JPS59186102A - Circuit for eliminating noise in sound signal - Google Patents

Abstract

PURPOSE:To eliminate noise by holding a level during the period of generation of noise produced at point of time of reproducing track changeover in advance by means of a signal level just before and adding an error signal corresponding to the signal level difference just before and after the period of generation of noise so as to interpolate the missing part in terms of area. CONSTITUTION:A reproducing signal (a) is given to a capacitor 5 through a switch circuit 2 controlled at the ON-period by a switching signal 1. A signal (b) holding the signal level just before the period of generation of noise is given to a mixing circuit 8. Since capacitors 6, 7 hold the signal level just before the OFF-period of switch circuits 3, 4, signals (e), (f) are given to a subtraction circuit 11 and its output signal (g) is given to a circuit 12. The switch circuit 12 is turned on only during the period of high level of the switching signal 1 so as to provide an error signal (h) to the mixing circuit 8. Thus, the mixing circuit 8 mixes the signals (b) and (h) so as to interpolate the missing part in terms of area during the period of generation of noise. An output of the mixing circuit 8 is given to an LPF9 and a reproducing signal from which noise is eliminated and with fidelity to the original signal is obtained at an output terminal 10.

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Field of Industrial Application The present invention relates to a device (such as a VTR) that alternately reproduces an audio signal from a recording medium recorded by frequency modulating the audio signal using a plurality of heads. The present invention relates to an audio signal noise removal circuit.

(B) Prior Art In recent VTRs, in view of the fact that it is difficult to guarantee sufficient sound quality with conventional audio signal tracks that are opened in the longitudinal direction of the magnetic tape, the video signal tracks are opened diagonally on the magnetic tape. Storing audio signals (frequency modulated audio signals) on signal tracks in a frequency multiplexed format is being considered. In a helical scan type VTR, adjacent tracks are alternately scanned using a plurality of heads (usually two heads). Therefore, periodic noise is unavoidable at the joints of the heads. In this case, the effect of the video signal can be relatively easily removed by setting the seam outside the frame of the display screen, but the effect of the audio signal is cyclic due to the nature of its information. This method cannot be adopted because it is not possible to tolerate the occurrence of intermittent parts.

Therefore, in the conventional method, by utilizing the fact that the noise generation period can be predicted from the switching timing of a plurality of heads, the signal immediately before the noise generation period is used instead of the transmission signal containing the noise generated by the discontinuity of the carrier wave during the noise generation period. A method of applying levels has been considered, but this method has the drawback of lacking fidelity of the reproduced signal. Note that this tendency becomes more pronounced as the signal frequency becomes higher since the noise generation period is constant.

Therefore, recently, in order to improve this fidelity, a method has been developed in which the interpolation signal is composed of an alternative signal having a level that connects each signal level immediately before and after the noise generation period with a straight line (linear approximation method).
has been proposed (Japanese Unexamined Patent Publication No. 176511/1983)
. However, this conventional example requires sophisticated circuit technology such as a delay circuit and a variable current source, and the number of elements inevitably increases. In particular, in delay circuits, unless the group delay characteristics are completely matched from low to high frequencies, new errors may occur in linear approximation, the noise generation period width may expand, and the interpolation time may increase. There will be a need to take it. Since approximation errors cannot be avoided in any approximation method, it goes without saying that the shorter the interpolation time, the more effective it is.

Figures 1 and 2 respectively show reproduced signal waveforms approximated by the two conventional methods described above, with solid lines (A) in both cases.
indicates the reproduced signal, and the part (A1)
8) indicates the original signal in the above portions (A1) ('B1), and period (T) indicates the noise generation period.

(c) Purpose of the Invention The present invention has been made with the above-mentioned points in mind, and provides an audio signal noise removal circuit that has a more versatile noise removal effect than the linear approximation method and has a simpler configuration. This is what we provide.

