JP2007281662A - Signal processor and am broadcast receiver equipped with it - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To lessen discontinuity of a signal caused by removing noise contained in an input signal. <P>SOLUTION: A timing regulation buffer 30 obtains a signal A3 after timing regulation by delaying a wideband signal A2 including the band of a target station and its peripheral band. A delay buffer 40 stores the signal A3 after timing regulation temporarily. A noise detecting section 80 performs noise detection for the wideband signal A2, and a gate signal generating section 90 generates a gate signal of the same length as the size of the delay buffer 40 when noise is detected. A switching section 50 outputs the signal A3 after timing regulation to a target station extracting section 60 under normal state, and outputs a delay signal stored in the delay buffer 40 in the section of a gate signal. Size of the delay buffer 40 is determined such that the phase becomes continuous at the joints of the gated portion and the before and behind portions thereof in a signal outputted from the switching section 50. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a signal processing technique, specifically to a technique for removing noise from an input signal.

  When receiving radio broadcasts, for example, when receiving AM broadcasts with an on-vehicle AM (Amplitude Modulation) receiver, noise is generated due to the influence of an electric mirror or wiper of the vehicle in addition to ignition noise generated from the engine of the automobile. Sometimes. These noises are pulse-like noises (hereinafter simply referred to as noises) having a short time width and a large amplitude, and when they are superimposed on a received signal, they have an undesirable effect on hearing. Various techniques have been proposed to solve this problem.

  For example, Patent Literature 1 and Patent Literature 2 disclose techniques for detecting and removing noise from an intermediate frequency signal (hereinafter referred to as IF signal) before AM detection, and Patent Literature 3 and Non-Patent Literature 1 disclose. A technique for detecting and removing noise from an audio signal after AM detection is disclosed. Non-Patent Document 1 also discloses a technique for detecting and removing noise at the mixer output stage.

  In Patent Document 1, a gate signal having a predetermined width is generated when noise is detected from an IF signal in a receiving circuit for AM radio broadcast of a double conversion method, and immediately before the detected noise in a section of the gate signal. A technique for holding the IF signal is disclosed.

  Patent Document 2 discloses a technique for removing noise by switching and outputting an input signal and a prediction signal according to the presence or absence of noise. Specifically, when no noise is detected, the input signal is output and the input signal is stored in a signal storage memory for delaying the input signal for a certain period of time. When noise is detected, a prediction signal is formed from the waveform of a signal (hereinafter referred to as a storage signal) obtained by delaying the input signal stored in the signal storage memory, and this prediction signal is output. Since the storage signal is a signal obtained by delaying the input signal, it does not include noise at the timing when the noise is detected. That is, the input signal in the noise generation period is replaced with the prediction signal formed from the signal before the noise generation stored in the signal storage memory, and the noise is removed.

  Patent Document 3 discloses a technique for removing noise by switching and outputting an input signal and a synthesized signal in accordance with the presence or absence of noise. This technique also outputs an input signal (here, a sound signal after detection) and stores a signal obtained by delaying the input signal. When noise is detected, the stored signal before noise generation is used as a composite signal to replace the signal in the noise section.

Non-Patent Document 1 discloses two noise removal techniques. One is a technique for removing noise at the mixer output stage. According to this technique, when noise is detected from the mixer output, gate processing, that is, a gate signal is generated, and processing for holding a signal immediately before the noise or signal deletion processing by inserting 0 is performed in the gate signal section. . The other is a technique for performing gate processing when noise is detected on an audio signal obtained by AM detection. Here, noise is removed by linear interpolation in the gate processing interval.
JP 2000-353974 A JP 2002-064389 A JP-A 61-150532 Technical information magazine "PIONEER R &D" 2004 Vol. 14, no. 3 (http://www.pioneer.co.jp/crd/rd/14-3.html)

  However, when the process of simply holding the noise interval at the signal level before the noise signal or the signal deletion by zero insertion is performed, the hold or zero insertion part is discontinuous at the connection point with the preceding and following signals. May occur, resulting in discomfort to the listener.

  In particular, when noise is removed before the process of extracting a signal in the target station band (hereinafter referred to as a target station signal) from the AM broadcast reception signal, the target station signal is extracted from the noise-removed signal. In some cases, the above-mentioned discontinuity portion spreads and is finally output as noise in the audible band of the audio signal obtained by detection.

  The advantage of gate processing at the mixer output is that processing is performed at a stage where the signal band is wide, so that the gate section can be as short as a few μsec, and the gate operation sound is almost a problem for hearing. There is nothing. However, since the signal band is wide, it is difficult to detect small-signal noise, and it is necessary to use it together with gate processing for the detected audio signal.

