Inversion Phenomenon Preventing Circuit

1. Field of the Invention

The present invention relates to an inversion phenomenon preventing circuit of a video recorder/reproducer in which a video signal is stored and reproduced through frequency modulation, and more particularly to a circuit for recovering the absence of zero crossing in an FM signal reproduced from a storage medium in the video recorder/reproducer.

2. Prior Art

FIG. 15 is a block diagram for illustrating a reproduction system of a conventional video signal storage/reproduction apparatus. A very small FM signal, reproduced from a magnetic tape 29 by means of a video head 30, is amplified by an amplifier 31 and is corrected frequency response thereof by an RF equalizer 32. The FM signal from the equalizer 32 is amplitude limited by a limiter circuit 33 and is frequency demodulated by an FM demodulator 34. Then the demodulated signal is fed to a de-emphasis circuit 35 to produce a video signal.

Generally, a normal video signal, as depicted by a waveform shown in FIG. 19b, which varies from a black level to a white level, produces a picture on a CRT as depicted in FIG. 19d. An abnormal video signal, as depicted by a waveform shown in FIG. 19a distorted by so-called "inversion phenomenon," will cause a picture as depicted by a illustration of FIG. 19c. This is explained as follows.

In the video storage/reproduction apparatuses, the lower side band of the FM signal is emphasized due to electromagnetic transducing characteristics. The limiter circuit 33 in FIG. 15 functions to detect the zero crossing points of the signal therethrough. If an absence of the zero crossing points exists in the signal, then an abnormal FM signal will appear at the output of the limiter circuit 33. When the limiter output is demodulated, an abnormal video signal having an abrupt drop in level will be produced as depicted by the waveform in FIG. 19a.

One prior art inversion phenomenon preventing circuit for preventing the inversion phenomenon is a circuit called DL-FM method described on p 98-100 in "NHK home video techniques" written by Katsuya Yokoyama, published by Nippon Hoso Shuppan Kyokai. FIG. 16 is a block diagram of such an inversion phenomenon preventing circuit according to the DL-FM method. FIG. 18 is a waveform diagram for showing waveforms at various points of the block diagram in FIG. 16.

A normal FM signal (i) in FIG. 18 is an original FM signal to be stored, in which points A to J indicate the time at which the zero crossing points are supposed to exist. The signal (i) is more emphasized at the lower side band thereof than the upper side band thereof, as shown by a signal (ii) due to lower side band emphasis effect encountered when the signal (i) passes through the electromagnetic transducing system. It should be noted that the reproduced FM signal (ii) is missing zero crossing at points E and F. This absence of the zero crossing points will cause the previously described inversion phenomenon. In the mean time, a signal (iii) exiting through a high pass filter, or HPF 36 has been emphasized upper side band thereof, there being no absence of the zero crossing at points E and F. The signal (iii) is amplitude limited by the limiter 37 to thereby produce a signal (iv) in FIG. 18. However, it should be noted that the signal (iv) has its zero crossing points at positions in time axis slightly shifted from those of the original FM signal (i). The shift in zero crossing points is also caused by another source which will be described later. The shift in zero crossing points is a source of deteriorated SN ratio in the video signal when the FM signal is demodulated. The signal (ii) is also input to a low pass filter or LPF 39 and is emphasized its lower side band, resulting in a signal (v) which has a very little shift in zero cross points. The signal (iv) and the signal (v) are aligned by an equalizer 38 so that their phase is equal with respect to each other and are then added by an adding circuit 40, thereby producing a signal (vi). The signal (vi) now has less shift in zero cross points caused by noise as compared to the signal (iv), indicating that SN ratio of the video signal obtained will be better in demodulating the signal (vii) than in demodulating the signal (iv). Having a deteriorated SN ratio thereof by the HPF 36 that suppresses lower frequencies, the signal from the equalizer 38 is added by the adding circuit 40 to the signal (v). Thus adding these two signals places a limitation on the improvement of SN ratio of the reproduced video signal, though the circuit arrangement in FIG. 16 can provide better SN ratio than a circuit arranged only with the HPF 36 and the limiter 37. In other words, if recovery of the missing zero crossing points in the reproduced FM signal is a first consideration, then the SN ratio of the video signal will be poor; conversely if improving the SN ratio is a primary importance then the missing zero crossing points will not be adequately recovered.

