JP2014146941A - Noise reduction device, broadcast receiver and noise reduction method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the noise component in a reproduction sound signal simply and rationally.SOLUTION: A weighting coefficient generation unit 133A generates the weighting coefficients for each frequency component in a stereo difference signal (L-R) corresponding to the listening by spectral masking effect, based on a power spectral distribution of a stereo sum signal (R+L) calculated in a power spectrum calculation unit 132. Subsequently, a weighting unit 134 performs weighting for the time frequency conversion result SF(f) of the stereo difference signal (L-R). Thereafter, an IFFT unit 135performs frequency time conversion for the weighting results CSF(f) from the weighting unit 134 thus calculating a correction difference signal CST. A matrix unit 136 calculates an L channel signal for reproduction and an R channel signal for reproduction, based on the stereo sum signal (R+L) and the correction difference signal CST.

Description

  The present invention relates to a noise reduction device, a broadcast reception device, a noise reduction method, a noise reduction program, and a recording medium on which the noise reduction program is recorded.

  2. Description of the Related Art Conventionally, broadcast receiving apparatuses that receive and process FM stereo broadcast waves and reproduce sound are widely used. In such FM stereo broadcasting, a sum signal (L + R) and a difference signal (LR) of a left channel (hereinafter referred to as “L channel”) signal L and a right channel (hereinafter referred to as “R channel”) signal R are It is transmitted in different frequency bands. In the broadcast receiving apparatus, the L channel signal L and the R channel signal R are generated from the sum signal (L + R) and the difference signal (LR).

  In the following description, the sum signal is referred to as a “stereo sum signal”, and the difference signal is referred to as a “stereo difference signal”.

  As described above, in FM stereo broadcasting, since the transmission system of the stereo sum signal (L + R) and the stereo difference signal (LR) is different, the S / N of the stereo difference signal (LR) is the stereo sum signal. The S / N of (L + R) was deteriorated by 20 dB or more. In order to improve the S / N ratio of the stereo difference signal (LR), the stereo sum signal (L + R) is band-divided by a plurality of band pass filters (BPF), and the stereo difference signal (LR) is stereo-summed. A technique for extracting a component of a divided band in which a signal component of the signal (L + R) exists has been proposed (refer to FIGS. 4 and 5 and related descriptions in Patent Document 1; hereinafter referred to as “conventional example”).

Japanese Patent No. 3370716

  In the conventional technique described above, when determining whether or not the signal component of the stereo sum signal (L + R) exists in each divided band, the signal level of the component of the stereo sum signal (L + R) in each divided band is determined in advance. It is determined whether the value is equal to or greater than the reference value. However, since the noise level changes due to changes in reception time and reception position, it is difficult to appropriately determine such a reference value in advance. For example, when the reference value is set high in order to increase the S / N improvement effect, when the reproduced sound is a music sound, the divided band from which the component of the stereo difference signal (LR) is extracted frequently Will change. As a result, the frequency distribution of the noise component in the reproduced sound of each channel frequently changes, which may cause an adverse effect that the listener has a sense of instability with respect to the stereo reproduced sound.

  In the conventional technique, it is necessary to prepare BPFs having an accurate pass frequency range and steep frequency characteristics by the number of band divisions. However, it is not easy to prepare the required number of BPFs that satisfy these requirements.

  For this reason, there is a demand for a technique that can easily and efficiently reduce the noise component in the stereo difference signal while suppressing instability of stereo playback sound. Meeting this requirement is one of the problems to be solved by the present invention.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a noise reduction device and a noise reduction method that can easily and rationally reduce noise components in a reproduced audio signal. And Another object of the present invention is to provide a broadcast receiver capable of reproducing stereo sound by simply and rationally reducing noise components based on the reception result of FM stereo broadcast waves. And

  The invention according to claim 1 is a noise reduction device included in a broadcast receiving device that receives FM stereo broadcast, and calculates a power spectrum distribution of the stereo sum signal based on a time-frequency conversion result for the stereo sum signal. A power spectrum calculation unit; a weighting coefficient generation unit that generates a weighting coefficient for each frequency component in the stereo difference signal corresponding to the audibility by a spectrum masking effect based on the calculated power spectrum distribution; and for the stereo difference signal A weighting unit that performs weighting on the time-frequency conversion result using the generated weighting coefficient; a frequency-time conversion unit that performs frequency-time conversion on the weighting result by the weighting unit; and the stereo sum signal; Based on the conversion result by the frequency time conversion unit, A noise reducing apparatus characterized by comprising; a channel signal calculation unit for calculating a Namayo left channel signal and right channel signal for playback.

  The invention according to claim 5 is a reception processing unit that generates a stereo sum signal and a stereo difference signal based on the reception result of the FM stereo broadcast wave; and the noise reduction according to any one of claims 1 to 4; A broadcast receiving device comprising: an apparatus;

  The invention according to claim 6 is a noise reduction method used in a noise reduction device included in a broadcast receiving device that receives FM stereo broadcast, wherein the stereo sum signal is based on a time-frequency conversion result for the stereo sum signal. A power spectrum calculation step of calculating a power spectrum distribution of the sound; and a weighting coefficient generation step of generating a weighting coefficient for each frequency component in the stereo difference signal corresponding to the audibility by the spectrum masking effect based on the calculated power spectrum distribution A weighting step of performing weighting using the generated weighting coefficient on the time frequency conversion result for the stereo difference signal; and a frequency time conversion step of performing frequency time conversion on the weighting result in the weighting step; Said stereo sum signal and said A noise reduction method, characterized in that it comprises a; based on the conversion result in the wave number h conversion process, and channel signal calculating step of calculating a left channel signal and right channel signal reproduced reproduction.

  The invention described in claim 7 is a noise reduction program that causes a computer provided in a noise reduction device included in a broadcast receiving apparatus that receives FM stereo broadcast to execute the noise reduction method according to claim 6. is there.

  The invention described in claim 8 is characterized in that the noise reduction program according to claim 7 is recorded so as to be readable by a computer provided in a noise reduction apparatus included in a broadcast receiving apparatus that receives FM stereo broadcast. Recording medium.