  An audio signal obtained by detecting an IF signal including noise has a long noise interval, and the above-described signal discontinuity becomes more prominent when noise is removed by a technique such as linear interpolation.

  As described above, the problem to be solved by the present invention is to reduce the occurrence of signal discontinuity due to the removal of noise included in an input signal.

  One aspect of the present invention is a signal processing apparatus. The signal processing apparatus includes a noise detection unit that detects noise from an input signal that is an IF signal including an AM broadcast signal, a delay buffer that accumulates the input signal for a predetermined time to obtain a delay signal, the input signal, and the delay signal And an output unit that outputs an input signal when noise is not detected by the noise detection unit, and outputs the input signal by replacing the input signal with a delay signal when noise is detected. The delay buffer accumulates the input signal for a predetermined time determined so that the phase is continuous at the connection point between the replaced part and the front and rear parts of the signal output from the output unit.

  According to the signal processing device of the present invention, it is possible to reduce signal discontinuity caused by removing noise included in an input signal.

Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings. First, the first embodiment will be described.
FIG. 1 shows a configuration of an AM broadcast receiver 100 according to the present embodiment. The AM broadcast receiver 100 includes an AD converter 10, a broadband signal extraction unit 20, a timing adjustment buffer 30, a delay buffer 40, a switching unit 50, a target station extraction unit 60, a detection unit 70, a noise detection unit 80, and a gate signal generation. Part 90.

  The AD converter 10 digitally converts an intermediate frequency signal (IF signal) including an AM broadcast signal received from an antenna (not shown) to obtain a digital IF signal A1. The wideband signal extraction unit 20 extracts a signal within a range including the target station band and its peripheral band from the digital IF signal A1 to obtain the wideband signal A2, and outputs it to the timing adjustment buffer 30 and the noise detection unit 80. To do.

  Although details of the timing adjustment buffer 30 will be described later, the timing adjustment buffer 30 outputs the timing adjusted signal A3 obtained by delaying the wideband signal A2 to the switching unit 50 and the delay buffer 40. The delay buffer 40 accumulates and temporarily stores the signal A3 after timing adjustment. In the present specification, the time during which the delay buffer 40 accumulates the signal A3 after timing adjustment, that is, the reception signal, is also referred to as the buffer size. Details of these will be described later.

  The noise detector 80 detects noise from the wideband signal A2 and outputs a noise detector output A8 that can indicate the occurrence of noise to the gate signal generator 90.

  Based on the noise detection unit output A8, the gate signal generation unit 90 generates a gate signal A9 according to the occurrence of noise and outputs the gate signal A9 to the switching unit 50.

  That is, when the gate signal creation unit 90 receives the noise detection output A8 from the noise detection unit 80, the switching unit 50 receives the gate signal A9, and the output from the delay buffer 50 is output to the target station extraction unit 60. Switch so that

  That is, the switching unit 50 outputs the signal A3 after timing adjustment in the normal state, that is, the state where noise is not detected, while the signal A3 after timing adjustment is output to the delay buffer 40 when noise is detected. Replace with the delay signal stored in the output. This replacement is performed in response to the gate signal A9.

  Furthermore, after receiving the gate signal A9, the switching unit 50 has a function of switching itself back to the normal state after the signal accumulation time of the delay buffer 40 has elapsed. Specifically, for example, a timer is provided, and switching management after a predetermined time has elapsed can be performed by the timer. Therefore, a delay signal corresponding to the signal accumulation in the delay buffer is output to the target station extraction unit 60. Thereby, the noise in the signal A3 after timing adjustment is removed, and the signal A5 after noise removal is obtained. Here, the gate signal creation unit 90 and the switching unit 50 function as the output unit 95.

  The target station extraction unit 60 extracts a desired target station band signal (hereinafter referred to as a target station signal) A6 from the noise-removed signal A5. The detector 70 detects the target station signal A6 and obtains an audio signal A7. The audio signal A7 is further output to a speaker (not shown) and provided to the listener.

  As described above, the first embodiment described with reference to FIG. 1 applies the present invention to an AM broadcast receiver and obtains an audio signal while removing noise contained in the received signal. The AM broadcast receiver according to the present embodiment removes noise by replacing the received signal from when noise is detected with the signal stored in the delay buffer and immediately before noise detection. The time for which the delay buffer accumulates the received signal corresponds to the noise width (hereinafter simply referred to as noise width). That is, the length of the signal (delayed signal) accumulated in the delay buffer is a value corresponding to the noise width, specifically, a value that is approximately equal to or greater than the maximum noise width of the noise width to be removed. It has become. The degree of this size will be described later. By this replacement, the noise portion in the received signal becomes a delayed signal that does not include noise, and the noise is removed. Further, the time for which the delay buffer accumulates the received signal is determined such that the phase of the signal waveform after replacement is continuous in the replaced portion and the connection portion of the signal before and after the replaced portion. By doing so, it is possible to eliminate the phase discontinuity caused by the replacement for removing the noise. Since the time for which the delay buffer accumulates the received signal is the same as the time for delaying the input signal, the “accumulation time” and the “delay time” are used in the same meaning in the following description.

