JPH04196889A - Time base correcting device - Google Patents

Abstract

PURPOSE:To attain the correction of a frequency characteristic and a phase characteristic by providing a thinning filter and an additional control means to calculate the coefficient of the thinning filter based on output from a time base error detecting means and the characteristic information of the frequency characteristic, the phase characteristic and so on. CONSTITUTION:An input video signal is converted into a digital signal by an A/D converter 1 and after that, is written in a memory 3. Then, for the video signal read from the memory 3, a time base is shifted with the period of a reference clock as a unit to a reading synchronous signal. On the other hand, a coefficient control circuit 7 controls the coefficient of a thinning filter 4 corresponding to a time base error smaller than the period of the reference clock and for the output of the thinning filter 4, the time base is shifted within the period of the reference clock to the output of the memory 3. A synchronization adding circuit 9 adds the reading synchronous signal to the output signal of the thinning filter 4 and last, a D/A converter 2 converts the signal into an analog signal and outputs. Thus, a time base correcting device to attain also the correction of a frequency characteristic and a phase characteristic without the broad increasing of a circuit scale can be realized.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a time axis correction device suitable for use in home VTRs (video tape recorders) and the like.

[Conventional technology]

In a VTR, fluctuations in the time axis occur in the reproduced video signal due to uneven rotation of the cylinder or the like. In order to correct this time base fluctuation, various types of time base correction devices (hereinafter referred to as TBC) have been used.
tor) has been proposed. An example of this is a device described in Japanese Patent Laid-Open No. 64-36278. The outline of this device will be explained with reference to FIG.

Figure 7 shows the TBC described in Japanese Unexamined Patent Publication No. 64-36278.
This is a diagram showing the characteristic parts extracted from the circuit blocks of the device. In the figure, an input video signal is converted into a digital signal by an A/D converter 1, and after passing through a signal delay unit 14 and temporarily stored in a main storage unit 15, a time axis error is corrected in a linear addition unit 16. It is processed as follows, D
/A converter 2 converts the signal into an analog video signal. The time axis error detection section 17 detects the time axis error of the input video signal at the same time as generating an address clear signal for the main storage section 15, and the error correction signal storage section 18 detects the time axis error until the data is read from the main storage section 15. is stored, and the addition control unit 1
9 generates coefficients for the linear addition section from the time axis error. The synchronization signal generating section 20 supplies a basic clock to each section and at the same time generates a read horizontal synchronization signal to control the read timing of the main storage section 15 and the like.

The biggest feature in FIG. 7 is the linear addition section 16. That is, the linear adder 16 restores the input video signal from the digital signal sampled with the basic clock, and further samples it again to obtain a signal free of time axis errors.

Here, the principle of time axis error correction of the TBC configured as shown in FIG. 7 will be explained.

Sample value obtained by sampling the original signal g (t) at time interval T...5g (2T) t g (T) +g
(0), g (T), g (2T)+... are given, and if the band W of the original signal satisfies W<7, then
The original signal g (t) can be calculated using the following equation (sampling theorem).

g(shi) = Σ g(i T)S(t, -i T)
・(1) i! −o＞, then perform an operation equivalent to resampling,
..., g(2T-△T), g(-T-△T), g(-
△T), g (T-△T), g (2T-△T),...
It is possible to calculate a new sample value of
This is obtained by shifting the time axis by ΔT from the above sample value. Using this, for example, detect how much the input signal's IH (horizontal scanning period) start timing deviates from the IH start timing of a reference signal with no time axis error, and correct the deviation. It becomes possible to do so.

Also, it is possible to vary the amount of shift for each sample value, so if we detect the expansion and contraction of the input signal and calculate the time shift amount for each sample value to convert it to a normal length, , expansion and contraction of the input signal can also be corrected.