The shift in zero crossing points is also caused if the frequency response characteristic of an FM transmission line is not flat. The frequency response of the FM transmission system varies, particularly in the DL-FM method, in accordance with the ratio of the level of the output from the LPF 39 to the level of the signal through the HPF 36 and the limiter circuit 37, thus causing shift in zero crossing points. Such a shift in zero crossing points is a source of changes in the frequency characteristic of the demodulated video signal which in turn causes a distorted video signal. Also, a change in the level ratio of the signals to be added causes a change in effect of suppressing the inversion phenomenon, which in turn causes a change in the frequency response characteristic of the demodulated video signal and distortion thereof.

An object of the present invention is to provide an inversion phenomenon preventing apparatus in which zero cross points can be sufficiently recovered, further the deterioration of the SN ratio of an FM demodulated signal can be minimized.

In the invention, an FM signal, which is partly missing zero crossing points thereof, is supplied to an input terminal, to which a second polarity discriminating means and a lower side band suppressing means followed by a first polarity discriminating means are coupled together. A pulse extracting means receives the outputs of the first polarity discriminating means and a second polarity discriminating means to produce pulses corresponding only to portions of the FM signal at which the zero crossing points are absent. The pulses thus obtained are combined in a waveform reforming means with the FM signal from the input terminal, thereby causing the FM signal to regain the zero crossing points once lost.

The pulse extracting means may further comprise an integration means for integrating the pulses to produce signals having amplitude substantially proportional to pulse widths of the pulses. The waveform reforming means may be constructed by a comparator having a threshold level that varies in accordance with the amplitude of the output of the integration means. Thereby, the waveform reforming means produces an FM signal that has optimally regained the missing zero cross points in accordance with the amount of shift of the zero crossing points from the time at which the FM signal is supposed to cross the zero level.

An apparatus of the invention may conveniently comprise a polarity control means for converting the pulses from the pulse extracting means into pulses having polarity in accordance with polarity of the portion of the FM signal in which the zero crossing points are absent.

Features and details of the invention will be apparent from the following description of specific embodiments with reference to the accompanying drawings in which:

FIG. 1 is a block diagram for illustrating a first embodiment of an inversion phenomenon preventing apparatus according to the invention;

FIG. 2 is a waveform diagram for showing waveforms at various points in FIG. 1;

FIG. 3 shows a modification of the first embodiment;

FIGS. 4a and 4b show the characteristics of a slicer in FIGS. 1 and 3;

FIG. 5 illustrates a second embodiment of the invention;

FIG. 6 shows an example of the second embodiment using a sine-type filter as a lower side band suppressing means;

FIG. 7 shows a third embodiment of the invention;

FIG. 8 is a waveform diagram for showing waveforms at various points in FIG. 5;

FIG. 9 is a waveform diagram for showing waveforms at various points in FIG. 7;

FIG. 10 shows another example of the third embodiment;

FIG. 11 shows a fourth embodiment of the invention;

FIG. 12 illustrates details of the signals in the embodiment shown in FIG. 10;

FIG. 13 illustrates waveforms at various points in FIG. 11;

FIG. 14 shows how pulses are converted into signals amplitudes of which are proportional to the pulse widths thereof;

FIG. 15 is a block diagram for showing a conventional video recorder/reproducer;

FIG. 16 is a prior art inversion phenomenon preventing system;

FIG. 17 shows a digital type inversion phenomenon preventing apparatus realized by combining the BPF of the sine-type filter in FIG. 6 with the pulse pick-up circuit formed of the comparator, the logic gates and the delay circuit;

FIG. 18 is a waveform diagram for showing waveforms at various points of the block diagram in FIG. 16;

FIGS. 19a and FIG. 19c show a normal video signal and its sample picture on a CRT, respectively;

