JPH0530383A - Waveform equalizer - Google Patents

Abstract

(57) [Summary] [Object] The present invention relates to a waveform equalizer that performs waveform equalization using a VITS signal, reduces the amount of residual error after waveform equalization, and also stabilizes waveform equalization. It is an object of the present invention to provide a waveform equalization device that improves the noise. A switching circuit 3 for switching an input of a waveform equalizing circuit 4 between an input television signal and an output signal of a VITS signal holding circuit 2, and a circuit 1 for detecting a change in the state of the input signal at an input stage of the switching circuit 4. And a memory 5 for input / output signals of the waveform equalization circuit 4, and a CPU 6 for fetching the signals of the memory 5 and performing waveform equalization calculation.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a waveform equalizer for performing waveform equalization using a VITS signal superimposed on a sampled television signal.

[0002]

2. Description of the Related Art MUSE (Multiple Sub-nyquist Sampling Encodin) developed as a high-definition transmission system
g) is sample value transmission, which is transmitted as an analog sample value sampled at 16.2 MHz. If the characteristics of the entire transmission system including the transmission system to the reception system satisfy Nyquist's distortion-free transmission condition, data can be reproduced without distortion by correctly resampling on the reception side. However, if the transmission path characteristics deviate from the Nyquist distortion-free transmission condition, the data cannot be correctly reproduced even if the resampling is performed at the correct timing. For this reason, it becomes necessary to correct the transmission path characteristics on the receiving side and perform waveform equalization so as to satisfy the Nyquist distortion-free transmission condition. In order to perform this waveform equalization, some waveform equalizers have already been developed. This method is roughly classified into a method for equalizing waveforms at a sampling rate equal to the transmission rate and a method for equalizing waveforms operating at twice the sampling rate. The former includes, for example, IEICE Spring National Convention, B-584,
Nakamura et al .: "Waveform Equalizer for HDTV Receivers" (198
9 years). For the latter, the National Conference of the Television Society 16-5 Iwadate et al .: "MUSE Decoder Built-in Waveform Equalizer" (1988), and IEEE Transactionson Cons
umer Electronics, Vol.36, No.3, AUGUST 1990 Noboru K
ojima, et al .; `` A WAVEFORM EQUALIZER FOR THE MUSE TR
ANSMISSION SYSTEM ”is available. The former uses a transversal filter operating at 16 MHz, which is equal to the transmission rate, to perform waveform equalization using data resampled at 16 MHz. If the Nyquist distortion-free transmission condition is satisfied in the sample value transmission, the received VITS signal is all 0 except the center value. If this is slightly deviated from the distortion-free condition, a constant signal is detected at a position other than the center. This is the waveform distortion itself. This state is shown in FIG. (A) is an ideal VITS signal waveform.
A mark "○" indicates a resampling phase at 16 MHz, and a mark "x" indicates a resampling phase offset by 32 MHz by one clock. (B) shows reproduction data when resampled at the correct phase, (c) shows only Δτ. This is the reproduced data when the sampling phase is shifted.
E (Mean Square Error) method or ZF
A constant algorithm such as (ZeroForcing) method is used to perform calculation on the time axis, the tap coefficient is sequentially corrected to finally obtain the optimum tap coefficient, and the transversal filter coefficient is set so that values other than the center become 0. Is to control. The output signal of the transversal filter is {Yk}, the reference signal is {Rk}, and the difference signal between the output signal of the transversal filter and the reference signal is {Ek}.
k} and the total number of taps is M + N + 1, the nth tap coefficient C {i} ⁽ⁿ⁾ of the transversal filter is MSE.
The method is modified based on the following (Equation 1) and the ZF method is based on (Equation 2). However, α and β are coefficients for determining the correction amount. Here, {Ek} is the residual error.

[0003]

[Equation 1]

[0004]

[Equation 2]

On the other hand, in the method of sampling at 32 MHz, which is twice the transmission rate, a transversal filter operating at 32 MHz is used, and the transversal filter is controlled so that the error with the ideal VITS signal interpolated at the 32 MHz rate is minimized. To do.

The operation of the conventional waveform equalizer will be described below with reference to the drawings. FIG. 5 shows an example of a conventional waveform equalizer. In FIG. 5, 20 is an A / D, 10 is a transversal filter, 6 is a CPU, and 22 is a waveform memory. The input MUSE signal is 1 by A / D20.
It is sampled at 6.2 MHz. The VITS signal inserted in the 216 to 316 samples of the 1st and 2nd lines of each frame of this MUSE signal is fetched into the waveform memory 22. The waveform distortion is detected from this VITS signal, and the tap coefficient of the transversal filter 10 is calculated and corrected by the CPU 6 based on the ZF algorithm of the following (Equation 2). Then, the calculation is repeated until the residual waveform distortion becomes sufficiently smaller than the preset value. This residual waveform distortion is treated as a residual error in the arithmetic expression of the algorithm. When the residual error {Ek} is detected, C {i} is updated based on (Equation 1). Therefore, {Ek}
Waveform equalization performance depends greatly on how to improve the detection accuracy. Therefore, it is necessary to add the VITS signals synchronously a sufficient number of times to improve the S / N. FIG. 6 shows a flowchart of a series of these processes.