  Next, details of each component of the AM broadcast receiver 100 will be described.

  FIG. 2 is a diagram for explaining the broadband signal extraction unit 20. The wideband signal extraction unit 20 is, for example, a wideband extraction filter. As illustrated in the figure, the wideband signal extraction unit 20 thereby detects signals within the wideband extraction range, that is, signals in the target station band (450 KHz), and around the target station band. A signal is extracted from the digital IF signal A1, and a signal in a band of a station other than the target station (non-target station signal) is excluded.

  The timing adjustment buffer 30 is for delaying the wideband signal A2 for a predetermined time and ensuring that the noise part is surely replaced later by the switching unit 50. First, the delay buffer 40 responsible for noise removal, the noise detection unit 80, the gate signal creation unit 90, and the switching unit 50 will be described.

  The delay buffer 40 delays the timing-adjusted signal A3 for a predetermined time. This predetermined time corresponds to the size of the delay buffer 40. In the present embodiment, since the time during which the delay buffer can store the received signal is important, the concept will be described first.

  The time that the delay buffer can store the received signal is related to the number of bits of the received signal, the capacity of the delay buffer, and the structure. Here, the relationship between them will be described.

  FIG. 9 shows an example of a delay buffer used in the first embodiment of the present invention. This delay buffer is composed of a plurality of flip-flop blocks (hereinafter also simply referred to as blocks), and one block is further composed of a plurality of flip-flops. In the figure, one square represents one flip-flop, and one row represents one block. The number of columns is the number of block stages.

  Since one flip-flop holds 1 bit, the number of flip-flops included in one block of the delay buffer corresponds to the number of bits (or bit width) of the received signal. Since the present embodiment is an AM broadcaster, the number of bits of the received signal is the same as the word length necessary for the detection processing performed later, and is 21 bits in the illustrated example. Note that the number of flip-flops included in one block of the delay buffer is also referred to as the bit width of the buffer.

  The delay buffer shown in FIG. 9 is used as a ring buffer, and each stage is cyclically used in the order indicated by the arrow A in the figure. That is, the received signal is accumulated from the first stage (the uppermost stage in the illustrated example) of the delay buffer, and when the received signal is accumulated up to the last stage (the lowest stage in the illustrated example), the subsequent received signal is accumulated. Therefore, the accumulated signal is discarded from the first stage due to the FIFO (First In First Out) property of the buffer.

The time during which the delay buffer according to the first embodiment of the present invention can accumulate the received signal, that is, the time during which the received signal can be delayed can be calculated from the following equation (1).

Accumulation time = number of stages x sampling time = number of stages / sampling frequency (1)

Here, the sampling frequency means a sampling frequency when digitally converting the received signal. Therefore, it can be seen that when the sampling frequency is constant, the accumulation time of the received signal in the delay buffer increases in proportion to the number of stages of the delay buffer.

For example, in order to delay a received signal having a sampling frequency and a bit width of 3.8 MHz and 21 bits, respectively, by 80 μsec, the necessary number of stages X of a delay buffer having a bit width of 21 bits is calculated using the above equation (1). As a result, 304 stages are obtained as follows, and the capacity required for the delay buffer is 6384 bits.
80 × 10 ⁻⁶ = X / (3.8 × 10 ⁶ )
X = (80 × 10 ⁻⁶ ) × (3.8 × 10 ⁶ ) = 304

  Thus, the time for which the delay buffer accumulates the received signal is determined by the sampling frequency of the received signal, the number of bits of the received signal, the bit width of the delay buffer, and the number of stages. This time corresponds to the number of flip-flop blocks in a one-to-one relationship when the sampling frequency and number of bits of the received signal and the bit width of the buffer are constant.

  As described above, the delay buffer 40 delays the timing-adjusted signal A3 for a predetermined time. This predetermined time corresponds to the size of the delay buffer 40. In the present embodiment, the size of the delay buffer 40 is determined so as to satisfy the following two conditions.

Condition 1: The size of the delay buffer has a value corresponding to the noise width.
Here, the “value according to the noise width” means a value corresponding to a width that can replace the noise to be removed, and is a value in the vicinity of the maximum noise width to be removed. Can do. In the present embodiment, it is assumed that it is equal to or larger than the maximum noise width to be removed. Specific examples will be described later.