FIG. 19b and FIG. 19d show an abnormal video signal and its sample picture on the CRT, respectively;

FIG. 20 shows a fifth embodiment of the invention;

FIG. 21 shows a specific embodiment of the polarity control circuit;

FIGS. 22 and 23 show specific embodiments of the polarity decision circuits 522,523;

FIG. 24 shows a block diagram of an example of an inversion phenomenon detecting & pulse adding circuit;

FIG. 25 shows a sixth embodiment of the invention;

FIG. 26a shows polarity discriminating circuits preceded by a delay circuit and a high pass filter, respectively;

FIG. 26b shows an example of a lower side band suppressing means using sine-type filters.

FIG. 27 illustrates waveforms at various points in FIG. 20;

FIG. 28 shows an example of the fifth embodiment to which an equilizer has been added in FIG. 20;

FIG. 29a shows an example of combining various embodiments;

FIG. 29b shows an arrangement when the polarity decision circuits are constructed as shown in FIGS. 22 and 23;

FIGS. 30a (1)-(2) and 30b (1)-(3) show circuit examples of a slicer; and

FIG. 31 and FIG. 32 show another examples of the arrangement in FIG. 29b using the subtracter and the adder instead of the exclusive OR circuits.

FIG. 1 is a block diagram for illustrating an embodiment of an inversion phenomenon preventing apparatus according to the invention.

An HPF 111, and a second limiter 117 are coupled together at inputs thereof to an input terminal 120. The output of the HPF 111 is connected to a first limiter 112, which in turn is connected to a first equalizer 113. The output of the equalizer 113 is connected to one input terminal of a subtracter 119 and the output of the second limiter 117 is connected to the other input terminal of the subtracter 119. In the subtracter 119, the signal designated by a negative sign (-) is subtracted from the signal designated by a positive sign (+). The output of the subtracter 119 is fed to one of the input terminals of an adder 110 through an LPF 114, a slicer 115 and, a second equalizer 116. To the other input terminal is supplied the signal at the input terminal 120. The output of the equalizer 116 is supplied to the adder 110. The adder provides an output to a subsequent third limiter 118.

The operation of the first embodiment will now be described as follows with reference to FIG. 2.

A waveform 101 in FIG. 2 represents a normal FM signal. An FM signal supplied to the input terminal 120 has partially lost zero crossing points at time interval x as shown by a waveform 102 in FIG. 2. The FM signal is supplied to the second limiter 117 to be converted into a waveform 103 as shown, where the zero crossing points are still missing but amplitude is limited at a level L5 and L6. In the mean time, the waveform 102 is also input to the HPF 111 where the zero crossing points absent in the waveform have now been recovered as depicted by a waveform 104. The waveform 104 is fed to the first limiter 112 where the waveform has limited its amplitude at levels L7 and L8 as depicted by a waveform 105. The waveform 105 is supplied to the equalizer 113 to produce a signal sh, the phase of which has been to be in phase with the output of the second limiter or the waveform 103. The subtracter 119 subtracts the waveform 103 from the signal sh to thereby provide the difference between the two. The output of the subtracter is directed to an LPF 114 and then to an equalizer 116 via a slicer 115. Since the signal sh has poor SN ratio, the output of the subtracter has also poor SN ratio. The LPF 114 reduces high frequency components beyond the deviation of the FM signal that cause the poor SN ratio and subsequently the slicer 115, having characteristic in FIG. 4a or FIG. 4b, drops off the reduced high frequency components. Thus, this results in a waveform 106, having a component corresponding only to the portion of the waveform 102, where the zero crossing points are missing.

The output of the slicer 115 is aligned by the equalizer 116 to be in phase with the waveform 102 and is then fed to the adder 110. The waveform 106 is then added to the waveform 102, thereby producing a waveform 107 in which the zero crossing points have been regained at time interval x. A third limiter 118 provides a waveform 108 having its amplitude limited at limiting levels L9 and L10.