In general, when the S / N of the input television signal is low, if the number of times of synchronous addition is insufficient, the residual error is not sufficiently reduced, but becomes lower than the noise signal level. In this case, it is impossible to distinguish the residual error from the noise, and as a result, this level becomes the waveform equalization limit.
Therefore, in order to further reduce the residual error amount,
For the above reason, the S / N must be improved by sufficiently performing the synchronous addition. However, as a matter of course, the processing time increases in proportion to the number of times of this synchronous addition, and as a result, the waveform equalization time increases. Further, since the television signal is re-loaded into the CPU every time the tap coefficient is corrected, the possibility of being affected by factors that make the system unstable such as divergence increases. When obtaining the optimum tap coefficient by successive corrections on the time axis, the tap coefficient oscillates due to the influence of noise and other disturbances, although it follows a certain procedure. The filter 10 diverges.

[0008] On the other hand, the inventors have previously proposed Japanese Patent Application No. 2-3.
In No. 00320, the VITS signal subjected to synchronous addition and DC offset removal processing is held separately from the input / output waveform memory of the waveform equalization means, and removed by adding another memory used for waveform equalization calculation. It was shown that the problems of time and stability can be remarkably improved. However, in this configuration, there is no means for knowing the change in the state of the input television signal input to the waveform equalizer during the period in which the memory is read out and waveform equalization is performed. That is, when the switching means for inputting the input television signal to the waveform equalizing means is switched by the switching means after waveform equalization, it is impossible to judge whether or not the switching is allowed. If the input television signal has changed from the initial state of the waveform equalization process, the output from the waveform equalization means may be distorted rather than the waveform equalization before switching.

[0009]

In order to reduce the residual error amount as much as possible and improve the waveform equalization performance in the configuration of the conventional waveform equalizer, the S / N is improved by performing sufficient synchronous addition. I have to go. On the other hand, increasing the number of synchronous additions results in lengthening the waveform equalization time. Further, since the correction is performed sequentially on the time axis, not only the waveform equalization time is increased, but also the influence of disturbance or the like is more likely to occur and the possibility of divergence increases.

The present invention solves the above-mentioned problems by reducing the residual error amount after waveform equalization as much as possible, and significantly shortening the removal time as compared with the conventional method of fetching a signal for each tap coefficient modification. At the same time, it is an object of the present invention to provide a waveform equalizer that can improve stability during waveform equalization.

[0011]

According to the present invention, in order to achieve this object, a VITS signal is synchronously added by a CPU, DC offset is removed, and then transferred to a RAM to switch an input to a waveform equalizing means. The output of is input to the waveform equalizing means for waveform equalization, and by additionally having a means for detecting a change in waveform distortion of the input television signal,
If there is no change in waveform distortion of the input television signal, the input television signal is switched to the waveform equalizing means.

[0012]

According to the present invention, in the above-mentioned structure, the signal is once added to the RAM after synchronous addition and DC offset removal processing.
Is repeatedly read and waveform equalization calculation is performed. This significantly reduces the processing time without fetching a signal for each successive modification of the tap coefficient, and improves the S / N by sufficiently taking the number of times of synchronous addition once. Be unaffected. Further, by detecting a change in waveform distortion of the input television signal, the input television signal is switched to the waveform equalizing means without error after the waveform equalization.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A waveform equalizer according to an embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram of a circuit configuration of a waveform equalizer in one embodiment of the present invention. In FIG. 1, 1 is a waveform distortion change detection circuit, 2 is V
ITS signal holding circuit, 3 switching circuit, 4 waveform equalizing circuit, 5 memory, 6 CPU. The operation of the waveform equalizer having the above components will be described.

Before the waveform equalization, the a side of the switching circuit 3 is selected, and the input television signal is input to the waveform equalization circuit 4. The input / output of the waveform equalization circuit 4 is fetched by the CPU 6 via the waveform memory 5. The CPU 6 repeatedly takes in the input television signal, performs sufficient synchronous addition, and performs S addition.
/ N is improved. After that, the DC offset is removed and the data is transferred from the CPU 6 to the VITS signal holding circuit 2.
After that, the switching circuit 3 selects the side b and repeatedly reads the output signal of the VITS signal holding circuit 2. This VI
The signal of the TS signal holding circuit 2 is taken into the CPU 6 in the same manner as when the input television signal was taken in first.
The CPU 6 controls the waveform equalization circuit 4 by making a difference from a reference signal that is internally held in advance and performing the calculations shown in (Equation 1) and (Equation 2). FIG. 4 shows a flowchart of the processing at this time. If it is assumed that N frames are required for synchronous addition, M times are the number of corrections, and L frames are required for residual error detection + waveform equalization calculation + tap coefficient correction, the conventional method and the configuration of the present invention are used when all processing is performed. Comparing with, the processing time ratio α is as follows (Equation 3).

[0015]

[Equation 3]

Here, as a general value, synchronous addition is 64
Considering a case where the number of frames and the number of corrections is 20 and the residual error detection + waveform equalization calculation + tap coefficient correction is one field, α becomes 1/17, which is significantly reduced.