B) Structure of the Invention The present invention is characterized in that noise occurs immediately after the track switching signal, and that the time width of the noise generated due to discontinuity of the 7M carrier immediately after the 7M demodulator is very short and constant at several microseconds. Focusing on this, we did not use a fortune-telling filter for the cutoff frequency after FM demodulation, but carried out pre-holding for the minimum necessary time width, and calculated the approximation error caused by pre-holding immediately after the end of pre-holding by calculating the area. Interpolation is performed by compensating the signal and passing it through a low-pass filter.

According to the present invention, it is not impossible to set the interpolation time width to 15 μs or less, which is sufficiently shorter than one wavelength of the maximum audible frequency.
z), it is possible to perform interpolation with extremely few approximation errors.

FIG. 3 is a conceptual diagram of the present invention. In the figure, a shows the area-interpolated waveform before passing through the low-pass filter, and b shows the area-interpolated waveform before passing through the low-pass filter.

C2^ is an instruction pulse for designating the noise generation period, the period immediately following the noise generation period, and the period immediately after the noise generation period, which are created based on the reproduction track switching signal, respectively. Further, e is the waveform of an output signal obtained by passing the signal a through a low-pass filter. instruction pulse b,
If the high level periods of c are T, -)T, the area of the right triangle with T as the base is +T as shown in signal e.
It is equal to rectangle E, which has one side and the same height. In Fig. 3, by subtracting the signal level pre-held by the instruction pulse C and the signal held by the device by the instruction pulse change, and extracting only the period immediately after the above (the high level period of the instruction pulse 1). Since this corresponds to the error deleted (or increased) during the noise generation period, if this is added immediately after the signal held in advance by the instruction pulse b, the area becomes the original waveform as shown in waveform a. almost equal. Since the interpolation time is sufficiently small with respect to the maximum audible frequency, interpolation that is extremely faithful to the original signal can be performed by passing the signal a through a low-pass filter whose cut-off frequency is higher than the maximum audible frequency.

(E) Embodiment FIG. 4 is a circuit diagram of an embodiment of the present invention, and FIG. 5 shows signal waveforms at various parts thereof.

Input terminal 1+l is connected to an FM demodulator (not shown) which receives a frequency modulated audio signal formed by alternately splicing the outputs of multiple heads.
The demodulated signal, ie, the reproduced signal shown in FIG. 5a, is input. The portion α1)(N2) of the reproduced signal a is noise caused by discontinuity of the carrier wave that occurs when a plurality of head outputs are switched, and is generated only for a certain period every certain period (for example, one field period). In the figure, the period is shown to be extremely short compared to the reproduced signal to clarify the subsequent generation and collection of noise.

The reproduced signal a is transmitted through the first, second, and third switch circuits f, respectively.
21 [3) (attached to 41, and when each switch circuit is on, the momentary signal level is maintained by the first, second, and third capacitors (5) + 61 [7) attached to the output side of each switch circuit. That's what I do. First switch circuit (2
) output (FIG. 5B) is applied to a mixing circuit (8), where it is added with an error signal (FIG. 5H), which will be described later, to output an interpolation signal as shown in FIG. 5C. This interpolated signal is applied to a low-pass filter (9) whose cut-off frequency is higher than the maximum audible frequency, where the error signal is averaged and the signal shown in FIG. 5d is delivered to the output terminal (10).

5th output which is 2nd and 3rd switch circuit +31+41 output
The signals shown in θ and f in FIG. 5 are respectively applied to a subtraction circuit (Il), which outputs the signal shown in FIG. 5g as an output of the subtraction circuit. This signal (g) is applied to the fourth switch circuit (12), and the portion selected there is applied to the mixing circuit (8) as an error signal color.

The input terminal (1 (6) is connected to a first switching signal (Fig.
) is input. This switching signal corresponds to a so-called RF pulse for selectively deriving a plurality of head outputs. Therefore, in a two-head helical power scan type VTR, the state is changed for each field, and therefore, as mentioned above, each change causes discontinuity in the carrier wave, which causes periodic noise to be added to the audio signal output from the FM demodulator. Become.