1 is a block diagram showing a schematic configuration of a broadcast receiving apparatus according to a first embodiment of the present invention. It is a block diagram which shows the structure of the reception processing unit of FIG. It is a block diagram which shows the structure of the noise reduction apparatus of FIG. It is a figure for demonstrating operation | movement of the weighting coefficient production | generation part of FIG. It is a block diagram which shows the structure of the reception processing unit in the broadcast receiver which concerns on 2nd Embodiment of this invention. It is a block diagram which shows the structure of the noise reduction apparatus in the broadcast receiver which concerns on 3rd Embodiment of this invention. It is a figure for demonstrating operation | movement of the weighting coefficient production | generation part of FIG.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the following description and drawings, the same or equivalent elements are denoted by the same reference numerals, and redundant description is omitted.

[First Embodiment]
First, a first embodiment of the present invention will be described with reference to FIGS.

  In the first embodiment, the FM stereo broadcast wave is transmitted from the broadcast station via two transmission systems in which the sum signal and difference signal of the pre-emphasis processed L channel signal and R channel signal are in different frequency bands. It shall be broadcast.

<Configuration>
FIG. 1 is a block diagram illustrating a schematic configuration of a broadcast receiving device 100A according to the first embodiment. As shown in FIG. 1, the broadcast receiving apparatus 100A includes an antenna 110, a reception processing unit 120A, and a noise reducing apparatus 130A. The broadcast receiving apparatus 100A includes an analog processing unit 140 and speaker units 150 _L and 150 _R. Furthermore, the broadcast receiving apparatus 100A includes an input unit 180 and a control unit 190.

  The antenna 110 receives FM stereo broadcast waves. A reception result by the antenna 110 is sent to the reception processing unit 120A as a signal RFS.

  The reception processing unit 120A receives the signal RFS transmitted from the antenna 110. Then, the reception processing unit 120A, in accordance with the channel selection command CSL sent from the control unit 190, from the signal RFS, a stereo sum signal (R + L) and a stereo difference signal corresponding to the desired broadcast station specified by the channel selection command CSL ( RL). The stereo sum signal (R + L) and the stereo difference signal (R−L) thus generated are sent to the noise reduction device 130A. In other words, the reception processing unit 120A functions as a reception processing unit.

  Details of the configuration of the reception processing unit 120A will be described later.

  The noise reduction device 130A receives the stereo sum signal (R + L) and the stereo difference signal (RL) sent from the reception processing unit 120A. Then, the noise reduction device 130A generates a reproduction L channel signal CLD and a reproduction R channel signal CRD based on the stereo sum signal (R + L) and the stereo difference signal (R−L). The reproduction L channel signal CLD and the reproduction R channel signal CRD thus generated are sent to the analog processing unit 140.

  Details of the configuration of the noise reduction device 130A will be described later.

The analog processing unit 140 receives the reproduction L channel signal CLD and the reproduction R channel signal CRD sent from the noise reduction device 130A. Then, under the control of the control unit 190, the analog processing unit 140 generates the output sound signal AOS _L based on the reproduction L channel signal CLD, and outputs the output sound signal AOS based on the reproduction R channel signal CRD. Generate _R.

  The analog processing unit 140 having such a function includes a DA (Digital to Analogue) conversion unit, a volume adjustment unit, and a power amplification unit. Here, the DA converter receives the reproduction L channel signal CLD and the reproduction R channel signal CRD transmitted from the noise reduction device 130A. The DA converter converts the reproduction L channel signal CLD and the reproduction R channel signal CRD into analog signals. Note that the DA converter includes two DA converters configured similarly to each other, corresponding to the reproduction L channel signal CLD and the reproduction R channel signal CRD. The analog conversion result by the DA conversion unit is sent to the volume adjustment unit.

  The volume adjuster receives the analog conversion result signals of the L channel and the R channel sent from the DA converter. Then, the volume adjustment unit performs volume adjustment processing on the analog conversion result signals corresponding to the L channel and the R channel in accordance with the volume adjustment command VLC sent from the control unit 190. Note that the volume adjustment unit is configured to include two electronic volume elements configured in the same manner, corresponding to the L channel and the R channel. The signal of the volume adjustment result by the volume adjustment unit is sent to the power amplification unit.

The power amplification unit receives the signals of the volume adjustment results of the L channel and the R channel sent from the volume adjustment unit. The power amplifying unit power-amplifies the signal of the volume adjustment result to generate the output sound signal AOS _L and the output sound signal AOS _R. The output sound signal AOS _L generated in this way is sent to the speaker unit 150 _L , and the generated output sound signal AOS _R is sent to the speaker unit 150 _R.

The speaker unit 150 _L receives the output sound signal AOS _L sent from the analog processing unit 140. Then, the speaker unit 150 _L reproduces and outputs sound in accordance with the output sound signal AOS _L.

The speaker unit 150 _R receives the output sound signal AOS _R sent from the analog processing unit 140. Then, the speaker unit 150 _R reproduces and outputs sound in accordance with the output sound signal AOS _R.

  The input unit 180 includes a key unit provided in the main body of the broadcast receiving device 100A, or a remote input device including the key unit. Here, as a key part provided in the main body, a touch panel provided in a display unit (not shown) can be used. Moreover, it can replace with the structure which has a key part, and the structure which inputs voice can also be employ | adopted. The input result to the input unit 180 is sent to the control unit 190 as input data IPD.

  The control unit 190 receives the input data IPD sent from the input unit 180. If the content of the input data IPD is a channel selection designation, the control unit 190 generates a channel selection command CSL corresponding to the designated desired station and sends it to the reception processing unit 120A. If the content of the input data IPD is volume adjustment designation, the control unit 190 generates a volume adjustment command VLC corresponding to the designated volume adjustment designation and sends it to the analog processing unit 140.

<< Configuration of Reception Processing Unit 120A >>
Next, the configuration of the reception processing unit 120A will be described.

The reception processing unit 120A includes an RF processing unit 121 and a detection unit 122, as shown in FIG. The reception processing unit 120A includes a multiplexer (MPX) unit 123 and de-emphasis correction units 124 _M and 124 _S.

  The RF processing unit 121 performs channel selection processing for extracting a signal of a desired station to be selected from the signal RFS in accordance with the channel selection command CSL sent from the control unit 190, and has a component in a predetermined intermediate frequency band. Convert to an intermediate frequency signal IFD. The intermediate frequency signal IFD thus obtained is sent to the detection unit 122.

  The intermediate frequency signal IFD is a digital signal.

  The detection unit 122 receives the intermediate frequency signal IFD sent from the RF processing unit 121. And the detection part 122 performs a detection process with respect to the intermediate frequency signal IFD. The detection result by the detection unit 122 is sent to the MPX unit 123 as a detection signal DTD.

