are written to the memory. A still image can be reproduced even if the writing is stopped and the reading is repeated repeatedly, and this case is also covered by the present invention.

E. Means for Solving Problems In this invention, the sampling phase is changed for each field, sub-Nyquist sampling is performed once in multiple fields, and the image is transmitted including a motion correction vector indicating the motion of the entire screen. Memory for the above multiple fields to store high-definition television signals
(141) to (144), this memory address controller (19), and a motion correction vector detection unit (21)
The address controller (19) is controlled according to the motion correction vector detected by the detection unit (21), and the memories (14L) to (144)
When obtaining a still image signal by repeatedly reading out signals for the plurality of fields, an address controller ( 19), and the motion correction vector is set to zero.

One cycle of a plurality of fields is repeatedly read out from the F working memories (14i) to (144) after motion compensation is performed by the address controller (19). Therefore, no motion compensation is required in the decoder at this time, but the motion compensation vector is zeroed out to achieve this purpose. Therefore, if the decoder decodes the signal in the same way as a normal signal, a correctly motion-compensated still image signal can be obtained.

G Embodiment G1 Description of Processing Apparatus FIG. 1 shows an example of the apparatus of this invention, and this example is an example in which this processing apparatus is installed in the recording/reproducing apparatus described above.

In this example, the recording/reproducing device converts the MUSE T C1 signal (8, IMIIz) into FM and records it in analog form. In IMIIz, the carrier frequency increases and the C/N deteriorates, so one horizontal section is time-stretched,
By adding a synchronization signal and a burst signal to the clock stream of the reproduction TBC (time base collector) and distributing it to two channels for each horizontal section, the video signal band per channel can be lowered to about 4.5M1lz. There is.

During playback, the signal is supplied to an FM decoupling circuit through a head amplifier and an equalizer, and is FM demodulated, and the jitter for each channel is corrected by TBC.

The signal jitter-corrected by 'r B C is time-base compressed, mixed for each horizontal section, and then MtlSE T C
No. 11d is played.

In FIG. 1, (11) is this reproduction MUSE 'r C
1 signal input terminal, and the reproduced MtlSIE signal through this is supplied to the A/D converter (12) and converted to 16.2 MHz as in the case of the A/D converter (3) in the example in FIG.
After being sampled and converted into a digital signal by
) to (144).

These field memories (141) to (144) are
Reading is performed sequentially for each field without overlapping with writing, and the read A word is supplied to a switch circuit (15) that is switched in synchronization with the reading, and from there the field is read in the original field order. The output signal of this switch circuit (15), which is taken out as a lined signal, is supplied to a control signal correction circuit (16), and after the control signal such as a motion correction vector is corrected as described later, it is output through a switch circuit (17). Terminal (18
) is derived.

The output of the A/D converter (12) is also
141) to (144), but directly to the output terminal (1B) via this switch circuit (I7).

In addition, (19) is a memory address and timing controller, which controls switching timing signals SW1 and SW2 of switch circuits (13) and (16), and chip selects for memories (14x) to (144) of 41. Signals CB1 to CE4, write control signals W1 to W
4 and address signals AD1 to AD4 are obtained from this and supplied to each circuit.

(21) is the MUSE T from the switch circuit (15)
The decoder extracts and decodes the control signal in the C1 signal, and the decoded output is supplied to a controller (19) to provide division in units of four fields and timing in units of fields. Further, the decoder (21) detects a motion correction vector included in the control 1-roll signal, and controls the address and read memory address from the timing controller (19) according to this motion correction vector. There is.

Further, (20) is a system controller that manages this memory address and timing controller (19).

G2-Operation Description The reproduction signal processing operation in the above configuration will be explained.

First, the MUSE signal and its control will be explained.