In this manner, recovery of the zero crossing points is achieved by producing signal components corresponding only to the portions in the FM signal at which the zero crossing points are absent and then adding the signal components to the FM signal. Thus the SN ratio of the video signal reproduced remains unimpaired while recovery of the zero crossing points is adequately carried out.

Which one of the two inputs to the subtracter 119 is leading in phase relative to the other is up to individual design of the respective blocks. The equalizer 113 serves to align phase relation between the waveforms supplied to the subtracter 119 and to compensate for level frequency response in the transmission line. Thus the equalizer 113 must be placed as shown in FIG. 3 if total amount of delay time of the HPF 111 plus the first limiter 112 is greater than that of the second limiter 117. Likewise, the equalizer 116 may also be placed as shown in FIG. 3.

A band pass filter or BPF may be used in place of the HPF 111, providing that frequencies within the deviation can be passed. Further, the LPF 114 or the slicer 115 may be omitted provided that SN ratio in the video signal reproduced from the FM signal that has regained the missing zero crossing points is not significantly deteriorated.

While FIG. 2 illustrates only a case in which the zero crossing points are absent on the negative swing of the incoming FM signal, the absence of the zero crossing points on the positive swing can also be recovered in like manner, the waveform 106 being a negative value.

A second embodiment of the invention will now be described with reference to FIG. 5 and FIG. 8. FIG. 5 illustrates an arrangement of a second embodiment of the invention. FIG. 8 shows waveforms at various points in FIG. 5. A waveform 201 in FIG. 8 is a normal FM signal in which zero crossing points are not absent. An FM signal reproduced by the video signal storage/reproduction apparatus is missing zero crossing points at, for example, points E and F as depicted by a waveform 202 in FIG. 8. The waveform 202 is supplied to a BPF 211 and a delay circuits 213 and 218. A waveform 203 is produced by the delay circuit 213 and a limiter 214 with the zero crossing points being shifted very little from the waveform 202. It should be noted that demodulating the waveform 203 will produce a video signal having good SN ratio but at the same time will cause inversion phenomenon in the video signal.

In the mean time, the BPF 211 produces a waveform 204 with all the zero crossing points thereof being present but with the location of the zero crossing points being somewhat shifted from those of the waveform 202. The signal 204 is fed to a limiter 212 to produce a waveform 205. A subtracter 215 subtracts the waveform 205 from the waveform 203 to produce a waveform 206. Narrow pulses 206a are caused by the shift in zero crossing points due to noise while a wider pulse 206b, between the points E and F, is caused by the absence of the zero crossing points at the points E and F. The pulse 206b is of a pulse width nearly equal to the reciprocal of the instantaneous frequency between the points E and F. Thus the waveform 206 is a signal in which the pulses due to noise and the pulse due to missing zero crossing points are coexisting. Thus extracting only the pulse 206b out of the waveform 206 will recover, without causing deterioration of SN ratio, the zero crossing points which have been absent in the reproduced FM signal 202. An LPF 216 and a slicer 217 are circuits for extracting the pulse 206b. The LPF 216 has an integration characteristics by which, for example, pulse signals such as (L), (M), (H) shown in FIG. 14 are integrated to produce pulse waveforms (O), (P), (Q) in FIG. 14, amplitudes h1 to h3 of which being in proportion to the pulse widths p1 to p3. Thus the LPF 216 converts the signal 206 into a signal 207 in which narrow pulses 207a have been sufficiently decreased in their levels. The slicer 217 picks up only pulses having greater amplitudes than a predetermined level (shown by a dashed line L1 and L2) from the signal 207. The pulse picked up by the slicer 217 is subtracted by a subtracter 219 from the signal 202 exiting through the delay circuit 218, and is then supplied to a limiter 210 to produce a waveform 208 as shown in FIG. 8.

