JP5845954B2 - Noise reduction device, voice input device, wireless communication device, noise reduction method, and noise reduction program - Google Patents

Description

  The present invention relates to a noise reduction device, a voice input device, a wireless communication device, a noise reduction method, and a noise reduction program.

  Noise reduction processing technology is widely used, for example, in the field of mobile communication. In mobile communication, noise is mixed when speech is encoded at a low bit rate, and speech intelligibility deteriorates. Therefore, in order to avoid this, noise is detected by detecting noise before encoding the voice.

  In the noise reduction processing technology, for example, noise (for example, unnecessary sound other than the desired sound) is mainly obtained from a sound signal collected by a microphone that mainly collects sound (for example, desired sound such as a sound emitted from a caller). By subtracting the noise signal collected by the microphone that picks up the sound, the noise component contained in the audio signal can be removed.

  Japanese Patent Application Laid-Open No. 2004-228688 discloses a technique for preventing voice quality from being lowered and processing performance from being removed from noise, so that the voice can always be heard in the best state. Patent Document 2 discloses a technique for preventing reduction of desired speech and reducing only unnecessary sound to be reduced in an adaptive noise reduction type speech input device. Patent Document 3 discloses a technique for realizing higher voice quality in which noise is suppressed.

JP-A-7-168586 JP-A-6-67692 JP 2011-99967 A

  When noise reduction processing is performed using an audio microphone that mainly collects audio and a reference sound microphone that mainly collects noise, the amount of audio reduction (cancellation) increases depending on the direction of noise arrival. There was a problem. That is, depending on how the noise reduction device is used, the sound may be mixed into the reference sound microphone that collects noise. When sound is mixed in the reference sound microphone in this way, not only the noise component mixed in the sound collected by the sound microphone but also the sound itself is canceled, resulting in a problem that sound clarity is lowered. there were.

  In view of the above problems, an object of the present invention is to provide a noise reduction device, a voice input device, a wireless communication device, a noise reduction method, and a noise reduction program that can appropriately reduce noise components.

The noise reduction device according to the present invention includes a first sound collection signal corresponding to a sound collected by a first microphone and a second sound collection signal corresponding to a sound collected by a second microphone. An audio noise interval detection unit that detects an audio interval and a noise interval based on at least one of the above, and an audio power that is a difference between the magnitude of the first collected signal and the magnitude of the second collected signal in the audio section A power information acquisition unit that acquires a difference and a noise power difference that is a difference between the magnitude of the first sound pickup signal and the magnitude of the second sound pickup signal in the noise section; A noise reduction process determination unit that determines the state of the noise power difference, and a noise reduction process using the first sound collection signal and the second sound collection signal according to the determination result of the noise reduction process determination unit To implement Comprising a's reduction processing unit.
The noise reduction processing determination unit may determine a state of a difference between the audio power difference and the noise power difference.
The noise reduction processing determination unit may determine whether the state of the audio power difference and the noise power difference is a first state where noise reduction processing is performed, and the noise reduction processing unit When the noise reduction process determination unit determines that the first state is established, the noise reduction process may be performed using the first sound collection signal and the second sound collection signal.
The noise reduction processing determination unit determines whether the state of the audio power difference and the noise power difference is a second state in which noise reduction processing is performed weaker than the first state. The noise reduction processing unit may perform the noise reduction processing weaker than the first state when the noise reduction processing determination unit determines that the second state is present.
The noise reduction processing determination unit may determine whether the state of the audio power difference and the noise power difference is a second state in which noise reduction processing is not performed, and the noise reduction processing The unit may not perform the noise reduction process when the noise reduction process determination unit determines that the state is the second state.
The noise reduction processing determination unit may determine the second state when the absolute value of the difference between the audio power difference and the noise power difference is within a predetermined threshold.
The power information acquisition unit may update the audio power difference in the audio interval and update the noise power difference in the noise interval.
The noise reduction processing unit reduces a noise component included in the first collected sound signal using the second collected sound signal when the noise reduction processing determining unit determines that the state is the first state. The signal after the noise reduction processing may be output as an audio signal, and the first sound collection signal may be output as an audio signal when the noise reduction processing determination unit determines that the second state has occurred.
The noise reduction processing unit is configured such that the magnitude of the second collected sound signal in the voice section is earlier than the magnitude of the first collected sound signal, and the noise reduction processing determining unit If determined, the noise component included in the second sound collection signal is reduced using the first sound collection signal, the signal after the noise reduction processing is output as a sound signal, When the magnitude of the second collected sound signal is earlier than the magnitude of the first collected sound signal and the noise reduction processing determination unit determines that the second state is detected, the second collected sound signal May be output as an audio signal.
The voice input device according to the present invention includes the noise reduction device.
In the voice input device, the first microphone is provided on a first surface of the voice input device, and the second microphone is opposed to the first surface with a predetermined distance therebetween. It may be provided on the surface.
A wireless communication device according to the present invention includes the noise reduction device.
In the wireless communication device, the first microphone is provided on a first surface of the wireless communication device, and the second microphone is opposed to the first surface with a predetermined distance therebetween. It may be provided on the surface.
The noise reduction method according to the present invention includes a first sound collection signal corresponding to a sound collected by a first microphone and a second sound collection signal corresponding to a sound collected by a second microphone. A voice interval and a noise interval are detected based on at least one of the following: a voice power difference that is a difference between the magnitude of the first collected sound signal and the second collected sound signal in the voice interval; Obtaining a noise power difference that is a difference between the magnitude of the first collected sound signal and the magnitude of the second collected sound signal, determining a state of the sound power difference and the noise power difference, and Noise reduction processing is performed using the first sound collection signal and the second sound collection signal according to the result of determining the power difference and the noise power difference state.
A noise reduction program according to the present invention causes a computer to collect a first sound collection signal corresponding to a sound collected by a first microphone and a second sound collection corresponding to a sound collected by a second microphone. A voice section and a noise section are detected based on at least one of the signals, and a voice power difference that is a difference between a magnitude of the first collected signal and a magnitude of the second collected signal in the voice section; A noise power difference that is a difference between the magnitude of the first sound pickup signal and the magnitude of the second sound pickup signal in the noise section is acquired, and the state of the voice power difference and the noise power difference is determined. A noise reduction program for performing noise reduction processing using the first sound collection signal and the second sound collection signal according to a result of determining the state of the sound power difference and the noise power difference. It is.

  According to the present invention, it is possible to provide a noise reduction device, a voice input device, a wireless communication device, a noise reduction method, and a noise reduction program that can appropriately reduce noise components.

It is a block diagram which shows the noise reduction apparatus concerning Embodiment 1. FIG. It is a block diagram which shows an example of the audio | voice noise area detection part with which the noise reduction apparatus concerning Embodiment 1 is provided. It is a block diagram which shows the other example of the audio | voice noise area detection part with which the noise reduction apparatus concerning Embodiment 1 is provided. It is a block diagram which shows an example of the phase information acquisition part with which the noise reduction apparatus concerning Embodiment 1 is provided. It is a figure which shows an example of the position of the sound source of an audio | voice and a noise with respect to the audio | voice microphone with which the noise reduction apparatus concerning Embodiment 1 and a reference sound microphone are equipped. It is a figure which shows an example of the position of the sound source of an audio | voice and a noise with respect to the audio | voice microphone with which the noise reduction apparatus concerning Embodiment 1 and a reference sound microphone are equipped. It is a figure which shows an example of the position of the sound source of an audio | voice and a noise with respect to the audio | voice microphone with which the noise reduction apparatus concerning Embodiment 1 and a reference sound microphone are equipped. FIG. 3 is a block diagram illustrating an example of a noise reduction processing unit included in the noise reduction device according to the first exemplary embodiment; 3 is a flowchart for explaining an operation of the noise reduction apparatus according to the first exemplary embodiment; It is a figure which shows an example of the audio | voice input apparatus using the noise reduction apparatus concerning Embodiment 1. FIG. 1 is a diagram illustrating an example of a wireless communication device using a noise reduction device according to a first exemplary embodiment; It is a figure for demonstrating the effect of the invention concerning Embodiment 1. FIG. It is a figure for demonstrating the effect of the invention concerning Embodiment 1. FIG. It is a block diagram which shows the noise reduction apparatus concerning Embodiment 2. It is a block diagram which shows an example of the power information acquisition part with which the noise reduction apparatus concerning Embodiment 2 is provided. 10 is a flowchart for explaining the operation of the noise reduction apparatus according to the second exemplary embodiment; It is a block diagram which shows the other example of the noise reduction apparatus concerning Embodiment 2. FIG. FIG. 6 is a block diagram illustrating a noise reduction device according to a third exemplary embodiment. It is a block diagram which shows an example of the noise reduction process part with which the noise reduction apparatus concerning Embodiment 3 is provided. 10 is a flowchart for explaining the operation of the noise reduction apparatus according to the third exemplary embodiment; It is a figure which shows an example of the position of the sound source of an audio | voice and a noise with respect to the audio | voice microphone with which the noise reduction apparatus concerning Embodiment 3 and a reference sound microphone are equipped.

<Embodiment 1>
Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram of the noise reduction apparatus according to the first embodiment. As shown in FIG. 1, the noise reduction apparatus 1 according to the present embodiment includes an audio microphone 11, a reference sound microphone 12, AD converters 13 and 14, an audio noise section detection unit 15, a phase information acquisition unit 16, and noise. A reduction processing determination unit 17 and a noise reduction processing unit 18 are included.

  Each of the sound microphone 11 and the reference sound microphone 12 can collect a sound including a sound component and a noise component. The sound microphone 11 picks up a sound mainly containing a sound component, converts it into an analog signal, and outputs the converted analog signal to the AD converter 13. The reference sound microphone 12 collects a sound mainly including a noise component, converts it into an analog signal, and outputs the converted analog signal to the AD converter 14. For example, the noise component included in the sound collected by the reference sound microphone 12 is used to reduce the noise component contained in the sound collected by the sound microphone 11.

  In the noise reduction device according to the present embodiment, the configuration in the case of two microphones (that is, the voice microphone 11 and the reference sound microphone 12) will be described. For example, a reference sound microphone is further added to the microphone. Three or more may be provided.

  The AD converter 13 samples the analog signal output from the audio microphone 11 at a predetermined sampling rate and converts it into a digital signal, and generates a sound pickup signal 21. The collected sound signal 21 generated by the AD converter 13 is output to the audio noise section detection unit 15, the phase information acquisition unit 16, and the noise reduction processing unit 18.

  The AD converter 14 samples the analog signal output from the reference sound microphone 12 at a predetermined sampling rate, converts the analog signal into a digital signal, and generates a sound collection signal 22. The collected sound signal 22 generated by the AD converter 14 is output to the phase information acquisition unit 16 and the noise reduction processing unit 18.

In the present embodiment, the frequency band of the sound input to the sound microphone 11 and the reference sound microphone 12 is approximately 100 Hz to 4000 Hz. Therefore, by setting the sampling frequency in the AD converters 13 and 14 to about 8 kHz to 12 kHz, an analog signal including an audio component can be handled as a digital signal.
In the present specification, a sound collection signal mainly including a sound component is also referred to as a sound signal, and a sound collection signal mainly including a noise component is also referred to as a reference signal.

  The voice noise section detector 15 detects a voice section and a noise section based on the collected sound signal 21 output from the AD converter 13. Then, the voice noise section detection unit 15 outputs the voice noise section information 23 and 24 indicating the voice section and the noise section to the phase information acquisition unit 16 and the noise reduction processing unit 18, respectively.