Figure 2 is an explanatory diagram of the sample point phase of the first to fourth fields of the MusE signal and the control signal.The numbers surrounded by circles in Figure A are the numbers surrounded by the circles for each pixel sample arranged in the horizontal line direction. indicates the pixel sample sent in that field among the pixel samples. In this case, the first and third fields are the odd-numbered internal pixel samples, and the second and fourth fields are the even-numbered tF pixel samples. , respectively, but the sample phase (sub-sample phase) is the same for the first field and the third field, and is the same for the second field and the fourth field. Therefore, as shown in FIG. 2B, information ``■'' and ``O'' indicating this sub-sample phase are added to the data of each field as a control signal. Further, FIG. 2C is a motion correction vector, where the numbers represent the amount of motion in terms of the number of pixel samples, and the arrows indicate the direction of motion.

Also, the second pattern is when the screen moves due to banning,
When the phases of the sample points transmitted in that field do not match, it is necessary to delay the entire sample of that field by one sample.Indicates whether the delay has been performed or not. OFF means no delay, and ON means there is a delay. It shows.

First, the reproduction signal processing when performing normal reproduction on a VTR will be explained.

In this case, the switch control signal from the system controller (20) causes the switch times 1i! 3' (17) is switched to the input terminal N side, and the ^/D converter (12
) is directly led out to the output terminal (18) through this switch circuit (17). In other words, during normal playback, the playback signal and control signal are sent directly to the MUS.
The E signal is supplied to the decoder circuit, where correct motion correction is performed, and the signal is returned to the RlG and B primary color signals.

- Note that this switch circuit (17) may not be provided and the following arrangement may be made. That is, the normal reproduction signal is sequentially written into the four field memories (141) to (144), and read out sequentially from a point delayed by one field, for example, and in the control fM correction stage II (16), the control signal is If you output it as is without making any modifications, the memory (141) to (
144) acts as a mere buffer memory, and the output terminal (18) simply receives a delayed normal playback signal. If the '1USIE signal is supplied to the decoder circuit, correct decoding processing will be performed.

In addition, at this time, the control signal decoder (21)
is disabled by the output of the system controller (20).

Next, to explain about still playback, at this time,
The switch circuit (17) is connected to the system commy roller (20
) is switched to the input end S side.

In this case, the signal state when there is no movement of the entire screen is shown in Figure 3, and the signal state is 1i for the first frame and the second frame.
FIG. 4 shows the signal states when the entire lj plane moves by one pixel to the right due to camera panning.

In this case, the first and third fields of the 4-field (2-frame) signal obtained by repeatedly binding the still image signal
The player represents the field.

883 Figure A and Figure 4 A show the camera output, respectively. In Figure 3 A, there is no movement, so the pixel positions of the first and third fields match, and in Figure 4 A, there is movement, so the pixel positions match. In the third field, it is shifted by one pixel to the right (in the delay direction).

FIG. 3B and FIG. 4B show sample sequences when this camera output is output as a MUSE T CI signal by the MUSE encoder, and in the figures, the pixel samples surrounded by the mouth are the transmitted pixel samples.

In FIG. 3B, there is no movement, so the motion correction vector is zero and the delay is also OFF. Further, the sub-sample phase is ■ for the first field and O for the third field.

On the other hand, in Figure 4B, there was a movement of one pixel in the camera output, so in order to match the sample phase, a delay of one pixel was applied to the third field, and the delay of the control signal of this third field was turned on. . Therefore, since the first field and the third field are shifted by two pixels, a rightward vector with a magnitude of 2 is added to the signal of the third field as a motion correction vector. Note that since the sample phases at this time match in the first field and the third field, both become ■.

Next, FIGS. 3C and 4C show the output states when the above-mentioned signals are repeatedly played back (still playback) for two frames each by the recording/reproducing device, and this signal is the same as that shown in FIG. The first field is supplied to the input terminal (11), the third field is supplied to the memory (141), and the third field is supplied to the memory (141).
(143), respectively.

When there is no movement, this reproduced signal (input signal) is directly output to the output terminal as shown in the third diagram.