The size of the delay buffer 40 can be expressed by the following equation (2).

Delay buffer size (received signal accumulation time)
= Number of input signal cycles / input signal frequency (2)

The input signal means a signal input to the delay buffer 40, and here is the signal A3 after timing adjustment. “Number of cycles of input signal” is the number of cycles of the input signal that the delay buffer 40 can accumulate to the maximum. When the frequency of the input signal is fixed, the number of cycles of the input signal has a one-to-one relationship with the size of the delay buffer 40.

Condition 2: Corresponds to “number of periods of input signal” in the expression satisfying that the calculation result of Expression (3) is almost an integer.

(Sampling frequency / input signal frequency) × number of input signal cycles (3)

The “sampling frequency” is a sampling frequency when the input signal is converted from an analog signal to a digital signal, and corresponds to the sampling frequency used by the AD converter 10 here.

  Here, the number of periods of the input signal that satisfies that the calculation result of Expression (3) is almost an integer will be described below. As for “satisfying that it is almost an integer”, it is most preferable that the value of the first decimal place becomes 0 when the second decimal place of the calculation result of Expression (3) is rounded off. In this case, that is, the value after the decimal point is less than 0.05 or 0.95 or more. In these cases, even if replacement by noise removal is performed, the problem of discontinuity does not occur, and it is fully satisfied that the phase is continuous.

  In addition, even if it remove | deviates from these, depending on the grade, a problem may not arise. For example, even if the value after the decimal point is less than 0.1 or 0.9 or more, the quality of the demodulated output signal satisfies quality target values such as S / N of 30 dB or more and distortion of 1% or less. In such cases, these values can be regarded as almost integers.

  In general, it is known from experience that the width of pulse noise in a digital IF signal is 100 μsec or less before the target station signal is extracted.

  The frequency of the digital IF signal is generally a frequency used in the AM demodulator, here, the detection unit 70, and is usually 450 KHz or 455 KHz.

  The sampling frequency is a value at least twice as high as that of the IF signal, preferably at least 4 times, so that the digital IF signal can be sufficiently reproduced.

In the present embodiment, as an example, the sampling frequency is 3.8 MHz, and the frequency of the timing-adjusted signal A3 is 450 KHz. When removing noise having a maximum width of 80 μsec, the size of the delay buffer 40 according to the noise width is a size that accumulates a signal in the vicinity of 80 μsec, preferably 80 μsec or more. According to the equation (2), the number of periods Y of the input signal corresponding to 80 μsec can be calculated as 36 as follows.
80 × 10 ⁻⁶ = Y / (450 × 10 ³ )
Y = (80 × 10 ⁻⁶ ) × (450 × 10 ³ )

  Next, among each “number of periods of the input signal” in which the calculation result of the expression (3) is almost an integer, 36 is obtained as a value corresponding to the noise width.

For example, when the number of periods of the input signal is 36, specifically,
It can be seen that (3.8 × 10 ⁶ / (450 × 10 ³ )) × 36 = 304 and the calculation result is an integer 304.

  Thereby, the size of the delay buffer 40 is determined to be a size capable of accumulating a 450 kHz signal for 36 cycles. For example, when the bit width of the delay buffer 40 is equal to the bit width of the input signal, if the size of the delay buffer 40 is expressed by the number of stages, the number of stages is 304.

The delay buffer 40 thus accumulates and temporarily stores the signal A3 after timing adjustment for 36 cycles.
In addition, although it was stated that 36 was obtained from Expression (3), the frequency of the input signal such that the calculation result becomes an integer using Expression (3) can be obtained from simulation, for example, 18, 45, 72, and the like. However, if 18 is selected, the size of the delay buffer (reception signal accumulation time) is 40 μsec from Expression (2), which is significantly smaller than the noise width, so Condition 1 is not satisfied. In the case of 72, the noise is 160 μsec and noise having a width of 80 μsec can be removed. However, as can be seen from the equation (1), the number of stages of the delay buffer is increased, that is, more hardware resources are required. This is not preferable.

  Note that 45 is a value that can be selected as an embodiment, although the size of the buffer is 100 μsec, which is larger than 80 μsec and the redundancy is high, from Equation (2).

  Also, if it is desired to remove noise having a maximum width of 90 μsec, the number of cycles of the input signal is calculated to be 40.5 from the equation (2). Therefore, it is necessary to adopt 41 as the number of periods. In this case, the calculation result in Expression (3) is 346.2, which does not satisfy the condition of an integer.

When the number of periods is increased to 42, 43,..., The calculation result of Expression (3) becomes an integer of 380 with a period number of 45.
This value of 380 corresponds to a noise width of 100 μsec from equation (2), and the size of the delay buffer in condition 1 also satisfies that it is equal to or larger than the maximum noise width to be removed. In conclusion, this value is desirable.