Based on the above principle, the time axis error detection section 17 in FIG. 2 detects the phase shift and period fluctuation of the input video signal in units of IH, and the addition control section 19 calculates the time for each sample value from these values. The time axis error is corrected by calculating the axis soft amount and further executing the above-mentioned operation to calculate the new sample value in the linear addition unit 16.

[Problem to be solved by the invention]

Incidentally, during playback on a home VTR, it may be necessary to correct frequency characteristics and phase characteristics. For example, there are processes such as emphasis/de-emphasis to reduce high-frequency noise in FM signal processing, an enhancer to emphasize the outline of an image, and an equalizer to improve waveform response. These processes may be due to differences in recording/playback systems (for example, differences in standards such as VH8 and 8mm video, or differences in modes such as SP and EP), differences in tape and head performance, or user preferences. It is necessary to change the characteristics accordingly.

In the conventional TBC, no particular consideration was given to correction of these frequency characteristics and phase characteristics, and the correction was performed using a separate circuit. Therefore, there was a problem that the circuit scale increased.

An object of the present invention is to provide a time axis correction device that can correct these frequency characteristics and phase characteristics.

[Means to solve the problem]

In order to achieve the above object, the present invention supplies information regarding frequency characteristics and phase characteristics to a coefficient control section, and the coefficient control section calculates coefficients of an interpolation filter based on a time axis error signal and characteristic information. .

[Effect]

The addition control unit calculates a comprehensive characteristic from the conventional time axis error signal and characteristic information, and determines coefficients in the interpolation filter.

The interpolation filter shifts the time axis and also corrects the frequency characteristics or phase characteristics.

This makes it possible to realize a time axis correction device that can also correct frequency characteristics and phase characteristics without significantly increasing the circuit scale.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of a time axis correction device according to the present invention, in which 1 is an A/D converter, 2 is a D/A converter, 3 is a memory, 4 is an interpolation filter, and 5 is a A write control circuit, 6 a read control circuit, 7 a coefficient control circuit, 8 a time axis error circuit, and 9 a synchronization addition circuit.

Although the circuit configuration in FIG. 1 is somewhat different from that of a conventional TBC, the basic principle of time axis correction is the same as that of the conventional TBC.

The input video signal is converted into a digital signal by the A/D converter 1 and then written into the memory 3. The write operation at this time is controlled by the write control circuit 5 so as to start writing based on the horizontal synchronization signal of the input video signal. The time axis error detection circuit 8 calculates the time axis error from the write start timing and the horizontal synchronization portion of the input video signal. By the way, the device shown in Fig. 1 operates based on a reference clock (not shown), or the time axis error in units of the period of the reference clock is supplied to the readout control circuit 6, and the time axis error smaller than the period is controlled by coefficients. It is supplied to the circuit 7. The readout control circuit 6 uses a synchronization signal (with no time axis fluctuation) obtained by dividing the reference clock (
A readout synchronization signal (hereinafter referred to as a readout synchronization signal) is generated, and the video signal is controlled to be read out from the memory 3 based on the readout synchronization signal. However, the readout timing is determined according to the time axis error in units of the period of the reference clock. shift. As a result, the time axis of the video signal read from the memory 3 is shifted in units of the period of the reference clock with respect to the read synchronization signal. On the other hand, the coefficient control circuit 7 controls the coefficients of the interpolation filter 4 in accordance with a time axis error that is eight times larger than the period.

As a result, the time axis of the output of the interpolation filter 4 is shifted with respect to the output of the memory 3 within the period of the reference clock. A synchronization addition circuit 9 adds a read synchronization signal to the output signal of the interpolation filter 4. Finally, the D/A converter 2 converts the signal into an analog signal and outputs it.

The principle of time axis error correction described above is the same as the conventional one, so its details will be omitted. The characteristics of this embodiment are the coefficient control circuit 7 and the interpolation filter 4.

These operations will be explained below.