On the other hand, during the period in which the switching circuit 3 is switched to the b side and the waveform equalization operation is performed, since the signal taken into the CPU is not the input television signal, the state of waveform distortion of the input television signal, In other words, I don't know if the signal is changing. After the tap coefficient is corrected M times and the waveform equalization is sufficiently performed, when the input television signal is switched to the waveform equalizing means, the waveform distortion change detection circuit 1 detects the signal change of the input television signal. If there is no change, there is no problem in switching as it is if there is no change. If the input television signal changes, the CPU 6 may be reset and the process may be restarted from that point.

Specific embodiments of the present invention will be described below with reference to the drawings. FIG. 2 is a block diagram showing a specific embodiment of the present invention. In FIG. 2, 10 is a transversal filter, 11 and 12 are memories, 13 is a correlation circuit, 14 is a determination circuit, and 15 is a VITS signal sampling circuit.

The VITS signal sampling circuit 15 is a VITS
A gate pulse signal is generated while the signals are superimposed. The input television signal is input to the transversal filter 10 with the c side selected by the switch 3, and the input / output signal is written in the memory 5. After that, the memory 5 is read by the CPU 6 to perform synchronous addition and DC offset removal processing. All of these are processed by software inside the CPU. After performing the series of processes a required number of times, the data is transferred to the memory 11. After the transfer is completed, the switch 3 is switched to the d side, the memory 11 is read and the signal written by the CPU 6 is repeatedly output, and the output of the memory 11 is input to the transversal filter 10. After being transferred to the memory 11 in this way, the signal of the memory 11 can be used to process the signal which has already been subjected to synchronous addition and DC offset removal processing each time.
The processing time can be shortened, and stable operation can be performed as if a fixed signal source is used.

In parallel with the waveform equalization processing, the correlation circuit 13
Performs a correlation operation between the VITS signal of the current frame and the VITS signal of the previous frame output from the memory 12,
The correlation output is compared with the threshold level preset by the decision circuit 14. This determination result is sent to the CPU 6
To determine the change of the input signal. As a result, when it is determined that there is no change in the input signal, the switch 3 is switched to the c side.

As described above, according to the waveform equalizer of the embodiment of the present invention, the VITS signals are synchronously added and DC offset removed by the CPU, stored in the memory, and the memory contents are input to the transversal filter. Since the waveform equalization is performed by using the waveform equalization, the processing time can be shortened and the waveform equalization can be stably performed. Since the change in the input signal is detected in parallel with the equalization processing, the input television signal can be input to the waveform equalizing means when there is no change in the input signal.

[0022]

As is apparent from the above embodiments, according to the present invention, after the synchronous addition and DC offset removal processing is once performed, this signal is switched to the input television signal to be used, thereby significantly reducing the processing time. And the stability at the time of waveform equalization can be improved. By detecting a change in the input signal during this processing period, it is possible to provide a waveform equalizer capable of switching the input television signal to the waveform equalizing means after the waveform equalization without any problem.

[Brief description of drawings]

FIG. 1 is a schematic block diagram of a waveform equalizer according to an embodiment of the present invention.

FIG. 2 is a block diagram of a specific embodiment of the waveform equalizer of the present invention.

FIG. 3A is a signal waveform diagram showing how the ideal VITS signal is resampled, and FIG. 3B is a reproduced VIT in the case of a correct resampling phase.
S signal waveform diagram (c) Reproduction VIT in case of deviated resampling phase
S signal waveform diagram

FIG. 4 is a flowchart of ghost removal processing according to the present invention.

FIG. 5 is a block diagram of a conventional waveform equalizer.

FIG. 6 is a flowchart of conventional ghost removal processing.

[Explanation of symbols]

 1 Waveform distortion change detection circuit 2 VITS signal holding circuit 3 Switching circuit 4 Waveform equalization circuit 5 Memory 6 CPU 10 Transversal filter 11 Memory 12 Memory 13 Correlation circuit 14 Judgment circuit 15 VITS signal sampling circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI technical display location H04N 11/08 9187-5C

Claims (1)

Claim: What is claimed is: 1. A waveform equalizing device for performing waveform equalization using a VITS signal superimposed on a television signal having a sampled value transmitted, means for performing waveform equalization, and the waveform and the like. Means for switching the input of the equalizing means between an input television signal and an output signal of the RAM; means for detecting a change in the state of the input television signal at the input stage of the switching means; and an input signal for the waveform equalizing means, An input / output waveform memory that captures an output signal and a CPU that performs synchronous addition of VITS signals, DC offset removal processing, and waveform equalization calculation together so that the signals of the waveform memory can be captured and waveform equalization calculation can be performed, After the CPU has sent a signal to which synchronous addition and DC offset removal processing have been performed from the CPU, the switching means outputs the signal to the waveform equalization means. To perform waveform equalization processing by inputting to, and switching to input the input television signal to the waveform equalization means only when it is determined that there is no signal state change in the state change detection means of the input television signal, A waveform equalizer that performs processing from the beginning when it is determined that there is a change in the input signal.