The detection signal DTD is a sum signal (L _P + R _P ) and a difference signal (L _P −R _P ) of the received L channel signal L _P and the received R channel signal R _P in which the effect of pre-emphasis processing remains. Is a stereo decoded signal reflected in different frequency bands.

The MPX unit 123 receives the detection signal DTD sent from the detection unit 122. Then, the MPX unit 123 extracts the sum signal (L _P + R _P ) and the difference signal (L _P −R _P ) from the detection signal DTD. The extracted sum signal (L _P + R _P ) is sent to the de-emphasis correction unit 124 _M. Further, the extracted difference signal (L _P −R _P ) is sent to the de-emphasis correcting unit 124 _S.

The de-emphasis correcting unit 124 _M receives the sum signal (L _P + R _P ) sent from the MPX unit 123. Then, the de-emphasis correction unit 124 _M performs de-emphasis correction on the sum signal (L _P + R _P ) to generate a stereo sum signal (L + R). The stereo sum signal (L + R) generated in this way is sent to the noise reduction device 130A.

The de-emphasis correcting unit 124 _S receives the difference signal (L _P −R _P ) sent from the MPX unit 123. Then, the de-emphasis correcting unit 124 _S performs de-emphasis correction on the difference signal (L _P −R _P ) to generate a stereo difference signal (L−R). The stereo difference signal (LR) generated in this way is sent to the noise reduction device 130A.

<< Configuration of Noise Reduction Device 130A >>
Next, the configuration of the noise reduction device 130A will be described.

As illustrated in FIG. 3, the noise reduction device 130 </ b> A includes Fourier transform units (FFT units) 131 _M and 131 _S , a power spectrum calculation unit 132, a weighting coefficient generation unit 133 </ b> A, and a weighting unit 134. . The noise reduction device 130 </ b> A includes inverse Fourier transform units (IFFT units) 135 _M and 135 _S and a matrix unit 136.

The FFT unit 131 _M receives the stereo sum signal (L + R) sent from the reception processing unit 120A. Then, the FFT unit 131 _M performs Fourier transform (time frequency conversion) on the stereo sum signal (L + R). The conversion result MF (f) by the FFT unit 131 _M is sent to the power spectrum calculation unit 132.

The FFT unit 131 _S receives the stereo difference signal (LR) sent from the reception processing unit 120A. Then, the FFT unit 131 _S performs a Fourier transform on the stereo difference signal (LR). The conversion result SF (f) by the FFT unit 131 _S is sent to the weighting unit 134.

Additional power spectrum calculating unit 132, the conversion result sent from the FFT unit 131 _M receives the MF (f). Then, the power spectrum calculation unit 132 calculates the power spectrum distribution PSD (f) of the stereo sum signal (L + R). The power spectrum distribution PSD (f) calculated in this way is sent to the weighting coefficient generation unit 133A. The power spectrum calculator 132, the conversion result MF to (f), and sends to the IFFT section 135 _M.

  Said weighting coefficient production | generation part 133A hold | maintains the predetermined masking curve calculation parameter inside. Such masking curve calculation parameters are determined based on experiments, simulations, experiences, etc. while utilizing the results in auditory psychology regarding the spectrum masking effect.

  The weighting coefficient generation unit 133A receives the power spectrum distribution PSD (f) sent from the power spectrum calculation unit 132. Then, the weighting coefficient generation unit 133A applies the power spectrum distribution PSD (f) to each frequency component in the stereo difference signal (LR) based on the masking curve calculated using the masking curve calculation parameter. Weighting coefficient information WCA is generated. The weighting coefficient information WCA thus generated is sent to the weighting unit 134.

  Details of the processing by the weighting coefficient generation unit 133A will be described later.

The weighting unit 134 receives the conversion result SF (f) sent from the FFT unit 131 _S and the weighting coefficient information WCA sent from the weighting coefficient generation unit 133A. The weighting unit 134 weights each frequency component in the stereo difference signal (LR) represented by the conversion result SF (f). The weighting result CSF (f) by the weighting unit 134 is sent to the IFFT unit 135 _S.

The IFFT unit 135 _M receives the conversion result MF (f) sent from the power spectrum calculation unit 132. Then, IFFT unit 135 _M performs inverse Fourier transform (frequency time transform) on conversion result MF (f) to reproduce a stereo sum signal (L + R). The stereo sum signal (L + R) reproduced in this way is sent to the matrix unit 136.

The IFFT unit 135 _S receives the weighting result CSF (f) sent from the weighting unit 134. Then, IFFT section 135 _S performs inverse Fourier transform on weighting result CSF (f) to calculate corrected difference signal CST in which the noise component in stereo difference signal (LR) is reduced. The correction difference signal CST calculated in this way is sent to the matrix unit 136.

The matrix unit 136 receives the stereo sum signal (L + R) sent from the IFFT unit 135 _M and the correction difference signal CST sent from the IFFT unit 135 _S. The matrix unit 136 adds the stereo sum signal (L + R) and the correction difference signal CST, and calculates a reproduction L channel signal CLD based on the addition result. Further, the matrix unit 136 subtracts the correction difference signal CST from the stereo sum signal (L + R), and calculates the reproduction R channel signal CRD based on the subtraction result. The reproduction L channel signal CLD and the reproduction R channel signal CRD thus calculated are sent to the analog processing unit 140. In other words, the matrix unit 136 functions as a channel signal calculation unit.

<Operation>
Next, the operation of the broadcast receiving apparatus 100A configured as described above will be described mainly focusing on the processing by the weighting coefficient generation unit 133A.

  As a premise, the channel selection designation is already input to the input unit 180 by the user, and the channel selection command CSL corresponding to the designated desired station is sent to the reception processing unit 120A (more specifically, the RF processing unit 121). Suppose that it is sent. Further, it is assumed that a volume adjustment designation has already been input to the input unit 180 by the user, and a volume adjustment command VLC corresponding to the designated volume adjustment mode has been sent to the analog processing unit 140 (FIG. 1). reference).

  In this state, when a broadcast wave is received by the antenna 110, the signal RFS is transmitted from the antenna 110 to the reception processing unit 120A. In the reception processing unit 120A, the RF processing unit 121 receives the signal RFS. Then, in the RF processing unit 121, after the signal of the desired station to be selected is converted into a signal in the intermediate frequency band, AD conversion is performed. The result of this AD conversion is sent to the detector 122 as an intermediate frequency signal IFD (see FIG. 2).