  The noise detection unit 80 performs noise detection on the broadband signal A2 and outputs a noise detection unit output A8 that can indicate the occurrence of noise to the gate signal creation unit 90. Based on the noise detection unit output A8, the gate signal generation unit 90 generates a gate signal A9 for a predetermined section and outputs it to the switching unit 50. Here, the width of the gate signal A9 is equal to the size of the delay buffer 40, and is the width of the signal A3 after timing adjustment for 36 cycles accumulated in the delay buffer 40, that is, 80 μsec.

  The switching unit 50 outputs the timing adjusted signal A3 to the target station extracting unit 60 in the normal state. When the gate signal A9 is received, the delay signal in the delay buffer 40 is output to the target station extraction unit 60 instead of the timing-adjusted signal A3 in response to the gate signal. Since the delay signal stored in the delay buffer 40 is a signal immediately before noise, it does not contain noise. As a result, the noise included in the timing-adjusted signal A3 is removed. Hereinafter, the signal output from the switching unit 50 is referred to as a noise-removed signal A5.

  Note that, as described above, the switching unit 50 switches to the delay signal and outputs the signal A3 after timing adjustment when a predetermined time has elapsed since receiving the gate signal A9, that is, when the time for accumulating the received signal by the delay buffer has elapsed. The data is output to the extraction unit 60. Hereinafter, for convenience of explanation, a section during which the switching unit 50 outputs the delay signal after receiving the gate signal A9 is also referred to as a gate processing section.

  Thus, the size of the delay buffer 40 is set to a value that satisfies the above two conditions, and the noise portion included in the timing-adjusted signal A3 using the delay signal accumulated in the delay buffer 40 when noise is detected. By replacing and outputting the signal, noise can be surely removed, and in the signal A5 after noise removal, phase discontinuity caused by the substitution is caused at the connection point between the replaced portion and the portion before and after the replaced portion. Can be reduced.

  Note that the switching unit 50 can switch the bit signal of the number of bits of the received signal, and specifically includes a known means such as a number of switching elements (not shown) corresponding to the number of bits. The Further, as described above, a timer (not shown) is also provided, and has a function of returning to the normal state.

  A timer may be provided in the gate signal creation unit 90 instead of the switching unit 50, and a signal for returning to the normal state may be transmitted from the gate signal creation unit 90 to the switching unit 50 after a predetermined time has elapsed. Good.

Returning to the description of the timing adjustment buffer 30.
First, a case where there is no timing adjustment buffer 30 will be considered with reference to FIG. In this case, the signals transmitted to the delay buffer 40, the switching unit 50, and the noise detection unit 80 are also the wideband signal A2, and are not shifted in time from each other. However, the noise detection unit 80 requires some time to detect noise for the wideband signal A2. Therefore, as shown in FIG. 3, the timing at which the noise is detected by the noise detection unit 80 is slightly later than the timing at which the noise is actually generated in the broadband signal A2. Since the gate signal A9 is created when noise is detected by the noise detection unit 80, the start timing thereof is slightly later than the timing at which noise is actually generated in the wideband signal A2. That is, when the gate signal A9 is created, the noise portion in the broadband signal A2 has already been output to the target station extraction unit 60 by the switching unit 50. Even if the switching unit 50 switches the output to the delay signal in the delay buffer 40 in response to the gate signal A9, the noise is not removed. Further, in this case, there is a high possibility that the delayed signal also includes a noise portion that has already been output, and there is a risk that the noise may be increased by replacement.

  Therefore, in the AM broadcast receiver 100 according to the present embodiment, the wideband signal A2 is output as it is to the noise detection unit 80, while the wideband signal A2 is delayed for a predetermined time using the timing adjustment buffer 30. The data is output to the buffer 40 and the switching unit 50. Here, the predetermined time can be a difference between the timing at which noise is detected by the noise detection unit 80 and the timing at which noise is actually generated in the wideband signal A2. As shown in FIG. 4, in this case as well, the timing at which noise is detected by the noise detector 80 or the timing at which the gate signal is generated is slightly later than the timing at which noise is actually generated in the wideband signal A2. Become. However, since the signal output to the switching unit 50 and the delay buffer 40 is a signal obtained by delaying the wideband signal A2 by the timing adjustment buffer 30, the replacement is performed on the noise portion in the signal A3 after timing adjustment. . As a result, noise can be reliably removed.

  The target station extraction unit 60 extracts the target station signal A6 from the noise-removed signal A5, and thereby the band range to be extracted is several KHz, which is narrower than the band range extracted by the wideband signal extraction unit 20. .