  Any technique can be used for the voice noise section detection processing in the voice noise section detector 15. When the noise reduction device is used in an environment where the noise level is high, it is preferable to determine the speech section and the noise section with high accuracy. For example, the speech noise section detection technique A or the speech noise section detection technique described later is used. By using B, it is possible to detect the voice section and the noise section with high accuracy. The sound includes sounds other than human voices, but in these examples, human voices are mainly detected. Note that the speech noise section detection technique A is also described as an example in Japanese Patent Application No. 2011-254578, which is an application claiming priority based on Japanese Patent Application No. 2010-260798. The voice noise section detection technique B is also described in Japanese Patent Application No. 2011-020659 as an example.

  First, the voice noise section detection technique A will be described. In the speech noise section detection technique A, the speech section is determined by paying attention to the frequency spectrum of the vowel component that is the main part of the speech. In the speech noise section detection technique A, an appropriate noise level is set for each band, a signal to noise level ratio with a peak of a vowel frequency component is obtained, and the signal to noise level ratio is a predetermined level ratio and a predetermined number of peaks. The voice section is determined by observing whether or not there is.

  FIG. 2 is a block diagram illustrating an example of a speech noise section detection unit 15 ′ using the speech noise section detection technique A. 2 includes a framing unit 31, a spectrum generating unit 32, a band dividing unit 33, a frequency averaging unit 34, a holding unit 35, a time averaging unit 36, a peak detecting unit 37, and a voice determination unit. The unit 38 is provided.

  The framing unit 31 sequentially cuts the sound pickup signal 21 in frame units (predetermined number of samples) having a predetermined time width, and generates an input signal in frame units (hereinafter referred to as a framed input signal).

  The spectrum generation unit 32 performs frequency analysis of the framing input signal output from the framing unit 31, converts the time-domain framing input signal into the frequency-domain framing input signal, and collects the spectrum. Is generated. The spectrum pattern is a collection of spectra for each frequency in which a frequency and energy at the frequency are associated with each other over a predetermined frequency band. The frequency transform method used here is not limited to a specific means, but requires a frequency resolution necessary for recognizing the spectrum of speech, and therefore has a relatively high resolution such as FFT (Fast Fourier Transform) or DCT (Discrete). It is recommended to use an orthogonal transformation method such as Cosine Transform. In the present embodiment, the spectrum generation unit 32 generates a spectrum pattern of at least 200 Hz to 700 Hz.

  A spectrum (hereinafter referred to as a formant) that indicates a feature of a voice, which is a target to be detected when a voice determination unit 38 to be described later determines a voice section, usually includes a harmonic part from a first formant corresponding to a fundamental tone. There are a plurality of nth formants (where n is a natural number). Of these, the first formant and the second formant often exist in a frequency band of less than 200 Hz. However, since this band contains a low-frequency noise component with relatively high energy, formants are easily buried. Also, a formant of 700 Hz or more is easily buried in a noise component because the formant itself has low energy. Therefore, by using a spectrum pattern of 200 Hz to 700 Hz that is difficult to be buried in the noise component for the determination of the voice section, the determination target can be narrowed down and the voice section can be determined efficiently.

  In order to detect a spectrum characteristic of speech in an appropriate frequency band unit, the band dividing unit 33 divides each spectrum of the spectrum pattern into a plurality of divided frequency bands that are frequency bands divided by a predetermined bandwidth. To divide. In the present embodiment, the predetermined bandwidth is about 100 Hz to 150 Hz.

  The frequency averaging unit 34 calculates average energy for each divided frequency band. In the present embodiment, the frequency averaging unit 34 averages the energy of all spectra in the divided frequency band for each divided frequency band. However, the maximum or average amplitude value of the spectrum is used instead of the spectrum energy in order to reduce the calculation load. (Absolute value) may be substituted.

  The holding unit 35 is configured by a storage medium such as a RAM (Random Access Memory), an EEPROM (Electrically Erasable and Programmable Read Only Memory), and a flash memory, and the average energy for each band is set to a predetermined number in the past (this embodiment). N frames in the form) are held.

  The time averaging unit 36 derives, for each divided frequency band, band-specific energy that is an average over a plurality of frames in the time direction of the average energy derived by the frequency averaging unit 34. That is, the band-specific energy is an average value over a plurality of frames in the time direction of the average energy for each divided frequency band. In addition, the time averaging unit 36 may obtain a substitute value of the band-specific energy by performing a process according to averaging using the weighting coefficient and the time constant on the average energy for each divided frequency band of the immediately preceding frame.

  The peak detector 37 derives an energy ratio (SNR: Signal to Noise ratio) between each spectrum of the spectrum pattern and the band-specific energy in the divided frequency band in which the spectrum is included. Then, the peak detection unit 37 compares the SNR for each spectrum with a predetermined threshold A and determines whether or not the threshold A is exceeded. If there is a spectrum whose SNR exceeds the threshold A, this spectrum is regarded as a formant, and information indicating that a formant has been detected is output to the voice determination unit 38.

  When receiving information from the peak detection unit 37 that the formant has been detected, the audio determination unit 38 determines whether the framed input signal of the corresponding frame is audio based on the determination result of the peak detection unit 37. When the speech determination unit 38 determines that the framed input signal is speech, the speech determination unit 38 outputs speech noise section information 23 and 24 indicating the speech section to the phase information acquisition unit 16 and the noise reduction processing unit 18, respectively. On the other hand, when the speech determination unit 38 determines that the framed input signal is not speech, the speech determination unit 38 outputs speech noise section information 23 and 24 indicating noise sections to the phase information acquisition unit 16 and the noise reduction processing unit 18, respectively.

  The audio noise section detection unit 15 ′ illustrated in FIG. 2 sets energy for each divided frequency band for each divided frequency band. Therefore, the voice determination unit 38 can accurately determine the presence / absence of a formant for each divided frequency band without being affected by noise components in other divided frequency bands.

  As described above, there are a plurality of formants from the first formant to the n-th formant, which is a harmonic part thereof. Therefore, even if the energy (noise level) of any divided frequency band is increased and a part of the formant is buried in noise, a plurality of other formants may be detected. In particular, since ambient noise is concentrated in the low range, even if the first formant corresponding to the fundamental tone and the second formant corresponding to the second overtone are buried in the low-frequency noise, the possibility of detecting a formant with a third or higher harmonic is possible. There is. Therefore, the speech determination unit 38 can determine a speech section that is more resistant to noise by determining that the framed input signal is speech when the number of spectrums whose SNR exceeds the threshold A is a predetermined number or more. .

  As described above, the voice noise section detection unit 15 ′ using the voice noise section detection technique A cuts the input signal in units of frames having a predetermined time width, and generates a framed input signal. Unit 31, a spectrum generation unit 32 that converts a framing input signal from the time domain to the frequency domain and generates a spectrum pattern in which spectra for each frequency are collected, each spectrum of the spectrum pattern, and a predetermined band Peak detection unit for determining whether the energy ratio of the divided frequency band including the spectrum among the plurality of divided frequency bands which are frequency bands divided by the width exceeds a predetermined threshold A 37 and voice determination for determining whether the framed input signal is voice based on the determination result of the peak detection unit Unit 38, frequency average unit 34 for deriving the average energy in the frequency direction of the spectrum in each divided frequency band of the spectrum pattern, and deriving the band-specific energy that is the average of the average energy in the time direction for each divided frequency band A time averaging unit 36.

  For example, the speech determination unit 38 can determine that the framed input signal is speech when the spectrum in which the energy ratio exceeds the threshold A is greater than or equal to a predetermined number.

  Next, the voice noise section detection technique B will be described. In the speech noise section detection technique B, the speech section is determined by paying attention to the property that the spectrum pattern that is a feature of the consonant tends to rise to the right. In the voice noise section detection technique B, the spectrum pattern of the consonant is measured in the mid-high frequency band, and the characteristics of the frequency distribution of the consonant that is partially buried by the noise component are not affected by the noise. This makes it possible to determine a speech segment with high accuracy.

  FIG. 3 is a block diagram illustrating an example of the voice noise section detection unit 15 ″ using the voice noise section detection technique B. The audio noise section detection unit 15 ″ includes a framing unit 41, a spectrum generation unit 42, a band division unit 43, an average derivation unit 44, a noise level derivation unit 45, a determination selection unit 46, and a consonant determination unit 47.

  The framing unit 41 sequentially extracts the sound pickup signal 21 in units of frames having a predetermined time width, and generates a framing input signal that is an input signal in units of frames.

  The spectrum generation unit 42 performs frequency analysis of the framing input signal output from the framing unit 41, converts the time-domain framing input signal into the frequency-domain framing input signal, and collects the spectrum. Is generated. The spectrum pattern is a collection of spectra for each frequency in which a frequency and energy at the frequency are associated with each other over a predetermined frequency band. The frequency conversion method used here is not limited to a specific means, but a frequency resolution necessary for recognizing a speech spectrum is necessary, and therefore, an orthogonal transformation method such as FFT or DCT having a relatively high resolution is used. Good.

  The band dividing unit 43 divides each spectrum of the spectrum pattern generated by the spectrum generating unit 42 for each predetermined bandwidth, and generates a plurality of divided frequency bands. In the present embodiment, the band dividing unit 43 divides the frequency range of, for example, 800 Hz to 3.5 kHz for each bandwidth of about 100 Hz to 300 Hz, for example.

  The average deriving unit 44 derives average energy for each band, which is an average energy for each divided frequency band (band) divided by the band dividing unit 43 in the spectrum pattern.

  The consonant determination unit 47 compares the band-by-band average energies derived by the average deriving unit 44. If the band-by-band average energy of the higher frequency band is higher, the consonant is included in the framed input signal. It is determined that

  In general, consonants tend to have a spectral pattern that rises to the right. Therefore, the speech noise section detection unit 15 ″ using the speech noise section detection technique B derives the average energy for each band in the spectrum pattern, and compares the energy for each band with each other, and the spectrum pattern characteristic to the consonant Detecting the upward trend in Therefore, the speech noise section detection unit 15 ″ can accurately detect a consonant section in which a consonant is included in the input signal.

  The consonant determination unit 47 counts a combination in which the average energy for each band between adjacent bands is higher than the adjacent low frequency band in the high frequency band, and the counted value is equal to or greater than a predetermined threshold A. If it is, the 1st determination means which determines that a consonant is contained is provided. In addition, the consonant determination unit 47 measures a combination in which the average energy for each band between adjacent bands is higher in the high frequency band than in the adjacent low frequency band, and when this combination continues across the bands And a second determination means for determining that a consonant is included when the counted value is equal to or greater than a predetermined threshold value B. The consonant determination unit 47 uses the first determination unit and the second determination unit in accordance with the noise level.

  Here, the noise level deriving unit 45 derives the noise level of the framed input signal so as to select the first determination unit and the second determination unit as appropriate. For example, the noise level can be an average value of average energy for each frequency band of the framed input signal. Further, the noise level deriving unit 45 may derive a noise level for each framed input signal, or may use an average value of noise levels of the framed input signal for a predetermined time. The determination selection unit 46 selects the first determination unit when the derived noise level is less than the predetermined threshold, and selects the second determination unit when the derived noise level is equal to or higher than the predetermined threshold.

  As described above, the speech noise section detection unit 15 '' using the speech noise section detection technique B includes the framing unit 41 that cuts out an input signal in units of predetermined frames and generates a framed input signal. A spectrum generation unit 42 that converts a framed input signal from a time domain to a frequency domain to generate a spectrum pattern in which spectra for each frequency are collected, and an average for each predetermined predetermined bandwidth in the spectrum pattern. The average derivation unit 44 for deriving the average energy for each band, which is energy, compares the derived average energy for each band, and if the average energy for each band in the higher frequency band is higher, the framed input A consonant determination unit 47 that determines that a consonant is included in the signal.