The detection unit 122 performs detection processing on the intermediate frequency signal IFD. And the detection part 122 sends the result of a detection process to the MPX part 123 as a detection signal DTD. Then, the MPX unit 123 extracts the sum signal (L _P + R _P ) and the difference signal (L _P −R _P ) from the detection signal DTD (see FIG. 2).

Subsequently, the de-emphasis correcting unit 124 _M performs de-emphasis correction on the sum signal (L _P + R _P ) to generate a stereo sum signal (L + R), and noise reduction is performed on the generated stereo sum signal (L + R). Send to device 130A. Further, the de-emphasis correcting unit 124 _S performs de-emphasis correction on the difference signal (L _P −R _P ) to generate a stereo difference signal (L−R), and the generated stereo difference signal (L−R). ) Is sent to the noise reduction device 130A (see FIG. 2).

In the noise reduction device 130A, the FFT unit 131 _M receives the stereo sum signal (L + R), and the FFT unit 131 _S receives the stereo difference signal (L−R). Then, the FFT unit 131 _M performs a Fourier transform on the stereo sum signal (L + R), and sends the conversion result MF (f) to the power spectrum calculation unit 132. Also, the FFT unit 131 _S performs a Fourier transform on the stereo difference signal (LR), and sends the transformation result SF (f) to the weighting unit 134 (see FIG. 3).

The power spectrum calculation unit 132 calculates the power spectrum distribution PSD (f) of the stereo sum signal (L + R) based on the conversion result MF (f), and the calculated power spectrum distribution PSD (f) is a weighting coefficient generation unit. Send to 133A. Also, the power spectrum calculation unit 132 sends the conversion result MF (f) to the IFFT unit 135 _M (see FIG. 3). An example of the power spectrum distribution PSD (f) calculated by the power spectrum calculation unit 132 is shown in FIG.

  When receiving the power spectrum distribution PSD (f), the weighting coefficient generation unit 133A first calculates the total power for each predetermined number of divided bands. An example of the total power calculated for each divided band in this way is indicated by a solid line arrow in FIG. FIG. 4B shows an example in which the number of divided bands is 10 and the center frequency of divided band Rj (j = 1 to 10) is fj.

  Next, the weighting coefficient generation unit 133A determines whether or not the sound component for each divided band is difficult to hear due to the spectrum masking effect. In making this determination, the weighting coefficient generation unit 133A first identifies the divided band having the maximum total power and sets it as the first specific divided band. Then, the weighting coefficient generation unit 133A determines that the weighting coefficient related to the first specific divided band is “1”. In the example shown in FIG. 4B, the divided band R5 is the first specific divided band, and as shown in FIG. 4C, the weighting coefficient for the divided band R5 is determined to be “1”.

  Subsequently, the weighting coefficient generation unit 133A, based on the total power in the first specific divided band and the masking curve calculation parameter held therein, the first masking curve corresponding to the sound component of the first specific divided band. Is calculated. In FIG. 4C, an example of the first masking curve is indicated by a one-dot chain line as a masking curve MC5.

  Next, the weighting coefficient generation unit 133A determines whether the total power in each of the divided bands other than the first specific divided band is less than the value of the first masking curve at the center frequency of the divided band. A first masking determination is made as to whether or not the sound components in each of the divided bands other than the first specific divided band are within the masking range determined by the first masking curve and are difficult to hear. Subsequently, the weighting coefficient generation unit 133A determines “0” as the weighting coefficient for the divided band for which the result of the first masking determination is positive. In the example shown in FIG. 4C, the weighting coefficient for the divided bands R4, R6, and R7 is determined to be “0”.

  Next, the weighting coefficient generation unit 133A identifies the divided band having the maximum total power among the divided bands for which the weighting coefficient has not been determined, and sets it as the second specific divided band. Then, the weighting coefficient generation unit 133A determines that the weighting coefficient related to the second specific divided band is “1”. In the example shown in FIG. 4B, the divided band R8 is the second specific divided band, and the weighting coefficient for the divided band R8 is determined to be “1” as shown in FIG. 4C.

  Subsequently, the weighting coefficient generation unit 133A, based on the total power in the second specific divided band and the masking curve calculation parameter held therein, the second masking curve corresponding to the sound component of the second specific divided band. Is calculated. In FIG. 4C, an example of the second masking curve is indicated by a one-dot chain line as a masking curve MC8.

  Next, the weighting coefficient generation unit 133A determines whether the total power in each of the divided bands for which the weighting coefficient has not been determined is less than the value of the second masking curve at the center frequency of the divided band. A second masking determination is made as to whether or not the sound components in each of the divided bands for which the weighting coefficient has not been determined are within the masking range determined by the second masking curve and are difficult to hear. Subsequently, the weighting coefficient generation unit 133A determines “0” as the weighting coefficient for the divided band for which the result of the second masking determination is positive. In the example shown in FIG. 4C, the weighting coefficient for the divided bands R9 and R10 is newly determined to be “0”.

  Thereafter, the weighting coefficient generation unit 133A determines (i) the identification of the divided band having the maximum total power among the divided bands for which the weighting coefficient is not determined, and (ii) the weighting coefficient related to the identified divided band. Determination of “1”, (iii) calculation of a masking curve based on the total power in the specified divided band and a masking curve calculation parameter held therein, and (iv) weighting using the masking curve The masking determination for each of the divided bands for which the coefficient has not been determined, and (v) the determination of setting the weighting coefficient for the divided band for which the result of the masking determination is affirmative as “0”, the weighting coefficient for all the divided bands Repeat until is determined. When the weighting coefficients for all the divided bands are determined, weighting coefficient generation section 133A sends the determined weighting coefficients as weighting coefficient information WCA to weighting section 134 (see FIG. 3).

  In the example shown in FIG. 4, the weighting coefficients for the divided bands R2, R5, and R8 are determined as “1”, and the weighting coefficients for the divided bands R1, R3, R4, R6, R7, R9, and R10 are determined as “0”. Is done.

The weighting unit 134 weights the conversion result SF (f) sent from the FFT unit 131 _{S according} to the weighting coefficient information WCA. In the first embodiment, since the weighting coefficient in the weighting coefficient information WCA is “1” or “0”, all the components of the divided bands in which the weighting coefficient is set to “1” in the weighting coefficient information WCA are left as they are. A weighted result CSF (f) is generated by removing all the components of the divided band whose weighting coefficient is set to “0”. Then, the weighting unit 134 sends the weighting result CSF (f) to the IFFT unit 135 _S (see FIG. 3).