The interpolation filter 4 has a transversal filter configuration as shown in FIG. In the same figure, 10-1~
10-n is a latch circuit, 11-1 to 11-n are multiplication circuits, and 12-1 to 12-(n-1) are addition circuits. The latch circuits 1O-1 to 10-n are connected in series, and use the reference clock as a latch pulse to sequentially delay the sample data of the signal read from the memory 3 by one cycle of the reference clock.

Multiplier circuits 11-1 to 11-n are latch circuits 10-1 to 10-1.
The outputs of 0-n are multiplied by the coefficients generated by the coefficient circuit 7. Addition circuits 12-1 to 12-(n-1) perform addition of multiplication circuits 11-1 to 11-n.

Here, as an example of the interpolation filter 4, the time axis accuracy is 7 (T
A case in which the time axis is shifted by (sampling period) will be explained. Sampling frequency f.

When restoring the original signal g (t) by interpolating the signal sampled at S at 4 fs, the interpolated data at these points can be obtained by substituting % in the above equation 4. . This will be explained with reference to FIG.

In the figure, QorQ4+Qs! indicated by O mark!・・・・・・
・ is sample data whose sampling frequency is fs, and is sample data Q. So, 1=0. When the original signal g (t) is restored in 4 fs using the above equation (1) for the signal sampled in this way, sample data Q is obtained. , Q4 between 1/4f s (=t/4
) at equal time intervals, the interpolation data work v q21 is indicated by the mark .
q3 is interpolated, and similarly sample data Q, ,Q,
Interpolated data Qs+qstq7 is interpolated in between, and in the same way, three pieces of interpolated data are interpolated at equal time intervals between each sample data.

Note that if the above equation (1) is calculated faithfully, the interpolated data will match the value of the original signal g (t), but since the sum of infinite terms is calculated, in reality, the error is within the allowable range. Generally, the number of terms in the above equation (1) is cut off to the extent possible, and the sampling function S (t) is also approximated to simplify the circuit configuration of the digital filter.

Further, when the sampling frequency fs is sufficiently large with respect to the signal band W (ie, fs>2W), there is no problem in practical use even if the sampling frequency is simply linear interpolation.

By the way, in a digital filter in which interpolation is performed as shown in FIG. 11, Qe+Q, q2゜qat Q4
If the data is not output in the order of 1 qat, etc., but is output at the time of 1/fs=T, then t=o, T, 2T.

...Sample data Q. l Q4t Q6f
Instead of ......, interpolation data work + qat qat should be interpolated with a delay of T/4 behind the sample data.
..., the signal sampled at the sampling frequency fs has been shifted (advanced) in time by T/4. Similarly, t=o, T
.

2T, ... sample day Qo+Q4+Qs+.
...Instead of interpolated data q2. q61 qyu. ,
When outputting ......, the time axis has been shifted by '/4T, and the interpolated data qay q7+ qll
When l... is output, the time axis has been shifted by j/4T.

In this way, it is possible to shift the input haiku name on the time axis using a digital filter, and the interpolation filter 4 shown in FIG. 1 is based on this principle, and FIG. 2 shows a specific example thereof.

By the way, it is known that the sampling function of equation (2) is also the output when an impulse is applied to a band-limited ideal LPF (low-pass filter). That is, when the above processing is expressed on the frequency axis, it becomes as shown in FIG. The original signal (a) is A
It is sampled by the /D converter 1, resulting in a spectrum as shown in (b).

In the processing in the interpolation filter 4, the sampling frequency is first quadrupled and a zero value is assigned to the added sample points. In this case, there is no change in the frequency spectrum and it remains as shown in (b). Next, when the ideal LPF (C) shown by the sampling function is applied, the spectrum (d) is obtained. Finally, when sampling is performed again at fs, the spectrum becomes the form shown in (b) again (however, as described above, the time axis is shifted before and after the interpolation filter). The important point here is that the LPF of ('C)
This is to prevent aliasing to the baseband that occurs during resampling. On the other hand, if the purpose can be achieved, the filter (c) may have the characteristics as shown in FIG. 5(a) or (C).