  For example, the consonant determination unit 47 counts combinations in which the average energy for each band between adjacent bands of the spectrum pattern is larger in the higher frequency band than in the adjacent lower frequency band, and the counted value is determined in advance. It is possible to determine that a consonant is included if it is equal to or greater than the threshold value.

  In addition, when applying said audio | voice noise area detection technology A and B to the noise reduction apparatus concerning this Embodiment, a parameter can be set for every product. That is, when the speech noise section detection techniques A and B are applied to a product that requires more reliable speech section determination, a stricter threshold can be set as a speech section determination parameter.

  The phase information acquisition unit 16 of the noise reduction apparatus 1 illustrated in FIG. 1 calculates a sound phase difference that is a phase difference between the sound collection signal 21 and the sound collection signal 22 in the sound section when the sound noise section information 23 indicates a sound section. get. Moreover, the phase information acquisition part 16 acquires the noise phase difference which is a phase difference of the sound collection signal 21 and the sound collection signal 22 in a noise area, when the audio | voice noise area information 23 shows a noise area. The acquired audio phase difference and noise phase difference are supplied as phase information 25 to the noise reduction processing determination unit 17.

  For example, when the noise reduction device according to the present embodiment is applied to a portable device (wireless communication device) such as a transceiver or a small device such as a speaker microphone (voice input device) used in the wireless communication device (FIG. 10). 11), the sound microphone 11 is provided on the front side where it is easy to pick up the sound, and the reference sound microphone 12 is provided on the back side where it is difficult to pick up the sound. Thereby, the sound microphone 11 can mainly collect sound components, and the reference sound microphone 12 can mainly collect noise components.

  The above-described wireless communication device and voice input device are generally a little smaller than a human fist. Therefore, the difference between the distance between the sound source and the sound microphone 11 and the distance between the sound source and the reference sound microphone 12 is considered to be about 5 to 10 cm, although it differs depending on the device and the arrangement of the microphones. Here, assuming that the spatial transmission speed of sound is 34000 cm / s, the distance that the sound is transmitted between one sample is 34000 ÷ 8000 = 4.25 when the sampling frequency is 8 kHz, so that 4.25 cm. If the distance between the sound microphone 11 and the reference sound microphone 12 is 5 cm, a sampling frequency of 8 kHz is insufficient to estimate the sound direction.

  In this case, if the sampling frequency is set to 24 kHz, which is three times 8 kHz, 34000 / 24000≈1.42 cm, and 3 to 4 phase difference points can be measured within 5 cm. Therefore, when detecting the direction of arrival of the sound based on the phase difference between the collected sound signal 21 and the collected sound signal 22, the sampling frequency of the collected sound signal 21 and the collected sound signal 22 input to the phase information acquisition unit 16 is set to 24 kHz. This should be done.

  In the noise reduction apparatus 1 shown in FIG. 1, for example, when the sampling frequency of the collected sound signals 21 and 22 output from the AD converters 13 and 14 is 8 to 12 kHz, the AD converters 13 and 14 and the phase information acquisition unit 16 Between them, a sampling frequency converter may be provided to convert the sampling frequency of the collected sound signals 21 and 22 supplied to the phase information acquisition unit 16 to 24 kHz or higher.

  On the other hand, for example, when the sampling frequency of the collected sound signals 21 and 22 output from the AD converters 13 and 14 is 24 kHz or more, between the AD converter 13 and the audio noise section detector 15, and between the AD converters 13 and 14, A sampling frequency converter is provided between the noise reduction processing unit 18 and the sampling frequency of the collected sound signals 21 and 22 supplied to the audio noise interval detection unit 15 and the noise reduction processing unit 18 is converted to 8 to 12 kHz. Also good.

  The phase difference between the collected sound signal 21 and the collected sound signal 22 indicates the arrival direction of sound or noise with respect to the position of the sound microphone 11. For example, when a speaker (sound source) speaks from the voice microphone 11 side on an extension line connecting the voice microphone 11 and the reference sound microphone 12 with a straight line, the phase difference becomes the largest in the positive direction. In other words, the time difference between the microphones when the sound reaches the sound microphone 11 and the reference sound microphone 12 becomes the largest in the positive direction (that is, the sound reaches the sound microphone 11 earliest).

  On the other hand, when the speaker (sound source) speaks from the reference sound microphone 12 side on the extension line connecting the sound microphone 11 and the reference sound microphone 12 with a straight line, the phase difference is greatest in the negative direction. In other words, the time difference between the microphones when the sound reaches the sound microphone 11 and the reference sound microphone 12 is greatest in the negative direction (that is, the sound reaches the sound microphone 11 the latest).

  In addition, a position on the vertical bisector of the line connecting the sound microphone 11 and the reference sound microphone 12 by the speaker (sound source) (that is, a position intermediate between the sound microphone 11 and the reference sound microphone 12). ), The sound reaches each microphone at the same time, so the phase difference (time difference) is zero.

  In this way, by detecting the position having the highest correlation using the collected sound signal 21 from the sound microphone 11 and the collected sound signal 22 from the reference sound microphone 12, the collected sound signal 21 and the collected sound signal are detected. The phase difference can be acquired with any one of 22 as a reference. Hereinafter, a case where the sound collection signal 21 from the sound microphone 11 is used as a reference will be described as an example.

  FIG. 4 is a block diagram illustrating an example of a phase information acquisition unit included in the noise reduction device 1 according to the present embodiment. The phase information acquisition unit 16 illustrated in FIG. 4 includes a reference signal buffer 51, a reference signal extraction unit 52, a comparison signal buffer 53, a comparison signal extraction unit 54, a cross correlation value calculation unit 55, a phase difference acquisition unit 56, and an audio phase difference storage. Unit 57, noise phase difference storage unit 58, and selector 59.

  The reference signal buffer 51 temporarily stores the collected sound signal 21 output from the AD converter 13. The comparison signal buffer 53 temporarily accumulates the sound collection signal 22 output from the AD converter 14.

  The sound and noise generated at the same time with a single sound source have different transmission paths to the microphones 11 and 12, and therefore the phases and amplitude values detected by the microphones 11 and 12 are different. However, when there is a single sound or noise source, the phases and amplitude values of the sound components detected by the microphones 11 and 12 are similar and the correlation is very high. In particular, in the present embodiment, since the voice is collected in the voice section and the noise is collected in the noise section, the correlation between the voice components detected by the microphones 11 and 12 and the correlation between the noise components are very high. It can be said. Therefore, the phase difference can be obtained by measuring this correlation, and the direction of the sound source can be estimated. The phase difference between the two microphones 11 and 12 can be calculated using, for example, a cross correlation function or a least square method.

In general, the cross-correlation function between two signal waveforms x1 (t) and x2 (t) can be expressed by the following equation.

  The reference signal extraction unit 52 extracts and fixes the signal waveform x1 (t) included in the collected sound signal (reference signal) 21. The comparison signal extraction unit 54 extracts the signal waveform x2 (t) included in the collected sound signal (comparison signal) 22, and moves the signal waveform x2 (t). The cross-correlation value calculation unit 55 performs a convolution operation (product-sum operation) on the signal waveform x1 (t) and the signal waveform x2 (t), so that the correlation between the sound collection signal 21 and the sound collection signal 22 is increased. Judge the high point. At this time, the convolution calculation value is calculated while shifting the signal waveform x2 (t) back and forth according to the maximum phase difference calculated from the sampling frequency of the sound pickup signal 22 and the spatial distance between the microphones 11 and 12. The point where the convolution calculation value is the maximum is the place where the codes match, and it can be determined that the correlation is the highest.

  More specifically, for example, when the time width (number of samples) for comparing the correlation is 200 [sample], the sound collection signal (reference signal) 21 is fixed and the sound collection signal (to be compared) ( The cross-correlation value can be calculated by moving the comparison signal) 22 from the head of the sample at the same time to the point of −L [sample] from the point of −L [sample]. Here, L can specify the maximum value from the sampling frequency when the sound pickup signal 21 is digitally converted and the distance between the microphones 11 and 12. The τ-th cross-correlation value (τ) can be obtained using Equation 1 above. At this time, the range of τ is from −L to + L, and N = 200.

  All cross-correlation values (τ) are obtained, and τ [sample] having the highest cross-correlation value is extracted. The resolution changes according to the sampling frequency of the AD converters 13 and 14. For example, since “time [sec] per 1 [sample] = 1 / sampling frequency”, when the sampling frequency is 96 [kHz], the time per 1 [sample] is about 10.42 [msec]. It becomes. The time corresponding to 1 [sample] multiplied by τ [sample] is the arrival time difference between the microphones, and it is possible to introduce a phase shift (phase difference).

When the least square method is used, the following equation can be used.

  When the least square method is used, the reference signal extraction unit 52 extracts and fixes a signal waveform included in the collected sound signal (reference signal) 21. The comparison signal extraction unit 54 extracts a signal waveform included in the collected sound signal (comparison signal) 22 and moves the signal waveform. The cross-correlation value calculation unit 55 calculates the sum of squares of the difference values between the signal waveform included in the collected sound signal 21 and the signal waveform included in the collected sound signal 22. The point at which the sum of squares is minimum is a place where the signal waveform included in the collected sound signal 21 and the signal waveform included in the collected sound signal 22 are similar (overlapping) to each other, and is determined to have the highest correlation. be able to. When the least square method is used, it is desirable to make the sizes of the reference signal and the comparison signal uniform, and it is preferable to normalize in advance based on one of them.

  The cross-correlation value calculation unit 55 outputs information related to the correlation between the reference signal and the comparison signal obtained by the above calculation to the phase difference acquisition unit 56. That is, the two signal waveforms determined to have high correlation by the cross-correlation value calculation unit 55 (that is, the signal waveform included in the sound collection signal 21 and the signal waveform included in the sound collection signal 22) have the same sound source. It is highly possible that the signal waveform is voice or noise. Therefore, the phase difference acquisition unit 56 obtains the phase difference between the two signal waveforms determined to have a high correlation, thereby obtaining the sound collected by the sound microphone 11 and the sound collected by the reference sound microphone 12. Can be obtained.

  The phase information acquisition unit 16 updates the phase difference (sound phase difference) between the sound pickup signal 21 and the sound pickup signal 22 when the sound noise interval detection unit 15 detects a sound interval. Moreover, the phase information acquisition part 16 updates the phase difference (noise phase difference) of the sound collection signal 21 and the sound collection signal 22 when the audio | voice noise area detection part 15 has detected the noise area.

  For example, when the audio noise interval information 23 supplied from the audio noise interval detection unit 15 indicates an audio interval, the probability that the phase difference acquired by the phase difference acquisition unit 56 is an audio phase difference (audio phase difference). Can be said to be expensive. At this time, since the selector 59 is supplied with a signal indicating the voice section as the voice noise section information 23, the selector 59 uses the phase difference (voice phase difference) output from the phase difference acquisition unit 56 as the voice phase difference storage unit 57. Output to. The audio phase difference storage unit 57 updates the already stored audio phase difference to the latest audio phase difference supplied from the selector 59. The updated audio phase difference is held until the next time when the audio noise interval information 23 indicates the audio interval (that is, the next update timing of the audio phase difference).

  Further, when the audio noise interval information 23 supplied from the audio noise interval detector 15 indicates a noise interval, the probability that the phase difference acquired by the phase difference acquisition unit 56 is a noise phase difference (noise phase difference). Can be said to be expensive. At this time, the selector 59 is supplied with a signal indicating the noise interval as the audio noise interval information 23, so the selector 59 converts the phase difference (noise phase difference) output from the phase difference acquisition unit 56 into the noise phase difference storage unit 58. Output to. The noise phase difference storage unit 58 updates the already stored noise phase difference to the latest noise phase difference supplied from the selector 59. The updated noise phase difference is held until the next time when the audio noise interval information 23 indicates the noise interval (that is, the next update timing of the noise phase difference).