The detection unit 70 performs AM detection on the target station signal A6 to obtain an audio signal A7.
FIG. 5D shows the waveform of the audio signal A7 obtained by the AM broadcast receiver 100 shown in FIG. FIG. 5A shows an audio signal obtained when noise removal is not performed on the same IF signal. FIG. 5C and FIG. 5D show audio signals obtained when noise is removed by a conventionally used noise removal method. FIG. 5C shows an audio signal obtained when noise is removed by inserting 0 before extracting the target station signal before detection, and FIG. 5D shows the audio signal after detection. An audio signal obtained when noise removal is performed on a signal by linear interpolation is shown.

  Consider the case of FIG. If noise is removed by inserting 0 before extracting the target station signal before detection, discontinuity occurs at the connection between the portion where 0 is inserted and the signal before and after that. This discontinuity includes a phase discontinuity and an amplitude discontinuity. When obtaining the target station signal through the target station extraction filter or the like after the noise is removed by this method, the band range extracted by the target station extraction filter is narrow. The interval of a certain part will spread. As shown in FIG. 5B, as a result of further detecting such a target station signal, a discontinuous portion can appear as noise in the audible band.

  Consider the case of FIG. When noise is removed by performing linear interpolation on the audio signal after detection, the width of the noise may reach several hundred μsec. Therefore, if the portion is linearly interpolated, the audio signal is shown in FIG. As shown in Fig. 2, distortion remains and the listener feels uncomfortable. For example, when noise having a width of 500 μsec in a signal after detection is removed by linear interpolation, a signal component of 2 kHz or higher in the audio band is deleted by linear interpolation, and a part of audio information is lost only in the noise section.

  On the other hand, the AM broadcast receiver 100 eliminates noise in a relatively narrow section before detection, and avoids interpolation of a wide noise section in the sound signal after detection, thereby eliminating audio information in the noise section. It is mitigating.

  Further, in the AM broadcast receiver 100, a value satisfying that the calculation result of the above equation (3) is almost an integer is used as the size of the delay buffer 40. Therefore, in the signal A5 after noise removal, The phase discontinuity between the front and back portions is suppressed. Further, since the signal immediately before the noise is used for the replacement, the amplitude discontinuity between the replaced portion and the portions before and after the replaced portion is also suppressed. As described above, since noise is removed while suppressing both phase and amplitude discontinuities, the discontinuity between the replaced portion and the portions before and after the replaced portion can be minimized. Therefore, even if the signal A5 after noise is passed through the target station extraction filter or the like, the possibility that noise is generated in the detected voice signal is reduced. Even if noise is generated, the noise is small enough not to give a sense of incongruity.

  Furthermore, before removing noise, the AM broadcast receiver 100 extracts a signal including a target station and a band around the target station from the digital IF signal A1 by the wideband signal extraction unit 20. Noise is removed from the extracted broadband signal A2. By this extraction, even if a non-target station signal is included in the digital IF signal A1, the effect of noise removal can be improved.

  FIG. 6 shows a processing result in the case where the AM broadcast receiver 100 does not perform extraction by the broadband signal extraction unit 20. Since the size of the delay buffer 40 is determined based on the frequency of the target station signal, if the non-target station signals are mixed in the signals input to the delay buffer 40 and the switching unit 50, FIG. As shown in a) and FIG. 6B, phase discontinuity occurs in the signal after noise removal using the delay signal accumulated in the delay buffer 40. When the target station signal is extracted from this signal, as shown in FIG. 6 (c), the portion having the phase discontinuity spreads in the target station signal. Then, the sound signal obtained by performing AM detection on the target station signal as shown in FIG. 6C is disturbed in a portion where there is a phase discontinuity as shown in FIG. It will make the listener feel uncomfortable.

  On the other hand, when the extraction by the broadband signal extraction unit 20 is performed, this problem is reduced. FIG. 7 shows the result in this case. Since the non-target station signal is not mixed in the signals input to the delay buffer 40 and the switching unit 50, as shown in FIGS. 7A and 7B, the signal A5 after noise removal includes Phase discontinuity is suppressed. As a result of extracting the target station signal A6 from this signal and detecting the extracted target station signal A6, the audio signal A7 is disturbed to give a sense of incongruity to the listener as shown in FIG. 7 (d). Absent.