In this case, the output signal is the analog L after the D/A converter 2.
Since harmonic components are removed by PF (not shown), the result is as shown in FIG. 6(b) or (d).

The coefficient control circuit 7 controls the coefficients given to the interpolation filter 4 according to the characteristic information K so that the filter characteristics become as shown in FIG. 4(C), FIG. 5(a), or FIG. 5(c). do.

The specific configuration of the coefficient control circuit 7 is to use a microcomputer to calculate the coefficients from the frequency characteristics, or to assume several frequency characteristics in advance and store the coefficients in a ROM.
(Read Only Memory) or the like, and a method of reading out the coefficients according to the characteristic information K is conceivable.

According to the embodiment shown in FIG. 1, the interpolation filter 4 can simultaneously perform time axis shift and frequency characteristic correction, so that the time axis error can be reduced while correcting the frequency characteristic without significantly increasing the circuit scale. Can be corrected.

FIG. 6 is a circuit block diagram showing another embodiment of the present invention,
The difference from FIG. 1 is that a waveform detection circuit 13 is provided.

In this case, it is assumed that the input video signal includes a reference waveform signal in a vertical blanking period or the like.

For example, in the current clear vision broadcasting, OCR (GhostCanc) is used to remove ghosts.
el Reference) signal is inserted into the vertical blanking period. The waveform detection circuit 13 detects the OCR signal and compares it with an internal reference OCR signal.

Based on the difference, the amount to be corrected is determined and the information is supplied to the coefficient control circuit. The coefficient control circuit uses a microcomputer to determine the coefficients of the interpolation filter 4 so that the output signal has an optimal waveform. However, in this case as well, Figures 4 (C) and 5
It goes without saying that the characteristics should be such that the 2fs and 3fs components that become aliasing noise during resampling are sufficiently attenuated, as shown in Figures (a) or (c).

According to the configuration shown in FIG. 6, it is possible to optimally correct the waveform while correcting the time axis error.

〔Effect of the invention〕

According to the present invention, since the interpolation filter can simultaneously perform time axis shift and correction of frequency characteristics or waveforms, etc., the time axis correction device can correct frequency characteristics, waveforms, etc. without significantly increasing the circuit scale. can be configured.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the time axis correction device according to the present invention, FIG. 2 is a block diagram showing an example of the configuration of the interpolation filter in FIG. 1, and FIG. 3 is a diagram explaining the principle of the interpolation filter. FIG. 4 is a diagram explaining the principle of the interpolation filter on the frequency axis, FIG. 5 is a characteristic diagram of the interpolation filter, and FIG. 6 is a block diagram showing another embodiment of the time axis correction device according to the present invention.
FIG. 7 is a block diagram showing a conventional time axis correction device. 1... A/D converter, 2... D/A converter, 3... Memory, 4... Interpolation filter, 5... Write control circuit, 6... Read control circuit, 7...・Coefficient control circuit, 8
...Time axis error detection circuit, 9...Synchronization addition circuit, 1
3... Characteristic detection circuit. 1st distribution 2nd turn: 3 Kyoyu 吾between → t 4th no r conroshi λ article - Comparison 5 Haichishuu ni ni - Nenpi L

Claims (1)

[Claims] 1. In a time axis correction device that corrects time axis fluctuations of a video signal by digitizing the video signal, writing it to a memory, and reading it out, detecting the time axis fluctuation of the video signal. a plurality of latch circuits that delay the video signal read from the memory by a unit time; a multiplication circuit that multiplies the output of each of the latch circuits by a coefficient; and a sum of the outputs of the multiplier circuits. an interpolation filter consisting of an adder circuit that calculates; and an addition control means that calculates coefficients of the interpolation filter based on the output from the time axis error detection means and characteristic information such as frequency characteristics and phase characteristics. A time axis correction device characterized by: 2. The time-base error correction device according to claim 1, further comprising characteristic detection means for detecting frequency characteristics, waveforms, etc. of the video signal.