  The audio phase difference stored in the audio phase difference storage unit 57 and the noise phase difference stored in the noise phase difference storage unit 58 are supplied as phase information 25 to the noise reduction processing determination unit 17. At this time, the audio phase difference and the noise phase difference are separately recognized by the noise reduction processing determination unit 17.

  5-7 is a figure which shows an example of the position of the sound source of an audio | voice and noise with respect to the microphone 11 for audio | voices with which the noise reduction apparatus concerning this Embodiment and the microphone 12 for reference sounds are equipped. 5-7, the sound microphone 11 is provided on the front side of the wireless communication apparatus 600, and the reference sound microphone 12 is provided on the back side. Usually, the speaker speaks toward the voice microphone 11 provided on the front side of the wireless communication apparatus 600.

  As shown in FIG. 5, when the sound source (speaker) is on the sound microphone 11 side, the phase of the sound collected by the sound microphone 11 is the same as that of the sound collected by the reference sound microphone 12. Faster than phase. Therefore, in this case, the phase difference (sound phase difference) between the sound collection signal 21 and the sound collection signal 22 is positive.

  On the other hand, when the noise source is on the reference sound microphone 12 side, the phase of the noise collected by the sound microphone 11 is slower than the phase of the noise collected by the reference sound microphone 12. Therefore, in this case, the phase difference (noise phase difference) between the sound collection signal 21 and the sound collection signal 22 is negative.

  Further, as shown in FIG. 6, when both the sound source (speaker) and the noise sound source are on the sound microphone 11 side, the phase of the sound collected by the sound microphone 11 is the reference sound microphone. 12 is earlier than the phase of the sound collected. The phase of noise collected by the sound microphone 11 is earlier than the phase of noise collected by the reference sound microphone 12. Therefore, in this case, both the phase difference (sound phase difference) between the sound collection signal 21 and the sound collection signal 22 in the sound section and the phase difference (noise phase difference) between the sound collection signal 21 and the sound collection signal 22 in the noise section are positive. It becomes.

  Further, as shown in FIG. 7, when both the sound source (speaker) and the noise sound source are on the reference sound microphone 12 side, the phase of the sound collected by the sound microphone 11 is It is later than the phase of the sound collected by the microphone 12. The phase of noise collected by the sound microphone 11 is slower than the phase of noise collected by the reference sound microphone 12. Therefore, in this case, the phase difference (sound phase difference) between the sound collection signal 21 and the sound collection signal 22 in the voice section and the phase difference (noise phase difference) between the sound collection signal 21 and the sound collection signal 22 in the noise section are both negative. It becomes.

  The noise reduction processing determination unit 17 illustrated in FIG. 1 determines the state of the audio phase difference and noise phase difference acquired by the phase information acquisition unit 16. For example, whether the state of the audio phase difference and the noise phase difference is the first state where the noise reduction process is performed, whether the noise reduction process is not performed, or the noise reduction process is performed from the first state. It is also determined whether or not the second state, which is a weak implementation.

  For example, when the absolute value of the difference between the audio phase difference acquired by the phase information acquisition unit 16 and the noise phase difference is within a predetermined threshold (first threshold), the noise reduction processing determination unit 17 The state can be determined. Hereinafter, the second state, which is a case where the noise reduction process is not performed or a case where the noise reduction process is performed weaker than the first state, may be simply described as “a case where the noise reduction process is not performed”.

  It should be noted that by using the audio phase difference and the noise phase difference, it is further determined whether the noise reduction processing is not performed or whether the noise reduction processing is performed weaker than in the first state. Also good. If there is a higher possibility that the voice component itself is reduced by providing a new threshold, the noise reduction process may not be performed. Further, the strength of noise reduction processing may be changed by always adapting to the difference between the audio phase difference and the noise phase difference. In this case, the determination operation performed by the noise reduction processing determination unit 17 is an operation of calculating the absolute value of the difference between the audio phase difference and the noise phase difference, and the noise reduction processing unit has a strength corresponding to the absolute value of the difference. Noise reduction processing is performed. For example, the noise reduction processing may be weaker as the absolute value of the difference is smaller. The above is the same when the power difference is used as in the second embodiment if the phase difference and the power difference are replaced.

  Here, the predetermined threshold can be arbitrarily set. For example, as the predetermined threshold value is decreased, the criterion for performing the noise reduction process becomes loose (in other words, the range in which the noise reduction process is not performed is narrowed). In other words, the difference between the audio phase difference and the noise phase difference is, for example, the angle of entry of the sound into the sound microphone 11 (the sound entry angle of the sound with respect to the main surface of the sound microphone 11) and the approach of the noise into the sound microphone 11. This corresponds to the difference between the angle (the noise entry angle with respect to the main surface of the voice microphone 11). Therefore, the smaller the predetermined threshold is, the narrower the difference between the sound and noise approach angles determined not to perform the noise reduction process.

  Conversely, the larger the predetermined threshold value, the stricter the criteria for performing the noise reduction process (in other words, the wider the range in which it is determined not to perform the noise reduction process). That is, the larger the predetermined threshold value, the wider the difference between the sound and noise approach angles determined not to perform the noise reduction process.

  As the difference between the sound and noise approach angles approaches 0, the sound (sound and noise) collected by the sound microphone 11 and the reference sound microphone 12 is approximated. For this reason, when noise reduction processing is performed in the noise reduction processing unit 18, there is a problem that the noise component included in the collected sound signal 21 is reduced and the sound component is also reduced at the same time. In order to solve such a problem, in the noise reduction device according to the present exemplary embodiment, the difference between the audio phase difference acquired by the phase information acquisition unit 16 and the noise phase difference (corresponding to the difference between the audio and noise approach angles). To determine whether or not to perform noise reduction processing. That is, when the absolute value of the difference between the audio phase difference and the noise phase difference is within a predetermined threshold, it can be determined that the noise reduction process is not performed.

  For example, if it is determined that the noise reduction process is to be performed (first state), the noise reduction process determination unit 17 invalidates the determination flag 26 (low level) and does not perform the noise reduction process or performs the noise reduction process first. If it is determined that the operation is performed weaker than the state (second state), the determination flag 26 is set valid (high level).

  The noise reduction processing unit 18 performs noise reduction processing using the sound collection signal 21 and the sound collection signal 22 according to the determination result of the noise reduction processing determination unit 17. That is, the noise reduction processing unit 18 is included in the sound collection signal 21 when the noise reduction processing determination is performed (first state) in the noise reduction processing determination unit 17 (when the determination flag 26 is low level). The noise component is reduced using the collected sound signal 22, and the signal after the noise reduction processing is output as the output signal 27. Further, when the noise reduction process determination unit 17 determines that the noise reduction process is not performed or the noise reduction process is performed weaker than the first state (second state) (when the determination flag 26 is at a high level). The collected sound signal 21 may be output as it is as an audio signal, or the noise reduction process may be performed so that the effect of the noise reduction process becomes weaker than usual (that is, as shown in FIG. 8). The pseudo noise signal 83 may be set smaller).

  The noise reduction processing unit 18 collects a reference sound including a noise component by using the reference sound microphone 12 in order to reduce a noise component included in the collected sound signal (audio signal) 21, and uses the reference sound as a reference sound. Based on this, a noise component that may be included in the collected sound signal 21 is generated in a pseudo manner. And the noise reduction process part 18 can implement a noise reduction process by subtracting this pseudo noise component generated from the collected sound signal 21.

  For example, an output signal 27 (digital signal) output from the noise reduction processing unit 18 is converted into an analog signal by a DA converter (not shown), and the converted analog signal is output to a speaker or a voice at the output unit (not shown). Output from the terminal.

  FIG. 8 is a block diagram illustrating an example of the noise reduction processing unit 18. The noise reduction processing unit 18 includes delay elements 71_1 to 71_n, multipliers 72_1 to 72_n + 1, adders 73_1 to 73_n, an adaptive coefficient adjustment unit 74, a subtractor 75, and an output signal selection unit 76. The delay elements 71_1 to 71_n, the multipliers 72_1 to 72_n + 1, and the adders 73_1 to 73_n constitute an FIR filter. The pseudo noise signal 83 is generated by processing the sound collection signal 22 using the delay elements 71_1 to 71_n, the multipliers 72_1 to 72_n + 1, and the adders 73_1 to 73_n.

  The adaptive coefficient adjustment unit 74 adjusts the coefficients of the multipliers 72_1 to 72_n + 1 in accordance with the audio noise section information 24. That is, the adaptive coefficient adjustment unit 74 adjusts the coefficient so that the adaptive error is reduced when the audio noise interval information 24 indicates a noise interval. On the other hand, when the voice noise section information 24 indicates a voice section, the coefficient is maintained or only the coefficient is finely adjusted.

  The subtractor 75 generates a signal 84 after noise reduction processing by subtracting the pseudo noise signal 83 from the collected sound signal 21 and outputs the signal 84 to the output signal selection unit 76. Further, the subtractor 75 generates a feedback signal 85 by subtracting the pseudo noise signal 83 from the collected sound signal 21, and outputs the feedback signal 85 to the adaptive coefficient adjustment unit 74.

  The output signal selection unit 76 outputs the collected sound signal 21 as it is as the output signal 27 according to the determination flag 26 output from the noise reduction process determination unit 17 or outputs the signal 84 after the noise reduction process as the output signal 27. Select whether to output as. That is, the output signal selection unit 76 outputs the collected sound signal 21 as it is as the output signal 27 when the determination flag 26 output from the noise reduction processing determination unit 17 is valid (high level). On the other hand, when the determination flag 26 output from the noise reduction processing determination unit 17 is invalid (low level), the signal 84 after the noise reduction processing is output as the output signal 27.

  Next, operation | movement of the noise reduction apparatus 1 concerning this Embodiment is demonstrated. FIG. 9 is a flowchart for explaining the operation of the noise reduction apparatus 1 according to the present embodiment.

  First, the voice noise section detector 15 detects whether the sound collected by the voice microphone 11 (sound pickup signal 21) is a voice section or a noise section (step S1). At this time, it is possible to reliably detect the voice section and the noise section by tightening the conditions for detecting the voice section and the noise section.

  When the audio noise interval information 23 detected by the audio noise interval detector 15 indicates a noise interval (step S2: No), the phase information acquisition unit 16 uses the collected sound signal 21 and the collected sound signal 22 in the noise interval. The noise phase difference is acquired (step S3). And the phase information acquisition part 16 updates the noise phase difference already hold | maintained using the noise phase difference acquired by step S3 (step S4).

  On the other hand, when the speech noise section information 23 detected by the speech noise section detection unit 15 indicates a speech section (step S2: Yes), the phase information acquisition unit 16 performs the sound collection signal 21 and the sound collection signal 22 in the speech section. Is used to obtain the audio phase difference (step S5). And the phase information acquisition part 16 updates the audio | voice phase difference already hold | maintained using the audio | voice phase difference acquired by step S5 (step S6).

  Next, the noise reduction process determination unit 17 determines whether or not to perform the noise reduction process based on the audio phase difference and the noise phase difference acquired by the phase information acquisition unit 16. And the noise reduction process determination part 17 determines with implementing a noise reduction process, when the absolute value of the difference of an audio | voice phase difference and a noise phase difference is larger than a predetermined threshold value (step S7: No). At this time, since the determination flag 26 output from the noise reduction processing determination unit 17 is invalid (low level), the noise reduction processing unit 18 uses the sound collection signal 22 to reduce the noise component included in the sound collection signal 21. Then, the signal after the noise reduction processing is output as the output signal 27 (step S8).