This noise generation period is called a noise generation period (T').

The first switching signal (1) is applied to the first mono multivibrator (14+ (the mono multivibrator is hereinafter referred to as MM) and the second MJ151.The first MM+1
4] is selected so that its time constant is 1.5 times the noise generation period (g), while the second MMC+ (g) is configured to output the second and third switching signals shown in '' that match the noise generation period. There is.

The second switching signal (j) is applied to the third switching circuit (4) so as to turn off the switch during the period when the second switching signal (j) is at the level of (g)) is applied to the first switch circuit (2) and also to the logic circuit 0φ so as to turn off the switch during the high level period. This logic circuit (IQ) performs an exclusive OR operation of the second and third switching signals (j) and (k), and outputs the fourth switching signal shown in FIG. ) is applied to each of the second and fourth switch circuits +31 (+2), and during the high level period, the second switch circuit (3) is switched off, and the fourth switch circuit (switched on for one audience). I try to do this.

With the above configuration, the reproduced signal (a) applied to the input terminal (1) is applied to the first capacitor (5) through the first switch circuit (2) whose on period is controlled by the third switching signal (1). A signal (b) that maintains the signal level immediately before the noise generation period is applied to the mixing circuit (8).

Also, the second and third capacitors +61f? 2nd and 3rd
Since the previous signal level is held during the off period of each switch circuit [31[41], the signals shown in Fig. 4 e and f are applied to the subtraction circuit (Il), and the signal shown in Fig. 4 g is outputted from the subtraction circuit (Il). It is applied to the switch circuit (12). This fourth
Since the switch circuit (I2) is turned on only during the high level period of the fourth switching signal (1), the mixing circuit (8)
An error signal ←) is given to Therefore, the mixing circuit (8) mixes the signals (b) and (h), and acts to interpolate in area the portion deleted (or increased) during the noise generation period.

The output of the mixing circuit (8) is applied to a low-pass filter (9), and a reproduced signal from which noise has been removed and which is faithful to the original signal is derived at the output terminal (10).

(F) Effects of the Invention The present invention maintains the noise generation period that occurs periodically at the time of switching the playback track at the signal level of the signal level immediately before the noise generation period, and the period immediately following the noise generation period, The error signal between the signal level immediately before the noise generation period and the signal level immediately before the noise generation period is added to interpolate the missing (or increased) portion in the noise generation period in terms of area. , noise removal with extremely small approximation error noise is possible with a simple configuration.

[Brief explanation of the drawing]

FIGS. 1 and 2 are output waveform diagrams according to the first and second conventional methods. FIG. 3 is a conceptual diagram of the present invention. FIG. 4 is a block diagram of the configuration of an embodiment of the present invention, and FIG. 5 is a waveform diagram of each part in FIG. 4. Explanation of main drawing numbers +21 +31 f+] H...1st, 2nd, 3rd,
4th switch circuit, +51+81+71...first, second, third capacitors, (8)...mixing circuit, io)
...Subtraction circuit, (+4) (Iff) ...1st and 2nd MM, (1 (to) ... logic circuit, (9) ... low pass filter. Fig. 4 Fig. 5 - Shakuhachiko small J Tsuki Rokuhei 111 °□ ＾J So Nen Hari - Extension゛・, 7ha＼ ,, II 97 ・, 77゛・2./1, I'1゛,] 1 mouth 1.1 ・1-- --i; enemy, 5T, 5T ll "15-

Claims (1)

[Claims] m In an audio signal reproducing circuit that alternately reproduces a frequency modulated audio signal obtained by frequency modulating a carrier wave with an audio signal using a plurality of heads, a noise generation period in a demodulated signal caused by discontinuity of the carrier wave at the time of switching the reproduction track. At the same time, an error signal corresponding to the signal level difference immediately before and after the noise generation period is added to the period immediately after the noise generation period, and the deleted (or increased) portion based on the front retention is expressed as an area. An audio signal noise removal circuit characterized in that noise is removed by interpolation.