The IFFT unit 135 _M performs inverse Fourier transform on the conversion result MF (f) sent from the power spectrum calculation unit 132 to reproduce the stereo sum signal (L + R), and the reproduced stereo sum signal (L + R) Is sent to the matrix unit 136. The IFFT unit 135 _S performs an inverse Fourier transform on the weighting result CSF (f) sent from the weighting unit 134 to generate a corrected difference signal CST in which the noise component in the stereo difference signal (LR) is reduced. The calculated correction difference signal CST is sent to the matrix unit 136 (see FIG. 3).

  Next, the matrix unit 136 adds the stereo sum signal (L + R) and the correction difference signal CST, and calculates a reproduction L channel signal CLD based on the addition result. Further, the matrix unit 136 subtracts the correction difference signal CST from the stereo sum signal (L + R), and calculates the reproduction R channel signal CRD based on the subtraction result. Then, the matrix unit 136 sends the reproduction L channel signal CLD and the reproduction R channel signal CRD to the analog processing unit 140 (see FIG. 3).

In the analog processing unit 140, the DA conversion unit, the volume control unit, and the power amplification unit sequentially perform processing, generate output sound signals AOS _L and AOS _R from the reproduction L channel signal CLD and the reproduction R channel signal CRD, and the speaker. It is sent to the units 150 _L and 150 _R (see FIG. 1). Then, the speaker units 150 _L and 150 _R reproduce and output sound according to the output sound signals AOS _L and AOS _R.

  As described above, in the first embodiment, the power spectrum calculation unit 132 uses the power of the stereo sum signal (R + L) based on the time frequency transform result (Fourier transform result) MF (f) for the stereo sum signal (R + L). A spectral distribution PSD (f) is calculated. Subsequently, the weighting coefficient generation unit 133A generates a weighting coefficient for each frequency component in the stereo difference signal (LR) corresponding to the audibility by the spectrum masking effect based on the power spectrum distribution PSD (f).

Next, after the weighting unit 134 performs weighting using the weighting coefficient on the time-frequency conversion result SF (f) for the stereo difference signal (LR), the IFFT unit 135 _S performs weighting by the weighting unit 134. The result CSF (f) is subjected to frequency-time transform (inverse Fourier transform) to calculate a corrected difference signal CST in which the noise component in the stereo difference signal (LR) is reduced. Then, the matrix unit 136 performs the reproduction L channel signal and the stereo difference signal CST based on the stereo sum signal (L + R) reproduced by performing the time-frequency conversion on the stereo sum signal (L + R), and the correction difference signal CST. A reproduction R channel signal is calculated.

  That is, in the first embodiment, among the frequency components in the stereo difference signal (LR) having a low S / N, the contribution of the noise components in the frequency components that are difficult to hear due to the spectrum masking effect to the reproduced sound is suppressed. It has come to be. Therefore, according to the first embodiment, it is possible to easily and rationally reduce noise components in the reproduced audio signal.

  In the first embodiment, the weighting coefficient generation unit 133A calculates the power of each frequency component in the stereo sum signal (L + R) based on the power of other frequency components and a predetermined masking curve calculation parameter. A masking determination is made as to whether or not the masking range is determined by the masking curve corresponding to the other frequency component. Then, the weighting coefficient generation unit 133A determines the weighting coefficient to be “1” for the frequency component for which the masking determination result is negative. Also, the weighting coefficient generation unit 133A determines the weighting coefficient to be “0” for the frequency component for which the masking determination result is positive.

  That is, in the first embodiment, the weighting coefficient is set to “0” according to whether or not the power of each frequency component in the stereo sum signal (L + R) is within the masking range by the spectrum masking effect corresponding to the other frequency components. "Or" 1 ". Therefore, according to the first embodiment, the noise component in the reproduced audio signal can be reduced very simply and rationally.

[Second Embodiment]
Next, a second embodiment of the present invention will be described with reference mainly to FIG.

<Configuration>
The broadcast receiving apparatus according to the second embodiment is different from the broadcast receiving apparatus 100A of the first embodiment described above in that a reception processing unit 120B having the configuration shown in FIG. 5 is provided instead of the reception processing unit 120A. ing. In the following description, the broadcast receiving device of the second embodiment is referred to as “broadcast receiving device 100B”.

As shown in FIG. 5, the reception processing unit 120B does not include the de-emphasis correction units 124 _M and 124 _S as compared to the reception processing unit 120A described above, and the matrix unit 125 and the de-emphasis correction. The difference is that the units 126 _L and 126 _R and the matrix unit 127 are further provided. Hereinafter, description will be made mainly focusing on these differences.

In the reception processing unit 120B, the RF processing unit 121, the detection unit 122, and the MPX unit 123 operate in the same manner as in the case of the reception processing unit 120A. Then, the sum signal (L _P + R _P ) and the difference signal (L _P −R _P ) extracted by the MPX unit 123 are sent to the matrix unit 125.

The matrix unit 125 receives the sum signal (L _P + R _P ) and the difference signal (L _P −R _P ) sent from the MPX unit 123. Then, the matrix unit 125 adds the sum signal (L _P + R _P ) and the difference signal (L _P −R _P ), and calculates a received L channel signal L _P based on the addition result. In addition, the matrix unit 125 subtracts the difference signal (L _P −R _P ) from the sum signal (L _P + R _P ), and calculates the received R channel signal R _P based on the subtraction result.

The reception L channel signal L _P calculated in this way is sent to the de-emphasis correction unit 126 _L. Further, the calculated received R channel signal R _P is sent to the de-emphasis correcting unit 126 _R.

The de-emphasis correction unit 126 _L receives the reception L channel signal L _P sent from the matrix unit 125. Then, the de-emphasis correcting unit 126 _L performs de-emphasis correction on the received L channel signal L _P and generates an L channel signal L. The L channel signal L generated in this way is sent to the matrix unit 127.

The de-emphasis correction unit 126 _R receives the reception R channel signal R _P sent from the matrix unit 125. Then, the de-emphasis correction unit 126 _R performs de-emphasis correction on the received R channel signal R _P to generate an R channel signal R. The R channel signal R generated in this way is sent to the matrix unit 127.

The matrix unit 127 receives the L channel signal L sent from the de-emphasis correcting unit 126 _L and the R channel signal R sent from the de-emphasis correcting unit 126 _R. Then, the matrix unit 127 calculates a stereo sum signal (L + R) by adding the L channel signal L and the R channel signal R. The matrix unit 127 calculates a stereo difference signal (LR) by subtracting the R channel signal R from the L channel signal L. The stereo sum signal (L + R) and the stereo difference signal (LR) calculated in this way are sent to the noise reduction device 130A.