  On the other hand, when the absolute value of the difference between the audio phase difference and the noise phase difference is within a predetermined threshold (step S7: Yes), the noise reduction process determination unit 17 determines that the noise reduction process is not performed. At this time, since the determination flag 26 output from the noise reduction processing determination unit 17 is valid (high level), the noise reduction processing unit 18 outputs the sound collection signal 21 (audio signal) as it is (step S9).

  Next, a voice input device using the noise reduction device according to the present embodiment will be described. FIG. 10 is a diagram illustrating an example of a voice input device 500 using the noise reduction device according to the present embodiment. FIG. 10A is a front view of the voice input device 500, and FIG. 10B is a rear view of the voice input device 500. As shown in FIG. 10, the voice input device 500 is configured to be connectable to the wireless communication device 510 via a connector 503. The wireless communication device 510 can use a general wireless device, and is configured to be able to communicate with other wireless communication devices at a predetermined frequency. The voice of the speaker is input to the wireless communication device 510 via the voice input device 500.

  The voice input device 500 includes a main body 501, a code 502, and a connector 503. The main body 501 is configured to have a size and shape suitable for being held by a speaker's hand, and includes a microphone, a speaker, an electronic circuit, and a noise reduction device. As shown in FIG. 10A, a speaker 506 and an audio microphone 505 are provided on the front surface of the main body 501. As shown in FIG. 10B, a reference sound microphone 508 and a belt clip 507 are provided on the back surface of the main body 501. An LED 509 is provided on the top surface of the main body 501. A PTT (Push To Talk) 504 is provided on a side surface of the main body 501. The LED 509 notifies the speaker of the detection state of the speaker's voice by the voice input device 500. The PTT 504 is a switch for setting the wireless communication device 510 in a voice transmission state, and detects that the protruding portion is pushed into the housing.

  The noise reduction device 1 according to the present embodiment is built in the voice input device 500. The voice microphone 11 included in the noise reduction device 1 corresponds to the voice microphone 505 of the voice input device 500, and the noise reduction device 1 is provided. The provided reference sound microphone 12 corresponds to the reference sound microphone 508 of the voice input device 500. Further, the output signal 27 output from the noise reduction device 1 is supplied to the wireless communication device 510 via the code 502 of the voice input device 500. That is, the voice input device 500 supplies the wireless communication device 510 with the output signal 27 that has been subjected to noise reduction processing by the noise reduction device 1. Therefore, the sound transmitted from the wireless communication apparatus 510 to another wireless communication apparatus is a sound subjected to noise reduction processing. Note that the noise reduction device 1 may be configured to be built in the wireless communication device 510.

  Next, a radio communication apparatus (transceiver) 600 using the noise reduction apparatus according to this embodiment will be described. FIG. 11 is a diagram illustrating an example of a wireless communication device 600 using the noise reduction device according to the present embodiment. FIG. 11A is a front view of the wireless communication apparatus 600, and FIG. 11B is a rear view of the wireless communication apparatus 600. As shown in FIG. 11, the wireless communication apparatus 600 includes an input button 601, a display unit 602, a speaker 603, an audio microphone 604, a PTT (Push To Talk) 605, a switch 606, an antenna 607, a reference sound microphone 608, and A lid 609 is provided.

  The noise reduction device 1 according to the present embodiment is built in the wireless communication device 600. The voice microphone 11 included in the noise reduction device 1 corresponds to the voice microphone 604 of the wireless communication device 600. The reference sound microphone 12 provided corresponds to the reference sound microphone 608 of the wireless communication apparatus 600. Further, the output signal 27 output from the noise reduction device 1 is subjected to high frequency processing in an internal circuit of the wireless communication device 600 and is wirelessly transmitted from the antenna 607 to another wireless communication device. Here, since the output signal 27 output from the noise reduction apparatus 1 is a signal on which noise reduction processing has been performed, the voice transmitted to another wireless communication apparatus is the voice on which noise reduction processing has been performed. When sound transmission is started by the user pressing the PTT 605, the processing of the noise reduction apparatus 1 as shown in FIG. 9 is started, and the user stops pressing the PTT 608 and the sound transmission is ended. In addition, the processing of the noise reduction apparatus 1 as shown in FIG. 9 may be terminated.

  As described in the problem of the present invention, when the noise reduction processing is performed using the voice microphone 11 that mainly collects sound and the reference sound microphone 12 that mainly collects noise, depending on the arrival direction of the noise. However, there is a problem that the amount of voice reduction (cancellation amount) increases. That is, depending on the use situation of the noise reduction device, the sound may be mixed into the reference sound microphone 12 that collects noise. When sound is mixed into the reference sound microphone 12 as described above, not only the noise component mixed in the sound collected by the sound microphone but also the sound itself is canceled, and the clarity of the sound is lowered. was there.

  For example, as shown in FIG. 7, when both the sound source (speaker) and the noise sound source are on the reference sound microphone 12 side, the sound is also collected by the reference sound microphone 12. Here, the noise reduction device collects the reference sound collected by the reference sound microphone 12 and artificially generates a noise component that may be included in the collected sound signal 21 based on the reference sound. Then, noise reduction processing is performed by subtracting the pseudo-generated noise component from the collected sound signal 21. For this reason, when sound is mixed into the reference sound microphone 12, when the noise component mixed in the sound collected by the sound microphone 11 is reduced, the sound itself is canceled together with the noise component. there were.

  For example, as shown in FIG. 6, when both the sound source (speaker) and the noise sound source are on the sound microphone 11 side, the action of canceling the sound coming from the direction of the noise sound source works. As for the sound component having the same direction of arrival, the sound component mixed in the reference sound microphone 12 is reduced, for example, the sound collected by the sound microphone 11 is reduced, and the clarity of the sound is impaired. End up. In addition to the case where the sound source of noise and the sound source of noise are in the same direction (see FIGS. 6 and 7), mirror symmetry is performed with respect to an axis connecting the sound microphone 11 and the reference sound microphone 12 with a straight line. When sound or noise comes from such a direction (that is, when the incident angle of sound and noise to each microphone approximates), there is a problem that the sound component is reduced along with the reduction of the noise component. It was. In such an environment, the noise itself is canceled together with the noise component during the noise reduction process, and there is a problem that the noise reduction process cannot be performed appropriately.

  Moreover, in patent document 1, in order to distinguish an audio | voice signal and a noise signal, the audio | voice and noise are picked up using a directional microphone. At this time, the directional microphones are arranged in opposite directions. However, for example, when a speaker speaks 90 degrees laterally with respect to a directional microphone that collects sound (that is, an intermediate position between two directional microphones), the directional microphone appropriately collects sound. Can not do it. Further, when speaking from the lateral direction of the two directional microphones, the sound component is equally input to the sound microphone and the noise microphone. In this case, since the sound is canceled together with the noise, the quality of the output sound is deteriorated.

  Moreover, in patent document 2, in order to distinguish an audio | voice signal and a noise signal, it uses combining 2 types of directional microphones. However, the directional microphone has a fixed input gain. Therefore, when the speaker speaks at a position off the front of the voice microphone, the voice signal may not be collected because the range of directivity is exceeded. In addition, when sound is input to the noise microphone, the sound component is also reduced together with the noise component.

  As described above, there is a problem that the noise reduction device cannot appropriately perform the noise reduction processing depending on the voice and noise arrival directions.

  In order to solve such a problem, in the noise reduction apparatus according to the present embodiment, the sound phase difference that is the phase difference between the sound collection signal 21 and the sound collection signal 22 in the sound section and the sound collection signal in the noise section Whether or not noise reduction processing is to be performed is determined based on a noise phase difference that is a phase difference between 21 and the sound pickup signal 22. That is, when the absolute value of the difference between the audio phase difference and the noise phase difference is within a predetermined threshold, it can be determined that the noise reduction process is not performed.

  As described above, in the noise reduction device according to the present embodiment, it is possible to determine a case that is inappropriate for performing the noise reduction processing based on the audio phase difference and the noise phase difference. Therefore, it is possible to prevent the noise reduction process from being performed when it is inappropriate to perform the noise reduction process. Here, the case where the noise reduction processing is inappropriate is, for example, when the sound is mixed in the reference sound microphone 12 or on the axis where the sound microphone 11 and the reference sound microphone 12 are connected by a straight line. When voice or noise comes from a direction that is mirror-symmetric with respect to (that is, when the incident angle of the voice and noise to each microphone approximates).

  12 and 13 are diagrams for explaining the effect of the invention according to the present embodiment. The horizontal axis in FIGS. 12 and 13 is the sample time [sec], and the vertical axis is the sound pressure level [dB]. FIG. 12 shows a state in which sound is mixed into the reference sound microphone 12 and not only the noise component mixed in the sound collected by the sound microphone 11 but also the sound itself is canceled. FIG. 13 shows a case where the noise reduction apparatus according to the present embodiment is used. That is, FIG. 13 shows a state where it is determined that the noise reduction process is inappropriate and the noise reduction process is not performed. Note that the positions of the sections A and B shown in FIG. 12 correspond to the positions of the sections A and B shown in FIG.

  As shown in FIG. 12, when the noise reduction apparatus according to the present embodiment is not used, audio components are reduced in the sections A and B. On the other hand, when the noise reduction apparatus according to the present embodiment is used, the speech component is not reduced in the sections A and B as shown in FIG. Therefore, by using the noise reduction device according to the present embodiment, it is possible to appropriately perform noise reduction processing according to the voice arrival direction and the noise arrival direction.

  The invention according to the present embodiment described above can provide a noise reduction device, a voice input device, a wireless communication device, and a noise reduction method that can appropriately reduce noise components.

<Embodiment 2>
Next, a second embodiment of the present invention will be described. FIG. 14 is a block diagram of the noise reduction apparatus according to the second embodiment. The noise reduction apparatus 2 according to the present embodiment is different from the noise reduction apparatus 1 described in the first embodiment in that a power information acquisition unit 60 is provided. Other than this, since it is the same as the noise reduction apparatus 1 described in the first embodiment, the same components are denoted by the same reference numerals, and redundant description is omitted.

  When the voice noise section information 23 indicates a voice section, the power information acquisition unit 60 acquires a voice power difference that is a difference between the magnitude of the sound collection signal 21 and the sound collection signal 22 in the voice section. Further, when the audio noise section information 23 indicates a noise section, the power information acquisition unit 60 acquires a noise power difference that is a difference between the magnitude of the sound collection signal 21 and the sound collection signal 22 in the noise section. The acquired audio power difference and noise power difference are supplied as power information 28 to the noise reduction processing determination unit 17.

  FIG. 15 is a block diagram illustrating an example of the power information acquisition unit 60 included in the noise reduction device 2 according to the present embodiment. The power information acquisition unit 60 shown in FIG. 15 includes a sound collection signal buffer 61, a sound collection signal power calculation unit 62, a sound collection signal buffer 63, a sound collection signal power calculation unit 64, a power difference calculation unit 65, and a sound power difference storage unit. 67, a noise power difference storage unit 68, and a selector 69. The power information acquisition unit 60 shown in FIG. 15 can obtain power information (power difference in the case of FIG. 15) of the sound pickup signal 21 and the sound pickup signal 22 in a certain unit time.

  The sound collection signal buffer 61 temporarily accumulates the supplied sound collection signal 21 in order to accumulate the sound collection signal 21 for a unit time. The sound collection signal buffer 63 temporarily accumulates the supplied sound collection signal 22 in order to accumulate the sound collection signal 22 for a unit time.

  The collected sound signal power calculation unit 62 calculates a power value per unit time using the collected sound signals for unit time accumulated in the collected sound signal buffer 61. In addition, the sound collection signal power calculation unit 64 calculates a power value per unit time using the sound collection signals for the unit time accumulated in the sound collection signal buffer 63.