The waveform 208 is an FM signal in which zero crossing points absent at the points E and F in the waveform 202 have been recovered. Unlike the prior art method, an entire FM signal having poor SN ratio is not added to at the signal 202 but only the pulse 207b. Thus not only the SN ratio of the video signal reproduced from the waveform 208 remains unimpaired but also the frequency response of the video signal will not be altered, while at the same time greatly improving effect of suppressing inversion phenomenon. If the BPF 211 is of a sine-type filter that exhibits good phase linearity, the delay circuit 213 is not needed and the inversion phenomenon preventing apparatus can be arranged as shown in FIG. 6. The sine-type filter 219 in FIG. 6 is of a filter similar to an FIR filter and consists of delay circuits 219a, 219b, scalers 219e, 219c, 219d and an adder 219f. The frequency response of the sine-type filter 219 can be changed at will by changing the scaling factors as well as phase adjustment of the filter may be made easily.

In FIG. 6, the adding circuit 221 may be replaced by a subtracter depending on design choice of the polarities, positive or negative, of scaling factors of the sine-type filter 219 and the polarity of the two inputs of the subtracter 215. This is, however, providing that the signals are properly added or subtracted so that zero crossing points can be recovered using the signal 207. Only a single delay circuit is required if the delay circuits 219a, 219b are to be arranged by a delay circuit of reflection type.

A third embodiment will now be described as follows with reference to a block diagram in FIG. 7 and waveforms shown in FIG. 9.

A waveform 301 in FIG. 9 is the same as the normal FM signal 201 in FIG. 8. A waveform 302 is the same as the waveform 202 in FIG. 8. The waveform 302 is supplied to a BPF 311 and then to a comparator 323 to provide a waveform 304 in which the zero crossing points, absent at points E and F of the waveform 302, have been recovered with little shift in time axis from the waveform 302. The signal that has passed through the delay circuit 313 and a comparator 322 is of a waveform 303 in which the zero crossing are absent at points E and F. Then the signals 303 and 304 are EX-ORed by an EX-OR circuit 325 to produce a waveform 305. The signal 305 is supplied to an LPF 316 to produce a waveform 306, which in turn is sliced by a comparator 324 at a threshold level indicated by a dashed line L3. The output of the comparator 324 is of a waveform 307 in FIG. 9 and is EX-ORed together with the signal 303 from the delay circuit 318 by an EX-OR circuit 326, thereby providing a waveform 308 with the zero crossovers being recovered at points E and F. In this matter, digital processing can eliminate the limiters 210 in the embodiments shown in FIG. 5 and 6. The LPF 316 and comparator 324 in FIG. 7, for extracting pulses having wider pulse widths, may be replaced by a delay circuit 327 and an AND circuit 328 as shown in FIG. 10.

FIG. 12 illustrates details of the signals in the embodiment shown in FIG. 10. A waveform 305 in FIG. 12, which is the same as the waveform 305 in FIG. 9, is delayed T seconds by the delay circuit 327, thereby resulting in a waveform 309. Normally, the delay time T is nearly equal to 1/2 of the reciprocal of the highest instantaneous frequency of the reproduced FM signal. The waveform 309 and the waveform 305 are ANDed by an AND circuit 328 to produce a waveform 310 in FIG. 12 which has a pulse width much wider than the delay time T. The described delay circuit 327 the AND circuit 328 for extracting pulses having widths wider than the time T seconds is only exemplary and may also be arranged using a one shot multi-vibrator or a flip flop. Optimum delay time of the delay circuit 318 in FIG. 10 is the sum of the propagation delay time of the EX-OR circuit 325 and the AND circuit 328 and T/2.

While the zero crossing points are recovered by addition and subtraction i.e., a linear operation in the first embodiment in FIG. 1 and the second embodiment shown in FIGS. 5 and 6, and by Boolean algebraic operation in the third embodiment shown in FIGS. 7 and 10, the zero crossing points may also be recovered by dynamically changing the threshold level L3 in the above described third embodiment, which will be described as follows.