  FIG. 8 shows the configuration of an AM broadcast receiver 200 according to the second embodiment of the present invention. Similarly to the AM broadcast receiver 100, the AM broadcast receiver 200 also obtains an audio signal while removing pulse noise (hereinafter simply referred to as noise) contained in the received signal. A signal extraction unit 120, a timing adjustment buffer 130, a delay buffer 140, a first switching unit 152, a second switching unit 154, a target station extraction unit 160, a detection unit 170, a noise detection unit 180, and a gate signal creation unit 190 Have. The AD converter 110, the broadband signal extraction unit 120, the timing adjustment buffer 130, the target station extraction unit 160, the detection unit 170, and the noise detection unit 180 in the AM broadcast receiver 200 are the AD converters in the AM broadcast receiver 100. 10, the broadband signal extraction unit 20, the timing adjustment buffer 30, the target station extraction unit 60, and the detection unit 70 are the same, and here, the delay buffer 140, the first switching unit 152, and the second switching unit 154 are used. Only the gate signal creation unit 190 will be described.

  In the normal state, the first switching unit 152 connects the timing adjustment buffer 130 and the delay buffer 140. In this state, the timing-adjusted signal from the timing adjustment buffer 130 is the second switching unit. 154 and the delay buffer 140.

  The second switching unit 154 connects the timing adjustment buffer 130 and the target station extraction unit 160 in a normal state. In this state, the signal after timing adjustment from the timing adjustment buffer 130 is transmitted to the target station extraction unit. 160 is output.

In the second embodiment, the gate signal creation unit 190, the first switching unit 152, and the second switching unit 154 constitute an output unit 195.
Also in the second embodiment, the hardware configuration of the first switching unit 152 and the second switching unit 154 can be configured by well-known means in the same manner as the switching unit 50 of the first embodiment. So I will omit the detailed explanation about them.

  In the AM broadcast receiver 200 of the present embodiment, the size of the delay buffer 140 is determined so as to satisfy the following two conditions.

Condition 1: M of 1 (M: an integer of 2 or more) having a value corresponding to the noise width.
Condition 2: Corresponds to “number of periods of input signal” in the expression satisfying that the calculation result of Expression (4) is almost an integer.
Here, “satisfying being almost an integer” is the same as the content described in the description of the first embodiment.
(Sampling frequency / input signal frequency) × number of input signal cycles (4)

The equation (4) is the same as the above equation (3), and the input signal is a signal after timing adjustment output from the timing adjustment buffer 130.
Here, the sampling frequency, the frequency of the input signal, and the noise width are set to 3.8 MHz, 450 KHz, and 80 μsec, respectively, by applying the example used when explaining the AM broadcast receiver 100.

  The number of periods of the input signal satisfying the expression (3) is 9, 18, 27, 36,..., And the size of the buffer corresponding to each of them is 20 μsec, 40 μsec, 60 μsec, 80 μsec,.・ It is. Among these, there are 20 μsec (M: 4) and 40 μsec (M: 2) as sizes satisfying the above condition 1. Here, as an example, it is assumed that the value of M is 4 and the size of the delay buffer 140 is set to 20 μsec.

  The gate signal creation unit 190 creates a gate signal when noise is detected by the noise detection unit 180 and outputs the gate signal to the first switching unit 152 and the second switching unit 154. Here, the width of the gate signal created by the gate signal creation unit 190 is M times the size of the delay buffer 140. Since the size of the delay buffer 140 is 20 μsec and M is 4, the width of the gate signal is 80 μsec.

  The second switching unit 154 switches the signal output to the target station extraction unit 160 to the delay signal stored in the delay buffer 140 during the gate signal interval.

  The first switching unit 152 disconnects the connection between the timing adjustment buffer 130 and the delay buffer 140 in the period of the gate signal, and feeds back the output of the delay buffer 140 to the delay buffer 140.

  That is, the signal accumulated in the delay buffer 140 is repeatedly output to the target station extraction unit 160 in the gate signal interval. Since the width of the gate signal is M times the size of the delay buffer 140, the signal accumulated in the delay buffer 140 is output M times (in this example, 4 times).

  Thus, the AM broadcast receiver 200 uses a value satisfying that the calculation result of Equation (4) is almost an integer as the size of the delay buffer 40, and uses the signal immediately before the noise for replacement. The same effect as that of the AM broadcast receiver 100 of the first embodiment can be obtained. In addition, since the size of the delay buffer is reduced, the circuit scale can be reduced.

  The present invention has been described above based on the embodiment. The embodiment is an exemplification, and various changes and increases / decreases may be added without departing from the gist of the present invention. It will be understood by those skilled in the art that modifications to which these changes and increases / decreases are also within the scope of the present invention.

  For example, in the AM broadcast receiver 100 and the AM broadcast receiver 200, noise detection processing and removal processing are performed on an IF signal converted into a digital signal. For example, noise detection processing is performed on an analog IF. It may be performed on the signal, and only noise removal processing may be performed on the IF signal converted into the digital signal.