  Here, the power value per unit time is the magnitude of the sound pickup signals 21 and 22 in unit time. For example, the maximum value of the amplitude of the sound pickup signals 21 and 22 in unit time or the sound pickup in unit time. An integrated value of the amplitude of the signals 21 and 22 can be used. In the present embodiment, any value other than the above maximum value or integral value may be used as the power value as long as the value indicates the magnitude of the sound pickup signals 21 and 22.

  The power difference calculation unit 65 calculates the power difference between the power value of the sound collection signal 21 obtained by the sound collection signal power calculation unit 62 and the power value of the sound collection signal 22 obtained by the sound collection signal power calculation unit 64. .

  When the audio noise section detection unit 15 detects a voice section, the power information acquisition unit 60 determines the power difference between the sound collection signal 21 and the sound collection signal 22, that is, the magnitude of the sound collection signal 21 and the sound collection signal 22. The difference from the size (audio power difference) is updated. In addition, when the audio noise section detection unit 15 detects a noise section, the power information acquisition unit 60 determines the power difference between the collected sound signal 21 and the collected sound signal 22, that is, the magnitude of the collected sound signal 21 and the collected sound. The difference (noise power difference) from the magnitude of the signal 22 is updated.

  For example, when the audio noise interval information 23 supplied from the audio noise interval detector 15 indicates an audio interval, the probability that the power difference calculated by the power difference calculator 65 is an audio power difference (audio power difference). Can be said to be expensive. At this time, since the selector 69 is supplied with a signal indicating the voice section as the voice noise section information 23, the selector 69 uses the power difference (voice power difference) output from the power difference calculator 65 as the voice power difference storage 67. Output to. The audio power difference storage unit 67 updates the already stored audio power difference to the latest audio power difference supplied from the selector 69. The updated voice power difference is held until the next time when the voice noise section information 23 indicates the voice section (that is, the next update timing of the voice power difference).

  Further, when the audio noise interval information 23 supplied from the audio noise interval detector 15 indicates a noise interval, the probability that the power difference calculated by the power difference calculator 65 is a noise power difference (noise power difference). Can be said to be expensive. At this time, since the selector 69 is supplied with a signal indicating the noise interval as the audio noise interval information 23, the selector 69 uses the power difference (noise power difference) output from the power difference calculator 65 as the noise power difference storage 68. Output to. The noise power difference storage unit 68 updates the already stored noise power difference to the latest noise power difference supplied from the selector 69. The updated noise power difference is held until the next time when the audio noise section information 23 indicates the noise section (that is, the next update timing of the noise power difference).

  The audio power difference stored in the audio power difference storage unit 67 and the noise power difference stored in the noise power difference storage unit 68 are supplied to the noise reduction processing determination unit 17 as power information 28. At this time, the audio power difference and the noise power difference are separately recognized by the noise reduction processing determination unit 17.

  For example, as shown in FIG. 5, when the sound source (speaker) of the sound is on the sound microphone 11 side, the size of the sound collected by the sound microphone 11 is picked up by the reference sound microphone 12. It is larger than the loudness of the voice. Therefore, in this case, the power difference (sound power difference) between the sound collection signal 21 and the sound collection signal 22 is positive.

  On the other hand, when the noise source is on the reference sound microphone 12 side, the magnitude of the noise collected by the voice microphone 11 is smaller than the magnitude of the noise collected by the reference sound microphone 12. Therefore, in this case, the power difference (noise power difference) between the sound collection signal 21 and the sound collection signal 22 is negative.

  Also, as shown in FIG. 6, when both the sound source (speaker) and the noise sound source are on the sound microphone 11 side, the size of the sound collected by the sound microphone 11 is The volume of sound collected by the microphone 12 is larger. Further, the magnitude of noise collected by the audio microphone 11 is larger than the magnitude of noise collected by the reference sound microphone 12. Therefore, in this case, the power difference (sound power difference) between the sound collection signal 21 and the sound collection signal 22 in the sound section and the power difference (noise power difference) between the sound collection signal 21 and the sound collection signal 22 in the noise section are both positive. It becomes.

  Further, as shown in FIG. 7, when both the sound source (speaker) and the noise sound source are on the reference sound microphone 12 side, the size of the sound collected by the sound microphone 11 is the reference sound. The volume of sound collected by the microphone 12 is smaller. Further, the magnitude of noise collected by the audio microphone 11 is smaller than the magnitude of noise collected by the reference sound microphone 12. Therefore, in this case, the power difference (sound power difference) between the sound collection signal 21 and the sound collection signal 22 in the sound section and the power difference (noise power difference) between the sound collection signal 21 and the sound collection signal 22 in the noise section are both negative. It becomes.

  The noise reduction process determination unit 17 performs the first state in which the noise reduction process is performed based on the audio power difference and the noise power difference acquired by the power information acquisition unit 60, and the case where the noise reduction process is not performed. Or it determines with the 2nd state which is a case where a noise reduction process is implemented weaker than a 1st state. For example, when the absolute value of the difference between the audio power difference acquired by the power information acquisition unit 60 and the noise power difference is within a predetermined threshold (second threshold), the noise reduction processing determination unit 17 performs the second operation. It can be determined that Here, the sound power difference and the noise power difference are obtained by calculating a relative ratio between the microphones (for example, the power of the sound pickup signal 21 / the power of the sound pickup signal 22) and comparing the two, thereby obtaining a sound and noise difference. It can be determined whether or not the approach angle to the microphone is approximate.

  Here, the predetermined threshold can be arbitrarily set. For example, as the predetermined threshold value is decreased, the criterion for performing the noise reduction process becomes loose (in other words, the range in which the noise reduction process is not performed is narrowed). That is, the difference between the sound power difference and the noise power difference is, for example, the approach angle of speech to the speech microphone 11 (speech entrance angle with respect to the main surface of the speech microphone 11) and the approach of noise to the speech microphone 11. This corresponds to the difference between the angle (the noise entry angle with respect to the main surface of the voice microphone 11). Therefore, the smaller the predetermined threshold is, the narrower the difference between the sound and noise approach angles determined not to perform the noise reduction process.

  Conversely, the larger the predetermined threshold value, the stricter the criteria for performing the noise reduction process (in other words, the wider the range in which it is determined not to perform the noise reduction process). That is, the larger the predetermined threshold value, the wider the difference between the sound and noise approach angles determined not to perform the noise reduction process.

  As the difference between the sound and noise approach angles approaches 0, the sound (sound and noise) collected by the sound microphone 11 and the reference sound microphone 12 is approximated. For this reason, when noise reduction processing is performed in the noise reduction processing unit 18, there is a problem that the noise component included in the collected sound signal 21 is reduced and the sound component is also reduced at the same time. In order to solve such a problem, in the noise reduction apparatus according to the present embodiment, the difference between the sound power difference acquired by the power information acquisition unit 60 and the noise power difference (the difference between the sound and noise approach angles). It is determined whether or not to perform noise reduction processing based on (corresponding). That is, when the absolute value of the difference between the audio power difference and the noise power difference is within a predetermined threshold, it can be determined that the noise reduction process is not performed.

  For example, if it is determined that the noise reduction process is to be performed (first state), the noise reduction process determination unit 17 invalidates the determination flag 26 (low level) and does not perform the noise reduction process or performs the noise reduction process first. If it is determined that the operation is performed weaker than the state (second state), the determination flag 26 is set valid (high level).

  The noise reduction processing unit 18 performs noise reduction processing using the sound collection signal 21 and the sound collection signal 22 according to the determination result of the noise reduction processing determination unit 17. That is, the noise reduction processing unit 18 is included in the sound collection signal 21 when the noise reduction processing determination is performed (first state) in the noise reduction processing determination unit 17 (when the determination flag 26 is low level). The noise component is reduced using the collected sound signal 22, and the signal after the noise reduction processing is output as the output signal 27. Further, when the noise reduction process determination unit 17 determines that the noise reduction process is not performed or the noise reduction process is performed weaker than the first state (second state) (when the determination flag 26 is at a high level). The collected sound signal 21 may be output as it is as an audio signal, or the noise reduction process may be performed so that the effect of the noise reduction process becomes weaker than usual (that is, as shown in FIG. 8). The pseudo noise signal 83 may be set smaller).

  Next, operation | movement of the noise reduction apparatus 2 concerning this Embodiment is demonstrated. FIG. 16 is a flowchart for explaining the operation of the noise reduction apparatus 2 according to the present embodiment.

  First, the sound noise section detection unit 15 detects whether the sound (sound collected signal 21) collected by the sound microphone 11 is a sound section or a noise section (step S11). At this time, it is possible to reliably detect the voice section and the noise section by tightening the conditions for detecting the voice section and the noise section.

  The power information acquisition unit 60 uses the sound collection signal 21 and the sound collection signal 22 in the noise interval when the audio noise interval information 23 detected by the audio noise interval detection unit 15 indicates a noise interval (step S12: No). The noise power difference is acquired (step S13). And the power information acquisition part 60 updates the noise power difference already hold | maintained using the noise power difference acquired by step S13 (step S14).

  On the other hand, when the sound noise section information 23 detected by the sound noise section detection unit 15 indicates a sound section (step S12: Yes), the power information acquisition unit 60 collects the sound collection signal 21 and the sound collection signal 22 in the sound section. Is used to obtain the audio power difference (step S15). And the power information acquisition part 60 updates the audio | voice power difference already hold | maintained using the audio | voice power difference acquired by step S15 (step S16).

  Next, the noise reduction process determination unit 17 determines whether or not to perform the noise reduction process based on the audio power difference and the noise power difference acquired by the power information acquisition unit 60. And the noise reduction process determination part 17 determines with implementing a noise reduction process, when the absolute value of the difference of an audio | voice power difference and a noise power difference is larger than a predetermined threshold value (step S17: No). At this time, since the determination flag 26 output from the noise reduction processing determination unit 17 is invalid (low level), the noise reduction processing unit 18 uses the sound collection signal 22 to reduce the noise component included in the sound collection signal 21. Then, the signal after the noise reduction processing is output as the output signal 27 (step S18).

  On the other hand, when the absolute value of the difference between the audio power difference and the noise power difference is within a predetermined threshold (step S17: Yes), the noise reduction process determination unit 17 determines that the noise reduction process is not performed. At this time, since the determination flag 26 output from the noise reduction processing determination unit 17 is valid (high level), the noise reduction processing unit 18 outputs the collected sound signal 21 (audio signal) as it is (step S19).

  In the noise reduction apparatus according to the present embodiment, the sound power difference that is the difference between the magnitude of the collected sound signal 21 and the collected sound signal 22 in the sound section, and the magnitude of the collected sound signal 21 and the collected sound signal in the noise section. Whether or not to perform noise reduction processing is determined based on a noise power difference that is a difference from the size of 22. That is, when the absolute value of the difference between the audio power difference and the noise power difference is within a predetermined threshold, it can be determined that the noise reduction process is not performed.

  As described above, in the noise reduction device according to the present embodiment, it is possible to determine a case inappropriate for performing the noise reduction processing based on the audio power difference and the noise power difference. Therefore, it is possible to prevent the noise reduction process from being performed when it is inappropriate to perform the noise reduction process.

  In the noise reduction apparatus according to the present embodiment, the noise reduction processing is performed using the phase information acquired by the phase information acquisition unit 16 (see Embodiment 1) together with the power information acquired by the power information acquisition unit 60. You may determine whether to implement. In this case, for example, like the noise reduction device 2 ′ illustrated in FIG. 17, a phase power information acquisition unit 70 including a phase information acquisition unit and a power information acquisition unit can be provided.

  For example, the phase power information acquisition unit 70 acquires the audio phase difference and the noise phase difference in the phase information acquisition unit, acquires the audio power difference and the noise power difference in the power information acquisition unit, and uses these information as the phase power information 29. Is output to the noise reduction processing determination unit 17.