FIG. 11 shows a fourth embodiment of the invention, which is of the same arrangement as the second embodiment shown in FIG. 5 except for the adding circuit 219 and the limiter 210. Thus detailed description will be omitted. FIG. 13 illustrates waveforms at various points in FIG. 11. The output of the slicer 417 will be a waveform 402, the same as through the slicer 217 in FIG. 6 of the second embodiment, which is supplied as a threshold level to the comparator 420. The waveform 401, aligned in phase with the waveform 402 by the delay circuit 418 is also supplied to the comparator 420, thus producing a waveform 403. The comparator 420 outputs H level when the level of the waveform 401 is higher than the level of the waveform 402, and outputs L level when the level of the waveform 401 is lower than that of the waveform 402. Thus the waveform 403 is of a low level during a time interval between the zero crossing points E and F. In this manner, the zero crossing points at points E and F are recovered. Also, the comparator 420 may be readily constructed of, for example, a MIN circuit and a MAX circuit. Further, the first, second, and third embodiments may be combined. For example, a digital apparatus for preventing inversion phenomenon as shown in FIG. 17 can be realized by combining the BPF of the sine-type filter 219 in FIG. 6 with the pulse pickup circuit formed of the comparators 322, 323, the EX-OR circuits 325, AND circuit 328 and the delay circuit 327.

A fifth embodiment of the invention will now be described as follows.

FIG. 20 shows the fifth embodiment of an inversion phenomenon preventing apparatus and FIG. 27 illustrates waveforms at various points in FIG. 20. No zero crossing points are absent in the time axis of a normal FM signal 501 in FIG. 20 while the zero crossing points are absent at points E and F of a waveform 502 in FIG. 27. The waveform 502 represents an FM signal reproduced by means of the video signal storage/reproduction apparatus and is supplied to an input terminal IN of a block diagram in FIG. 20.

A delay circuit 513 has a delay time nearly equal to that of an HPF 511. When the output of the delay circuit 513 is fed to a first polarity discriminating circuit 522, amplitudes greater than a predetermined level, for example, O V, are converted into a high level and amplitudes smaller than O V are converted into a low level. Thus a waveform 503 as depicted in FIG. 27 is produced. In the mean time, the HPF 511 provides an output in which the zero crossing points are shifted in time from the waveform 502 but are not absent at all. A second polarity discriminating circuit 523 receives the output of the HPF 511 to provide a waveform 504 in FIG. 27. The waveforms 503 and 504 are supplied to an EX-OR circuit 525, which in turn provides an output waveform depicted by 505 in FIG. 27. It should be noted that the waveform 505 is at a high level both at portions where the zero crossing points are absent and at portions where the zero crossing points are shifted in time from the waveform 502. The missing zero crossing points may be recovered by extracting only pulses having pulse widths much wider than half the reciprocal of the highest instantaneous frequency of the waveform 502 and then adding these pulses, with their polarity being arranged properly, to the waveform 503. For this purpose, the waveform 505 is delayed by a delay circuit 527 to produce a waveform 506 which is delayed somewhat relative to the waveform 505. The waveform 505 and the waveform 506 are then supplied to an AND circuit 528 so as to produce a waveform as depicted by 507. Thus, a pulse appears only at a portion at which the zero crossing points are absent in the waveform 502. In order to combine the pulse 507 thus obtained with the waveform 502, the polarity of the pulse 507 must be carefully set. In the case where the waveform 502 is to be recovered as well as the missing zero crossing points thereof, a negative pulse must be added to the waveform 502 since the zero crossing points are absent at a high level amplitude of the waveform 503, while a positive pulse must be added if the zero crossing points are absent at a low level. The waveform 507 is supplied to a polarity control circuit 529 which reverses the polarity of an input thereto as shown by a waveform 508 in FIG. 27.

The waveform 508 and the waveform 502 are added by an adder 510 to produce a waveform 509 in FIG. 27 in which the zero crossing absent in the waveform 502 has now been recovered. The delay circuit 518 causes signal delay equal to the total amount of delay time of the signal path from the input IN to the output of the polarity control circuit 529.