  By the way, the configuration in which the matrix unit 127 is removed from the above-described reception processing unit 120B is a configuration that has been generally employed in a broadcast receiving apparatus, and has been configured as an LSI (Large Scale Integrated Circuit). That is, the configuration of the reception processing unit 120B in the broadcast receiving apparatus 100B is a configuration in which the matrix unit 127 is added on to the configuration of the reception processing unit generally employed.

<Operation>
Next, the operation of the broadcast receiving device 100B configured as described above will be described.

  As a premise, the channel selection designation is already input to the input unit 180 by the user, and the channel selection command CSL corresponding to the designated desired station is sent to the reception processing unit 120B (more specifically, the RF processing unit 121). Suppose that it is sent. Further, it is assumed that a volume adjustment designation has already been input to the input unit 180 by the user, and a volume adjustment command VLC corresponding to the designated volume adjustment mode has been sent to the analog processing unit 170 (FIG. 1). And FIG. 5).

When a broadcast wave is received by the antenna 110 in such a state, a sum signal (L _P + R _P ) and a difference signal (L _P −R _P ) are output from the MPX unit 123 as in the case of the broadcast receiving device 100A described above. Is done. The sum signal (L _P + R _P ) and the difference signal (L _P −R _P ) output in this way are input to the matrix unit 125 (see FIG. 5).

The matrix unit 125 adds the sum signal (L _P + R _P ) and the difference signal (L _P −R _P ), and calculates a received L channel signal L _P based on the addition result. In addition, the matrix unit 125 subtracts the difference signal (L _P −R _P ) from the sum signal (L _P + R _P ), and calculates the received R channel signal R _P based on the subtraction result. Then, the matrix unit 125 sends the reception L channel signal L _P to the de-emphasis correction unit 126 _L and sends the reception R channel signal R _P to the de-emphasis correction unit 126 _R (see FIG. 5).

The de-emphasis correcting unit 126 _L performs de-emphasis correction on the received L channel signal L _P sent from the matrix unit 125 to generate an L channel signal L, and the generated L channel signal L to the matrix unit 127. send. The de-emphasis correcting unit 126 _R performs de-emphasis correction on the received R channel signal R _P sent from the matrix unit 125 to generate an R channel signal R, and the generated R channel signal R is used as a matrix unit. 127 (see FIG. 5).

  The matrix unit 127 calculates a stereo sum signal (L + R) by adding the L channel signal L and the R channel signal R. The matrix unit 127 calculates a stereo difference signal (LR) by subtracting the R channel signal R from the L channel signal L. Then, the matrix unit 127 sends the stereo sum signal (L + R) and the stereo difference signal (LR) to the noise reduction device 130A (see FIG. 5).

Thereafter, the noise reduction device 130A and the analog processing unit 140 sequentially perform the same processing as in the first embodiment described above, and output sound signals AOS _L and AOS _R are generated. Then, the speaker units 150 _L and 150 _R reproduce and output sound according to the output sound signals AOS _L and AOS _R.

  As described above, in the second embodiment, the noise reduction device 130 </ b> A is a stereo having a low S / N ratio in the same manner as in the first embodiment while using the configuration of a reception processing unit that is generally employed. Among the frequency components in the difference signal (LR), noise components in the frequency components that are difficult to hear due to the spectrum masking effect are reduced. Therefore, according to the second embodiment, reproduction can be performed easily and rationally in the same manner as in the first embodiment while maximally utilizing the results related to the configuration of the reception processing unit that is generally employed. Noise components in the audio signal can be reduced.

[Third Embodiment]
Next, a third embodiment of the present invention will be described with reference mainly to FIGS.

<Configuration>
The broadcast receiving apparatus according to the third embodiment is different from the broadcast receiving apparatus 100A of the first embodiment described above in that a noise reducing apparatus 130C having the configuration shown in FIG. 6 is provided instead of the noise reducing apparatus 130A. ing. In the following description, the broadcast receiving apparatus of the third embodiment is referred to as “broadcast receiving apparatus 100C”.

  As shown in FIG. 6, the noise reduction device 130C is different from the noise reduction device 130A described above in that a weighting coefficient calculation unit 133C is provided instead of the weighting coefficient calculation unit 133A. Hereinafter, description will be made mainly focusing on these differences.

  The weighting coefficient calculation unit 133C holds a predetermined masking curve calculation parameter in the same manner as in the case of the weighting coefficient calculation unit 133A. The weighting coefficient calculation unit 133C holds “1” therein as an initial value of the weighting coefficient for each divided band.

  The weighting coefficient generation unit 133C receives the power spectrum distribution PSD (f) sent from the power spectrum calculation unit 132 as in the case of the weighting coefficient calculation unit 133A. Then, the weighting coefficient generation unit 133C performs power frequency distribution PSD (f) and a masking curve calculated using the masking curve calculation parameter with respect to each frequency component in the stereo difference signal (LR). Weighting coefficient information WCC is generated. The weighting coefficient information WCC thus generated is sent to the weighting unit 134.

<Operation>
Next, the operation of the broadcast receiving apparatus 100C configured as described above will be described mainly focusing on the processing by the weighting coefficient generation unit 133C.

  As a premise, the channel selection designation is already input to the input unit 180 by the user, and the channel selection command CSL corresponding to the designated desired station is sent to the reception processing unit 120A (more specifically, the RF processing unit 121). Suppose that it is sent. Further, it is assumed that a volume adjustment designation has already been input to the input unit 180 by the user, and a volume adjustment command VLC corresponding to the designated volume adjustment mode has been sent to the analog processing unit 170 (FIG. 1). reference).

When a broadcast wave is received by the antenna 110 in such a state, the power spectrum calculation unit 132 calculates the power spectrum distribution PSD (f) of the stereo sum signal (R + L), as in the case of the broadcast receiving device 100A described above. Then, it is sent to the weighting coefficient generation unit 133C. Also, the FFT unit 131 _S performs a Fourier transform on the stereo difference signal (LR), and sends the conversion result SF (f) to the weighting unit 134 (see FIG. 6).

  Upon receiving the power spectrum distribution PSD (f), the weighting coefficient generation unit 133C first calculates the total power for each predetermined number of divided bands, as in the case of the weighting coefficient generation unit 133A described above. Next, the weighting coefficient generation unit 133C determines whether or not the sound component for each divided band is difficult to hear due to the spectrum masking effect. In making this determination, the weighting coefficient generation unit 133C first identifies the divided band having the maximum total power and sets it as the focused divided band.