  The noise reduction processing determination unit 17 determines whether to perform the noise reduction processing based on the difference between the audio phase difference and the noise phase difference acquired by the phase power information acquisition unit 70 and the difference between the audio power difference and the noise power difference. Can be determined. For example, the absolute value of the difference between the audio phase difference and the noise phase difference is within a predetermined first threshold, and the absolute value of the difference between the audio power difference and the noise power difference is within a predetermined second threshold. In this case, it can be determined that the noise reduction process is not performed. At this time, by adjusting the first threshold and the second threshold, the determination using the difference between the audio phase difference and the noise phase difference and the determination using the difference between the audio power difference and the noise power difference are made. Weighting can be given.

  For example, in a portable device (wireless communication device) such as a transceiver or a small device such as a speaker microphone (voice input device) attached to the wireless communication device, the microphone opening may be blocked by a hand, For example, the microphone opening may be shielded. Therefore, when determining the presence or absence of noise reduction processing, a combination of a method using a phase difference and a method using a power difference can be used to increase the accuracy of cases that are inappropriate for noise reduction processing. Can be determined.

<Embodiment 3>
Next, a third embodiment of the present invention will be described. FIG. 18 is a block diagram showing the noise reduction device 3 according to the present embodiment. In the noise reduction device 3 according to the present embodiment, the configurations and operations of the audio noise section detection unit 95, the noise reduction processing determination unit 97, and the noise reduction processing unit 98 are the same as those of the noise reduction device 1 ( Different from FIG. Other than this, since it is the same as the noise reduction apparatus 1 described in the first embodiment, the same components are denoted by the same reference numerals, and redundant description is omitted.

  As shown in FIG. 18, the noise reduction device 3 according to the present embodiment includes an audio microphone 11, a reference sound microphone 12, AD converters 13 and 14, an audio noise section detection unit 95, a phase information acquisition unit 16, noise A reduction processing determination unit 97 and a noise reduction processing unit 98 are included.

  The voice noise section detector 95 detects a voice section and a noise section based on the collected sound signal 21 output from the AD converter 13 or the collected sound signal 22 output from the AD converter 14. Then, the voice noise section detector 15 outputs the voice noise section information 23 and 24 indicating the voice section and the noise section to the phase information acquisition section 16 and the noise reduction processing section 98, respectively.

  For example, the audio noise section detection unit 95 may include a circuit that determines whether or not the sound collection signal 21 includes sound, and a circuit that determines whether or not the sound collection signal 22 includes sound. In this case, the voice noise section detection unit 95 can detect a voice section using the collected sound signal that contains more voice. Note that the same technique as that described in the first embodiment can be used for detection of the voice section and the noise section in the voice noise section detection unit 15.

  In the noise reduction apparatus 1 described in the first embodiment, it is assumed that there is a high probability that sound is collected by the sound microphone 11, and the sound noise section detection unit 15 performs only the sound collection signal 21 of the sound microphone 11. The case where the speech section is determined based on the above is shown. However, depending on how the noise reduction device is used, there may be cases where the reference sound microphone 12 collects more sound than the sound microphone 11. Therefore, in the present embodiment, the voice noise section detection unit 95 can detect a voice section using the collected sound signal of the collected sound signal 21 and the collected sound signal 22 that contains more sound. It is configured as follows.

  The noise reduction process determination unit 97 is a first state where noise reduction processing is performed based on the audio phase difference and noise phase difference acquired by the phase information acquisition unit 16, and when noise reduction processing is not performed. Or it determines with the 2nd state which is a case where a noise reduction process is implemented weaker than a 1st state. For example, when the absolute value of the difference between the audio phase difference acquired by the phase information acquisition unit 16 and the noise phase difference is within a predetermined threshold (first threshold), the noise reduction processing determination unit 17 performs the second operation. It can be determined that

  As shown in FIG. 21, depending on the use situation of the wireless communication device 600 including the noise reduction device, it may be assumed that a sound source is present on the side where the reference sound microphone 12 is disposed. In this case, since the sound collection signal 22 from the reference sound microphone 12 includes many sound components, it is preferable to reduce the noise component contained in the sound collection signal 22 using the sound collection signal 21. Therefore, the noise reduction process can be performed more reliably.

  Therefore, when the sound source of the sound is on the reference sound microphone 12 side, the noise reduction processing determination unit 97 is a selection for switching between the sound collection signal 21 and the sound collection signal 22 used for the noise reduction processing in the noise reduction processing unit 98. A signal 99 (for example, a high level signal) is output. In this manner, by outputting the selection signal 99 to the noise reduction processing unit 98, the noise component included in the sound collection signal 22 can be reduced using the sound collection signal 21.

  For example, when the phase of the sound collected by the sound microphone 11 is slower than the phase of the sound collected by the reference sound microphone 12, it is determined that the sound source of the sound is on the reference sound microphone 12 side. be able to. In other words, when the phase of the sound pickup signal 22 in the sound section acquired by the phase information acquisition unit 16 is earlier than the phase of the sound pickup signal 21 (that is, the phase difference between the sound pickup signal 21 and the sound pickup signal 22 (sound level). When the phase difference is negative), the noise reduction processing determination unit 97 can determine that the sound source is the reference sound microphone 12 side.

  For example, as shown in FIG. 21, when the noise source is on the sound microphone 11 side, the phase of the noise collected by the sound microphone 11 is the noise collected by the reference sound microphone 12. Be faster than phase. Therefore, in this case, the phase difference (noise phase difference) between the sound collection signal 21 and the sound collection signal 22 is positive.

  Since the other configuration and operation of the noise reduction process determination unit 97 are the same as the configuration and operation of the noise reduction process determination unit 17 described in the first embodiment, a duplicate description is omitted.

  Moreover, in the noise reduction apparatus according to the present embodiment, as in the noise reduction apparatus described in the second embodiment, noise reduction processing is performed using the audio power difference and the noise power difference acquired by the power information acquisition unit. You may determine whether to implement. For example, when the volume of the sound collected by the sound microphone 11 is smaller than the volume of the sound collected by the reference sound microphone 12, it is determined that the sound source is on the reference sound microphone 12 side. be able to. In other words, when the magnitude of the sound pickup signal 22 in the voice section acquired by the power information acquisition unit is larger than the magnitude of the sound pickup signal 21 (that is, the power difference between the sound pickup signal 21 and the sound pickup signal 22 (voice power When the difference is negative), the noise reduction processing determination unit 97 can determine that the sound source is the reference sound microphone 12 side.

  Further, for example, when the noise source is the voice microphone 11 side, the magnitude of the noise collected by the voice microphone 11 is larger than the magnitude of the noise collected by the reference sound microphone 12. Therefore, in this case, the power difference (noise power difference) between the sound collection signal 21 and the sound collection signal 22 is positive.

  The noise reduction processing unit 98 performs noise reduction processing using the sound collection signal 21 and the sound collection signal 22 according to the determination result of the noise reduction processing determination unit 97. For example, when the noise reduction processing unit 98 performs the noise reduction processing in which the phase of the sound collection signal 21 in the voice section is earlier than the phase of the sound collection signal 22 (that is, the sound source of the sound is the sound microphone 11 side). When the noise reduction processing determination unit 97 determines (when the determination flag 26 is at a low level), the noise component included in the sound collection signal 21 is reduced using the sound collection signal 22, and the signal after the noise reduction process is output as an output signal. 27 is output. In addition, the noise reduction processing unit 98 has a phase of the collected sound signal 21 earlier than that of the collected sound signal 22 in the sound section (that is, the sound source of the sound is on the sound microphone 11 side) and does not perform the noise reduction process. And the noise reduction processing determination unit 97 (when the determination flag 26 is at a high level), the collected sound signal 21 is output as the output signal 27 as it is.

  On the other hand, the noise reduction processing unit 98 performs a noise reduction process in which the phase of the sound pickup signal 22 in the voice section is earlier than the phase of the sound pickup signal 21 (that is, the sound source of the voice is the reference sound microphone 12 side). Then, when it is determined by the noise reduction processing determination unit 97 (when the determination flag 26 is at low level), the noise component included in the sound collection signal 22 is reduced using the sound collection signal 21 and a signal after the noise reduction process is output. Output as signal 27. In addition, the noise reduction processing unit 98 performs the noise reduction process in which the phase of the sound collection signal 22 in the speech section is earlier than the phase of the sound collection signal 21 (that is, the sound source of the sound is the reference sound microphone 12 side). Otherwise, if the noise reduction processing determination unit 97 determines (when the determination flag 26 is at a high level), the sound collection signal 22 is output as it is as the output signal 27.

  When power information is used, the noise reduction processing unit 98, for example, has a larger sound collection signal 21 in a sound section than a sound collection signal 22 (that is, a sound source is a sound microphone 11). Side), and when the noise reduction processing determination unit 97 determines that the noise reduction processing is performed (when the determination flag 26 is low level), the noise component included in the sound pickup signal 21 is reduced using the sound pickup signal 22. Then, the signal after the noise reduction processing is output as the output signal 27. In addition, the noise reduction processing unit 98 performs the noise reduction processing in which the size of the collected sound signal 21 in the speech section is larger than the magnitude of the collected sound signal 22 (that is, the sound source of the sound is the sound microphone 11 side). Otherwise, if the noise reduction processing determination unit 97 determines (when the determination flag 26 is at a high level), the sound collection signal 21 is output as it is as the output signal 27.

  On the other hand, the noise reduction processing unit 98 has a larger sound collection signal 22 in the voice section than the sound collection signal 21 (that is, the sound source of the sound is the reference sound microphone 12 side), and performs noise reduction processing. When the noise reduction processing determination unit 97 determines that the noise reduction processing is performed (when the determination flag 26 is at a low level), the noise component included in the sound pickup signal 22 is reduced using the sound pickup signal 21, and the signal after the noise reduction process is obtained. Output as an output signal 27. In addition, the noise reduction processing unit 98 has a larger sound collection signal 22 in the voice section than the sound collection signal 21 (that is, the sound source of the sound is the reference sound microphone 12 side), and performs noise reduction processing. Otherwise, if the noise reduction processing determination unit 97 determines (when the determination flag 26 is at high level), the sound collection signal 22 is output as it is as the output signal 27.

  FIG. 19 is a block diagram illustrating an example of the noise reduction processing unit 98. The noise reduction processing unit 98 includes delay elements 71_1 to 71_n, multipliers 72_1 to 72_n + 1, adders 73_1 to 73_n, an adaptive coefficient adjustment unit 74, a subtractor 75, an output signal selection unit 76, and a selector 77.

  The selector 77 selects the sound collection signal 21 and the sound collection signal 22 as a sound signal 81 (a signal mainly including a sound component) and a reference signal 82 (noise), respectively, according to the selection signal 99 output from the noise reduction processing determination unit 97. A signal is output as a component for generating a pseudo component, and a case where the sound pickup signal 21 and the sound pickup signal 22 are output as a reference signal 82 and a sound signal 81, respectively. For example, the selector 77 outputs the sound collection signal 21 and the sound collection signal 22 as the sound signal 81 and the reference signal 82, respectively, when the sound source is the sound microphone 11 side (that is, when the selection signal 99 is low level). To do. On the other hand, when the sound source is the reference sound microphone 12 side (that is, when the selection signal 99 is at a high level), the sound collection signal 21 and the sound collection signal 22 are output as the reference signal 82 and the sound signal 81, respectively.

  The delay elements 71_1 to 71_n, the multipliers 72_1 to 72_n + 1, and the adders 73_1 to 73_n constitute an FIR filter. The pseudo noise signal 83 is generated by processing the reference signal 82 using the delay elements 71_1 to 71_n, the multipliers 72_1 to 72_n + 1, and the adders 73_1 to 73_n.