On the other hand, when there is movement, the control signal decoder (
At step 21), a motion correction vector is detected, and the read address signal from the address and timing controller (19) is controlled by the detection signal. That is, as shown in FIG. 41, during the first four-field period, the input four-field signals are sequentially read out as they are from the memory, but during the next four-field period, the first four field signals are
Due to the motion correction vector of the third field detected when reading the field, the signal of the first field is read out with a delay of 2 pixel samples, which is equivalent to the case that there was no movement for the entire screen (in terms of the Zamble phase, the signal of the first field is read out with a delay of 2 pixel samples. Hereafter, in this state, there is no movement of the entire image between the first frame and the second frame, and the first to
The fourth field is read out repeatedly.

From the second time onwards, the control signal correction circuit (16) sets the motion correction vectors of all fields to zero and corrects them to a state where there is no motion, and the fAl field is corrected due to the delay. The delay information power (is turned ON and the subsample phase for this first field is modified to O accordingly to the sample phase in its delayed state).

In this way, a still image signal with corrected control signals such as motion correction vector, subsample phase, delay, etc. is delivered to the output terminal (18) via the switch circuit (17).

Therefore, if the signal obtained at this output terminal (18) is supplied to the decoder circuit, the decoded output will be as shown in Fig. 3E in the case of a still image signal with no movement, and as shown in Fig. 3E when there is movement and the corrected still image signal. In the case of an image signal, it is as shown in FIG. 4E. In other words, even if there is movement, in the second and subsequent repetitions of the repeated 4-field signal, the circuit in Figure 1 treats it as if there is no movement, and the motion correction vector is made zero, so the decoder Now, if you decode it in exactly the same way as during normal playback, you will get a still image with no resolution degradation. In this case, the delay in the decoder is set to "OF I" if it was ON, and
By turning on -1 in F' and shifting it by one pixel, the relative positional relationship of 1tiI elements between each field is restored to the original, so the decoder output is the control signal of the signal obtained at the output terminal (18). The delay is exactly the opposite.

Then, if this decoded output is supplied to a monitor receiver, the results will be as shown in FIGS. 3F and 4F, taking into account the delay.

Similarly, an example of the operation when the first and third fields move by two pixels due to camera panning is shown in Figure 5, and an example of the operation when the camera moves by three pixels is shown in Figure 6. It comes to show.

In the above example, the circuit shown in FIG. 1 is provided in the output stage of the VTR, but this circuit may also be provided in the manual section of the MUSE decoder. Also, as a method for obtaining a stationary Pi-time signal, without using a VTR, as mentioned above, a memory for 4 fields is provided, and 4 specific fields are stored in this memory and read out repeatedly. Good too.

Effects of the Invention According to the present invention, when obtaining a still image by repeating a specific high-definition television signal having a fiJJ correction vector, for example, 2 frames = 4 fields, it is possible to perform motion correction by setting memory addresses in advance. By controlling the motion compensation vector and zeroing the motion compensation vector, it is possible to supply a signal similar to a motionless signal to a high-definition television signal decoder. It is possible to obtain a signal with no deterioration in resolution even if there is a problem.

[Brief explanation of drawings]

Fig. 1 is a block diagram of an example of this invention device, and Fig. 2 is a block diagram of an example of the device of this invention.
An explanatory diagram of the USE signal, FIGS. 3 to 6 are diagrams for explaining the operation of the present invention, FIG. 7 is an explanatory diagram of the MIJS [ET CI signal, and FIG. 8 is an example of a receiving device for the old SIE signal. FIG. (14z) to (144) are field memories, (1
6) is a control signal modification circuit, (19) is an address and timing controller, (20) is a system controller, and (21) is a control signal decoder.

Claims (1)

[Scope of Claims] (a) A high-definition television signal whose sampling phase is changed for each field, subjected to sub-Nyquist sampling in one round in multiple fields, and transmitted including a motion correction vector indicating the movement of the entire screen. (b) an address controller for this memory; (c) a detection section for the motion correction vector; and (d) a motion correction vector detected by the detection section. (e) When obtaining a still image signal by repeatedly reading signals for the plurality of fields from the memory, specifying one of the plurality of fields to be read repeatedly; A high-definition television signal processing device, wherein the address controller is controlled to match the signal of each field to the screen position of the field, and the motion correction vector is set to zero.