  In the embodiment described above, the size of the delay buffer is the time that the delay buffer actually delays the input signal, or the number of stages corresponding to this time, and is necessarily the maximum possible size of the delay buffer. Not exclusively. For example, in the AM broadcast receiver 100 according to the first embodiment, the size of a delay buffer suitable for removing noise having a width of 100 μsec is 380 stages for an input signal whose sampling frequency is 3.8 MHz. On the other hand, the size of the delay buffer suitable for removing noise having a width of 80 μsec is 304 stages. Therefore, the noise width is set so that the number of stages of the delay buffer is increased to, for example, 400, and 380 stages are used for noise having a width of 100 μsec and only 304 stages are used for noise having a width of 80 μsec. Depending on the delay time, the number of stages used here may be switched.

  On the contrary, in the AM broadcast receiver 200 of the second embodiment, only 76 stages may be provided as the number of stages of the delay buffer, and the number of replacements may be switched according to the noise width. For example, in the case of noise having a width of 20 μsec, replacement is performed only once, but in the case of noise having a width of 80 μsec or 100 μsec, replacement of the same content may be repeated 4 or 5 times, respectively.

It is a figure which shows the structure of AM broadcast receiver concerning 1st Embodiment. It is a figure for demonstrating the process of the wideband signal extraction part in AM broadcast receiver shown in FIG. FIG. 3 is a diagram for explaining a timing adjustment buffer in the AM broadcast receiver shown in FIG. 1 (part 1); FIG. 3 is a diagram for explaining a timing adjustment buffer in the AM broadcast receiver shown in FIG. 1 (part 2); It is a figure which compares the audio | voice signal obtained by the AM broadcast receiver shown in FIG. 1 with the audio | voice signal obtained by the conventional method. FIG. 3 is a diagram for explaining the operation of a wideband signal extraction unit in the AM broadcast receiver shown in FIG. 1 (part 1); FIG. 7 is a diagram for explaining the operation of a wideband signal extraction unit in the AM broadcast receiver shown in FIG. 1 (part 2). It is a figure which shows the structure of AM broadcast receiver concerning 2nd Embodiment. Diagram for explaining delay buffer

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 AD converter 20 Wideband signal extraction part 30 Timing adjustment buffer 40 Delay buffer 50 Switching part 60 Target station extraction part 70 Detection part 80 Noise detection part 90 Gate signal preparation part 95 Output part 100 AM broadcast receiver 110 AD converter 120 Broadband signal extraction unit 130 Timing adjustment buffer 140 Delay buffer 152 First switching unit 154 Second switching unit 160 Target station extraction unit 170 Detection unit 180 Noise detection unit 190 Gate signal creation unit 195 Output unit 200 AM broadcast receiver

Claims (6)

A noise detector that detects noise from an input signal that is an IF signal including an AM broadcast signal;
A delay buffer for accumulating the input signal for a predetermined time to obtain a delay signal;
An output that receives the input signal and the delay signal and outputs the input signal when no noise is detected in the noise detection unit, and outputs the input signal by replacing the input signal with the delay signal when the noise is detected And comprising
The delay buffer receives the input signal for a predetermined time determined so that a phase is continuous at a connection point between a portion where the replacement is performed and a portion before and after the portion in the signal output by the output unit. A signal processing device characterized by storing.   The signal processing apparatus according to claim 1, further comprising a target station extraction unit that is provided behind the output unit and extracts a signal of the target station. A wideband signal extraction unit for extracting signals from the IF signal including the AM broadcast signal to the target station and its peripheral band;
The signal processing apparatus according to claim 1, wherein the input signal is a signal extracted by the broadband signal extraction unit. An AD converter that converts the input signal that is an analog signal into a digital signal at a predetermined sampling frequency;
The delay buffer and the output unit perform respective processing on the digital signal,
The predetermined time has a value corresponding to the width of the noise on the time axis, among the times corresponding to the number of digital signal periods satisfying the condition that the calculation result of Equation (1) is substantially an integer. The signal processing device according to any one of claims 1 to 3.

(Sampling frequency / digital signal frequency) x number of digital signal cycles (1)
An AD converter that converts the input signal that is an analog signal into a digital signal at a predetermined sampling frequency;
The delay buffer and the output unit perform respective processing on the digital signal,
The predetermined time is 1 / M of a value corresponding to the width of the noise on the time axis in the time corresponding to the number of digital signal periods satisfying the condition that the calculation result of Expression (2) is almost an integer. However, M is an integer of 2 or more)
4. The signal processing apparatus according to claim 1, wherein the output unit performs replacement using the delay signal in the delay buffer continuously M times. 5.

(Sampling frequency / digital signal frequency) x number of digital signal cycles (2)
A signal processing device according to any one of claims 1 to 5;
A detection unit for performing detection processing provided behind the output unit;
An AM broadcast receiver comprising:

Images