  Next, the weighting coefficient generation unit 133C calculates a masking curve corresponding to the sound component of the target divided band based on the total power in the target divided band and the masking curve calculation parameter held therein. Subsequently, the weighting coefficient generation unit 133C determines whether or not the total power in each of the divided bands other than the target divided band is less than the value of the masking curve at the center frequency of the divided band. Masking determination is performed as to whether or not the sound components in each of the divided bands are within the masking range determined by the masking curve and are difficult to hear.

  Next, the weighting coefficient generation unit 133C is based on the ratio value (<1) between the total power in the divided band in which the masking determination result is positive and the value of the masking curve at the center frequency of the divided band. A provisional weighting coefficient related to the divided band is calculated using a predetermined formula (hereinafter referred to as “coefficient formula”). Here, the calculated temporary weighting coefficient is a value less than “1”, and the smaller the ratio value, the smaller the calculated temporary weighting coefficient value.

  The coefficient formula is determined based on experiments, simulations, experiences, etc. while utilizing the results in auditory psychology regarding the spectrum masking effect.

  Thereafter, the weighting coefficient generation unit 133C (i) uses the divided band in which the total power is maximum among the divided bands that have not been set as the target divided band and for which the provisional weighting coefficient has not been calculated. Identification of new target divided band, (ii) Calculation of new masking curve based on total power in new target divided band and masking curve calculation parameter held inside, (iii) Use of new masking curve And (iv) based on the ratio value (<1) between the total power of the divided band in which the masking determination result is positive and the value of the masking curve at the center frequency of the divided band. The calculation of a new provisional weighting coefficient using the coefficient formula is sequentially repeated until no new divided band of interest is found.

  Next, for each divided band, the weighting coefficient generation unit 133C multiplies “1”, which is an initial value of the weighting coefficient, by all the provisional weighting coefficients calculated for the divided band. Then, a multiplication result for each divided band is determined as a weighting coefficient for each divided band. When the weighting coefficients for all the divided bands are determined, weighting coefficient generation section 133C sends the determined weighting coefficients to weighting section 134 as weighting coefficient information WCC (see FIG. 6).

In FIG. 7, the coefficient formula is FW (r) (r: value of ratio), and the total power in the divided band Rj of the center frequency fj (j = 1, 2,...) Is “SPj”. An example of calculation processing of the weighting coefficient WCCj in FIG. Here, FIG. 7A shows an example in which the result of the masking determination regarding the divided band Rj is affirmative and the number of times of calculating the provisional weighting coefficient is one. In the example of FIG. 7A, an example in which the value at the center frequency fj of the masking curve MC when the masking determination result is affirmative is “MCj” is shown. In this case, the weighting coefficient WCCj is calculated by the following equation (1).
WCCj = 1 · FW (SPj / MCj) (1)

FIG. 7B shows an example in which the masking determination result regarding the divided band Rj is affirmative, and the provisional weighting coefficient is calculated twice. In the example of FIG. 7B, the values at the center frequency fj of the masking curves MC ₁ and MC ₂ when the masking determination result is affirmative are “MCj ₁ ” and “MCj ₂ ”. An example is shown. In this case, the weighting coefficient WCCj is calculated by the following equation (2).
WCCj = 1 · FW (SPj / MCj ₁ ) · FW (SPj / MCj ₂ ) (2)

  It should be noted that the initial weight “1” is used as the final weighting coefficient for the weighting coefficient related to the divided band for which the masking determination result has never been positive, that is, the divided band for which the provisional weighting coefficient has not been calculated. It becomes.

The weighting unit 134 weights the conversion result SF (f) sent from the FFT unit 131 _{S according} to the weighting coefficient information WCC, and generates a weighting result CSF (f). That is, the weighting unit 134 generates the weighting result CSF (f) by multiplying the component of each frequency band in the conversion result SF (f) by the corresponding weighting coefficient. Then, the weighting unit 134 sends the weighting result CSF (f) to the IFFT unit 135 _S (see FIG. 6).

Then, as in the case of the first embodiment, the IFFT unit 135 _M performs inverse Fourier transform on the conversion result MF (f) sent from the power spectrum calculation unit 132 to obtain a stereo sum signal (L + R). And the reproduced stereo sum signal (L + R) is sent to the matrix unit 136. Similarly to the case of the first embodiment, the IFFT unit 135 _S performs an inverse Fourier transform on the weighted result CSF (f) sent from the weighting unit 134 to obtain a stereo difference signal (LR). The correction difference signal CST in which the noise component is reduced is calculated, and the calculated correction difference signal CST is sent to the matrix unit 136 (see FIG. 6).

  Next, the matrix unit 136 adds the stereo sum signal (L + R) and the correction difference signal CST similarly to the case of the first embodiment, and calculates the reproduction L channel signal CLD based on the addition result. . Further, the matrix unit 136 subtracts the correction difference signal CST from the stereo sum signal (L + R), and calculates the reproduction R channel signal CRD based on the subtraction result. Then, the matrix unit 136 sends the reproduction L channel signal CLD and the reproduction R channel signal CRD to the analog processing unit 140 (see FIG. 6).

Thereafter, similarly to the case of the first embodiment, the analog processing unit 140 generates output sound signals AOS _L and AOS _R from the reproduction L channel signal CLD and the reproduction R channel signal CRD, and the speaker units 150 _L , Send to 150 _R. Then, the speaker units 150 _L and 150 _R reproduce and output sound according to the output sound signals AOS _L and AOS _R.

  As described above, in the third embodiment, the power spectrum calculation unit 132 determines the power of the stereo sum signal (R + L) based on the time frequency transform result (Fourier transform result) MF (f) for the stereo sum signal (R + L). A spectral distribution PSD (f) is calculated. Subsequently, the weighting coefficient generation unit 133C generates a weighting coefficient for each frequency component in the stereo difference signal (LR) corresponding to the audibility by the spectrum masking effect based on the power spectrum distribution PSD (f).

Next, after the weighting unit 134 performs weighting using the weighting coefficient on the time-frequency conversion result SF (f) for the stereo difference signal (LR), the IFFT unit 135 _S performs weighting by the weighting unit 134. The result CSF (f) is subjected to frequency-time transform (inverse Fourier transform) to calculate a corrected difference signal CST in which the noise component in the stereo difference signal (LR) is reduced. Then, the matrix unit 136 performs the reproduction L channel signal and the stereo difference signal CST based on the stereo sum signal (L + R) reproduced by performing the time-frequency conversion on the stereo sum signal (L + R), and the correction difference signal CST. A reproduction R channel signal is calculated.