FIG. 21 shows a specific embodiment of the polarity control circuit 529. FIG. 22 and FIG. 23 show specific embodiments of the polarity discriminating circuits 522, 523. Other circuit arrangements may also be used as polarity discriminating circuits providing that they operate in the above mentioned manner. Pulses having widths in accordance with how seriously the zero crossing points are absent may be added to the waveform 502 to optimally recover the absent zero crossing points. This is more advantageous than simply adding pulses having a fixed width produced by, for example, a monostable multivibrator or through differentiation of a signal with a fixed time constant. FIG. 24 shows a block diagram of an inversion phenomenon preventing apparatus using a circuit, called an inversion phenomenon detection pulse adding circuit 13. Here, pulses having widths in accordance with how seriously the zero crossing points are absent in the reproduced FM signal are added to the portions of the FM signal where the zero crossing points are actually absent.

FIG. 28 shows an example of the fifth embodiment to which an equalizer has been added in FIG. 20.

A sixth embodiment of the invention will now be described with reference to FIG. 25 as follows.

The operation of the sixth embodiment is generally the same as that of the arrangement in FIG. 20 except that the waveform 502 is not added directly to a pulse representative of absence of the zero crossing points, but through a delay circuit 618 and an amplitude limiter 614. This arrangement can prevent a variation in the envelope of the FM signal 502 caused during transmission before the waveforms are added by the adder 610. In addition, a BPF may be used in place of the HPF 611. A BPF which has good phase linearity may be used if phase characteristic is a most important design consideration. FIG. 26b shows an arrangement of such a BPF, where scaling factor indicates the ratio of levels of input signals supplied to the adder and may be set to 1/4, 1/2 and -1/4, respectively, for unity gain of the BPF. The circuit portions in FIGS. 20, 24, and 25, that correspond to a circuit in FIG. 26a, may be replaced by the circuit in FIG. 26b to obtain a filter having linear phase characteristics. Only a single filter is needed to form a circuit equivalent to that shown in FIG. 26b if the delay circuits in FIG. 26b are replaced by a reflection type.

While a combination of a delay circuit and an AND circuit has been used to obtain pulses having pulse widths in accordance with degree of absence of the zero crossing points, an LPF and a slicer may be combined so that pulses are converted into signals amplitude of which are proportional to the width thereof and then eliminates amplitudes below a predetermined threshold value by the slicer. FIG. 30a(1) shows a circuit example of a one side slicer. The output shown in FIG. 30a(2)=Max(E.sub.2, E.sub.1 +e.sub.1). The resultant output waveform is shown by a solid line. FIG. 30b(1) shows a circuit example of a double sided slicer. The output shown in FIG. 30b(2)=Max(E.sub.2, E.sub.1 +e.sub.1)+Min(E.sub.3, E.sub.1 +e.sub.1). The resultant output waveform at point (A) is shown by a solid line in FIG. 30b(3).

While the EX-OR circuit has been used as a means for comparing the outputs from the two polarity discriminating circuits 522 and 523, other circuit arrangement, for example, a subtracter, may be used for the same result as shown in FIG. 31, in which case the polarity control circuit 529 is not needed. Subtraction is performed by the single subtracter 530 in such a way that the output of the subtractor 530 will be of negative polarity when the output of the polarity discriminating circuit 522 is of H level and the output of the polarity discriminating circuit 523 is of L level while the pulse will be of positive polarity when the output is of L and H levels, respectively.

While combination of the LPF and the slicer simplifies circuit construction for extracting pulses in accordance with absence of the zero crossing points if the subtracter is used for comparing the outputs from the two polarity discriminating circuits 522, 523, the delay circuit 527 and the logic circuits may also be used to form the circuit.

Also the arrangement thus far described may be combined in any way, for example, an arrangement in FIG. 29a. An arrangement in FIG. 29b, where the exclusive OR circuit 525 has been replaced by subtractor 615 and the polarity control circuit 529 has been eliminated, may be used if the polarity discriminating circuits are constructed as shown in FIGS. 22 and 23. This is because the output of the polarity discriminating circuit 522 has the same information as the input FM signal in terms of the zero crossing points. FIG. 31 and FIG. 32 show such embodiments. The delay circuit used in FIG. 26b can of course be replaced by an LPF having a sufficient band width through which signals in the FM band can pass.