  The adaptive coefficient adjustment unit 74 adjusts the coefficients of the multipliers 72_1 to 72_n + 1 in accordance with the audio noise section information 24. That is, the adaptive coefficient adjustment unit 74 adjusts the coefficient so that the adaptive error is reduced when the audio noise interval information 24 indicates a noise interval. On the other hand, when the speech noise section information 24 indicates a speech section, the coefficient of the noise reduction processing unit 18 is maintained or only the coefficient is finely adjusted.

  The subtractor 75 generates a signal 84 after noise reduction processing by subtracting the pseudo noise signal 83 from the audio signal 81 and outputs the signal 84 to the output signal selection unit 76. Also, the subtractor 75 generates a feedback signal 85 by subtracting the pseudo noise signal 83 from the audio signal 81 and outputs the feedback signal 85 to the adaptive coefficient adjustment unit 74.

  The output signal selection unit 76 outputs the audio signal 81 as it is as the output signal 27 according to the determination flag 26 output from the noise reduction processing determination unit 97, or the signal 84 after the noise reduction processing as the output signal 27. Select whether to output. In other words, the output signal selection unit 76 outputs the audio signal 81 as it is as the output signal 27 when the determination flag 26 output from the noise reduction processing determination unit 97 is valid (high level). On the other hand, when the determination flag 26 output from the noise reduction processing determination unit 97 is invalid (low level), the signal 84 after the noise reduction processing is output as the output signal 27.

  Next, operation | movement of the noise reduction apparatus 3 concerning this Embodiment is demonstrated. FIG. 20 is a flowchart for explaining the operation of the noise reduction apparatus 3 according to the present embodiment. Note that steps S21 to S26 shown in FIG. 20 are the same as steps S1 to S6 (see Embodiment 1) shown in FIG.

  In step S <b> 27, the noise reduction processing determination unit 97 determines whether the sound source is on the reference sound microphone 12 side. When the sound source of the sound is the sound microphone 11 side (step S27: No), the noise reduction processing unit 98 collects the sound signal 21 of the sound microphone 11 as the sound signal 81 and the reference sound microphone 12. The sound signal 22 is used as a reference signal 82 (step 28). For example, when the phase difference (sound phase difference) between the sound pickup signal 21 and the sound pickup signal 22 is positive, the noise reduction processing determination unit 97 can determine that the sound source of the sound is on the sound microphone 11 side. .

  On the other hand, when the sound source of the sound is the reference sound microphone 12 side (step S <b> 27: Yes), the noise reduction processing unit 98 uses the sound signal 22 of the reference sound microphone 12 as the sound signal 81 and the sound microphone 11. The collected sound signal 21 is used as a reference signal 82 (step 29). For example, when the phase difference (sound phase difference) between the sound collection signal 21 and the sound collection signal 22 is negative, the noise reduction processing determination unit 97 may determine that the sound source of the sound is on the reference sound microphone 12 side. it can.

  Next, the noise reduction process determination unit 97 determines whether or not to perform the noise reduction process based on the audio phase difference and the noise phase difference acquired by the phase information acquisition unit 16. That is, when the absolute value of the difference between the audio phase difference and the noise phase difference is larger than the predetermined first threshold (step S30: No), the noise reduction process determination unit 97 determines to perform the noise reduction process. When power information is used, the noise reduction processing determination unit 97, when the absolute value of the difference between the audio power difference acquired by the power information acquisition unit and the noise power difference is larger than a predetermined second threshold value, It can be determined that the noise reduction process is performed.

  At this time, since the determination flag 26 output from the noise reduction processing determination unit 97 is invalid (low level), the noise reduction processing unit 98 uses the reference signal 82 for the noise component included in the audio signal 81 (see FIG. 19). The signal after the noise reduction processing is output as the output signal 27 (step S31).

  On the other hand, when the absolute value of the difference between the audio phase difference and the noise phase difference is within a predetermined first threshold (step S30: Yes), the noise reduction process determination unit 97 determines that the noise reduction process is not performed. . When power information is used, the noise reduction processing determination unit 97, when the absolute value of the difference between the audio power difference acquired by the power information acquisition unit and the noise power difference is within a predetermined second threshold, It can be determined that the noise reduction process is not performed.

  At this time, since the determination flag 26 output from the noise reduction processing determination unit 97 is valid (high level), the noise reduction processing unit 98 outputs the audio signal 81 as it is (step S32).

  As shown in FIG. 21, depending on the use situation of the wireless communication device 600 including the noise reduction device, it may be assumed that a sound source is present on the side where the reference sound microphone 12 is disposed. In this case, since the sound collection signal 22 from the reference sound microphone 12 includes many sound components, it is preferable to reduce the noise component contained in the sound collection signal 22 using the sound collection signal 21. Therefore, the noise reduction process can be performed more reliably.

  Therefore, in the noise reduction apparatus 3 according to the present embodiment, when the sound source of the sound is the reference sound microphone 12 side, the sound collection signal 21 and the sound collection signal 22 used for the noise reduction processing in the noise reduction processing unit 98 are obtained. Switching. In this way, by switching between the sound collection signal 21 and the sound collection signal 22, the noise component included in the sound collection signal 22 can be reduced using the sound collection signal 21, and noise reduction processing can be performed more reliably. Can be implemented.

  Although the present invention has been described with reference to the above embodiment, the present invention is not limited to the configuration of the above embodiment, and can be made by those skilled in the art within the scope of the invention of the claims of the claims of the present application. It goes without saying that various modifications, corrections, and combinations are included. For example, the sound microphone 11 and the reference sound microphone 12 may be provided at substantially the same position in the upper part (or lower part) of the device, and these microphones may be arranged so as to have different directivities. For example, it is preferable to arrange the sound microphone 11 and the reference sound microphone 12 so that the directivities thereof are different by 180 °.

1, 2, 3 Noise reduction device 11 Audio microphone 12 Reference sound microphone 13, 14 AD converter 15, 95 Audio noise section detection unit 16 Phase information acquisition unit 17, 97 Noise reduction processing determination unit 18, 98 Noise reduction processing unit 21, 22 Sound pickup signals 23, 24 Audio noise section information 25 Audio noise direction information 26 Determination flag 27 Output signal 28 Power information 29 Phase power information 60 Power information acquisition unit 70 Phase power information acquisition unit

Claims (14)

A voice section and noise based on at least one of a first sound collection signal corresponding to the sound collected by the first microphone and a second sound collection signal corresponding to the sound collected by the second microphone. A voice noise section detector for detecting a section;
A voice power difference that is a difference between the magnitude of the first sound pickup signal and the second sound pickup signal in the voice section, the magnitude of the first sound pickup signal in the noise section, and the second A power information acquisition unit that acquires a noise power difference that is a difference from the magnitude of the sound pickup signal of
A noise reduction processing determination unit that determines the state of the audio power difference and the noise power difference depending on whether a value based on the difference between the audio power difference and the noise power difference is within a predetermined threshold ;
A noise reduction processing unit that performs noise reduction processing using the first sound collection signal and the second sound collection signal according to a determination result of the noise reduction process determination unit;
A noise reduction device comprising: The noise reduction processing determination unit determines whether or not the state of the audio power difference and the noise power difference is a first state where noise reduction processing is performed,
When the noise reduction processing determination unit determines that the noise reduction processing determination unit is in the first state, the noise reduction processing unit performs noise reduction processing using the first sound collection signal and the second sound collection signal. The noise reduction device according to claim 1 . The noise reduction process determination unit is a second state where the state of the audio power difference and the noise power difference is a case where the noise reduction process is performed weaker than the first state where the noise reduction process is performed Whether or not
The noise reduction device according to claim 1 , wherein the noise reduction processing unit performs the noise reduction processing weaker than the first state when the noise reduction processing determination unit determines that the second state is present. . The noise reduction processing determination unit determines whether or not the state of the audio power difference and the noise power difference is a second state in which noise reduction processing is not performed.
The noise reduction processing unit, when the noise reduction processing determination unit determines that it is the second state, does not perform the noise reduction processing, the noise reducing device according to claim 1. The noise according to claim 3 or 4 , wherein the noise reduction processing determination unit determines that the second state is present when an absolute value of the difference between the audio power difference and the noise power difference is within a predetermined threshold. Reduction device. The noise reduction device according to any one of claims 1 to 5 , wherein the power information acquisition unit updates the audio power difference in the audio interval and updates the noise power difference in the noise interval. The noise reduction processing determination unit determines whether or not the state of the audio power difference and the noise power difference is a first state where noise reduction processing is performed,
When the noise reduction processing determination unit determines that the noise reduction processing determination unit is in the first state, the noise reduction processing unit performs noise reduction processing using the first sound collection signal and the second sound collection signal. And
The noise reduction processing unit
When the noise reduction process determination unit determines that the first state is present, the noise component included in the first sound collection signal is reduced using the second sound collection signal, and after the noise reduction process The signal is output as an audio signal,
When the noise reduction processing determination unit determines the second state, the first sound collection signal is output as an audio signal.
The noise reduction device according to any one of claims 3 to 5 . The noise reduction processing determination unit determines whether or not the state of the audio power difference and the noise power difference is a first state where noise reduction processing is performed,
When the noise reduction processing determination unit determines that the noise reduction processing determination unit is in the first state, the noise reduction processing unit performs noise reduction processing using the first sound collection signal and the second sound collection signal. And
The noise reduction processing unit
The magnitude of the second collected signal in the voice section is much larger than the size of the first voice collecting signal, and, if it is determined that the first state in the noise reduction processing determination unit, wherein Reducing a noise component included in the second collected sound signal by using the first collected sound signal, and outputting the signal after the noise reduction processing as an audio signal;
The magnitude of the second collected signal in the voice section is much larger than the size of the first voice collecting signal, and, if it is determined that the second state in the noise reduction processing determination unit, wherein Outputting the second collected sound signal as an audio signal;
The noise reduction device according to any one of claims 3 to 5 . Voice input device having a noise reduction device according to any one of claims 1 to 8. The first microphone is provided on a first surface of the voice input device;
The voice input device according to claim 9 , wherein the second microphone is provided on a second surface facing the first surface with a predetermined distance. Wireless communication device having the noise reducing device according to any one of claims 1 to 8. The first microphone is provided on a first surface of the wireless communication device;
The wireless communication device according to claim 11 , wherein the second microphone is provided on a second surface facing the first surface with a predetermined distance. A voice section and noise based on at least one of a first sound collection signal corresponding to the sound collected by the first microphone and a second sound collection signal corresponding to the sound collected by the second microphone. Detect the interval,
A voice power difference that is a difference between the magnitude of the first sound pickup signal and the second sound pickup signal in the voice section, the magnitude of the first sound pickup signal in the noise section, and the second To obtain the noise power difference that is the difference between
Whether or not a value based on the difference between the audio power difference and the noise power difference is within a predetermined threshold determines the state of the audio power difference and the noise power difference,
A noise reduction process is performed using the first sound collection signal and the second sound collection signal according to a result of determining the state of the sound power difference and the noise power difference.
Noise reduction method. A voice section based on at least one of a first sound collection signal corresponding to sound collected by the first microphone and a second sound collection signal corresponding to sound collected by the second microphone in the computer And detect the noise interval,
A voice power difference that is a difference between the magnitude of the first sound pickup signal and the second sound pickup signal in the voice section, the magnitude of the first sound pickup signal in the noise section, and the second To obtain the noise power difference that is the difference with the magnitude of the sound pickup signal of
Value based on the difference between the noise power difference between the sound power difference, depending on whether it is within a predetermined threshold, to determine the state of the noise power difference between the sound power difference,
In accordance with the result of determining the state of the audio power difference and the noise power difference, noise reduction processing is performed using the first sound collection signal and the second sound collection signal.
Noise reduction program.