  That is, in the third embodiment, among the frequency components in the stereo difference signal (LR) having a low S / N, the contribution of the noise components in the frequency components that are difficult to hear due to the spectrum masking effect to the reproduced sound is suppressed. It has come to be. Therefore, according to the third embodiment, as in the case of the first embodiment, the noise component in the reproduced audio signal can be reduced easily and rationally.

  In the third embodiment, the weighting coefficient generation unit 133C calculates the power of each frequency component in the stereo sum signal (L + R) based on the power of other frequency components and a predetermined masking curve calculation parameter. A masking determination is made as to whether or not the masking range is determined by the masking curve corresponding to the other frequency component. Then, the weighting coefficient generation unit 133C, for the frequency component for which the masking determination result is affirmative, the weighting coefficient generation unit 133C determines the power of the frequency component for which the weighting coefficient is generated and the frequency component for the generation target. A weighting coefficient value is generated based on the amount of difference from the masking curve value in frequency.

  Therefore, according to the third embodiment, the noise component in the reproduced audio signal can be reduced easily and very reasonably.

[Modification of Embodiment]
The present invention is not limited to the above-described embodiment, and various modifications are possible.

  For example, in the above third embodiment, the divided band to which the generation target frequency component belongs as the difference between the power of the generation target frequency component of the weighting coefficient and the value of the masking curve at the frequency of the generation target frequency component. And the value of the masking curve at the frequency of the frequency component to be generated and the value of the ratio. On the other hand, as the difference amount, other types of amounts may be adopted as long as they reflect the spectrum masking effect.

  In the third embodiment, the weighting coefficient is generated based only on the spectrum masking effect by the sound component of the divided band specified as the target divided band. On the other hand, the weighting coefficient may be generated based on the spectrum masking effect by the sound components of all the divided bands.

  Moreover, in said 1st-3rd embodiment, the stereo sum signal supplied to the matrix part in a noise reduction apparatus is obtained by carrying out an inverse Fourier transform, after carrying out the Fourier transform of the stereo sum signal sent from the reception processing unit. Thus, the stereo sum signal supplied to the matrix portion and the correction difference signal are synchronized. On the other hand, a delay corresponding to the sum of the Fourier transform time and the inverse Fourier transform time included in the correction difference signal generation process is applied to the stereo sum signal sent from the reception processing unit, and the noise reduction device You may make it supply to a matrix part.

  Moreover, you may perform the deformation | transformation to 2nd Embodiment with respect to said 1st Embodiment with respect to 3rd Embodiment.

  In addition, the noise reduction apparatus in the above embodiment is configured as a computer as a calculation unit including a DSP (Digital Signal Processor) and the like, and a program prepared in advance is executed on the computer, whereby in the above embodiment Some or all of the processing may be executed. This program is recorded on a computer-readable recording medium such as a hard disk, CD-ROM, or DVD, and is read from the recording medium and executed by the computer. The program may be acquired in a form recorded on a portable recording medium such as a CD-ROM or DVD, or may be acquired in a form distributed via a network such as the Internet. Also good.

100A to 100C ... Broadcast receiving device 120A, 120B ... Reception processing unit (reception processing unit)
130A, 130C ... Noise reduction device 132 ... Power spectrum calculation unit 133A, 133C ... Weighting coefficient generation unit 134 ... Weighting unit 136 ... Matrix unit (channel signal calculation unit)

Claims (8)

A noise reduction device included in a broadcast receiving device that receives FM stereo broadcast,
A power spectrum calculation unit for calculating a power spectrum distribution of the stereo sum signal based on a time-frequency conversion result for the stereo sum signal;
A weighting coefficient generator for generating a weighting coefficient for each frequency component in the stereo difference signal corresponding to the audibility by the spectrum masking effect based on the calculated power spectrum distribution;
A weighting unit that performs weighting using the generated weighting coefficient on the time-frequency conversion result for the stereo difference signal;
A frequency time conversion unit that performs frequency time conversion on the weighting result by the weighting unit;
A channel signal calculation unit that calculates a reproduction left channel signal and a reproduction right channel signal based on the stereo sum signal and a conversion result by the frequency time conversion unit;
A noise reduction device comprising: The weighting coefficient generator is
Whether the power of each frequency component is within the masking range determined by the masking curve corresponding to the other frequency component calculated based on the power of the other frequency component and a predetermined masking curve calculation parameter Make a masking decision,
For frequency components for which the result of the masking determination is negative, the weighting coefficient is “1”.
The noise reduction device according to claim 1.   The noise reduction apparatus according to claim 2, wherein the weighting coefficient generation unit sets a weighting coefficient to “0” for a frequency component for which the result of the masking determination is positive.   For the frequency component for which the result of the masking determination is affirmative, the weighting coefficient generation unit includes the power of the frequency component for which the weighting coefficient is generated and the value of the masking curve at the frequency of the frequency component for the generation The noise reduction device according to claim 2, wherein a value of a weighting coefficient is generated based on the difference amount of the noise. A reception processing unit that generates a stereo sum signal and a stereo difference signal based on the reception result of the FM stereo broadcast wave;
A noise reduction device according to any one of claims 1 to 4;
A broadcast receiving apparatus comprising: A noise reduction method used in a noise reduction device included in a broadcast reception device that receives FM stereo broadcast,
A power spectrum calculating step of calculating a power spectrum distribution of the stereo sum signal based on a time-frequency conversion result for the stereo sum signal;
A weighting coefficient generation step of generating a weighting coefficient for each frequency component in the stereo difference signal corresponding to the audibility by the spectrum masking effect based on the calculated power spectrum distribution;
A weighting step of performing weighting using the generated weighting coefficient on the time-frequency conversion result for the stereo difference signal;
A frequency time conversion step of performing frequency time conversion on the weighting result in the weighting step;
A channel signal calculation step of calculating a reproduction left channel signal and a reproduction right channel signal based on the stereo sum signal and the conversion result in the frequency time conversion step;
A noise reduction method comprising:   A noise reduction program for causing a computer included in a noise reduction device included in a broadcast reception device that receives FM stereo broadcast to execute the noise reduction method according to claim 6.   A recording medium on which the noise reduction program according to claim 7 is recorded so as to be readable by a computer provided in a noise reduction device included in a broadcast reception device that receives FM stereo broadcast.

Images