JP5958218B2 - Noise reduction device, voice input device, wireless communication device, and noise reduction method - Google Patents

Description

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

  There is a noise cancellation function (noise reduction device) that reduces the noise component contained in the audio signal and makes it easier to hear the audio.

  In the noise cancellation function, for example, the noise included in the audio signal is subtracted from the audio signal collected by the microphone that mainly collects sound from the audio signal collected by the microphone that mainly collects noise. Components can be removed.

  Patent Document 1 discloses a technology relating to a noise reduction device that can prevent a reduction in sound component itself even when a sound component that is not a reduction target is mixed in the noise component. Patent Document 2 discloses a technique related to a microphone system capable of obtaining a large S / N ratio improvement effect regardless of the position of a noise source. Patent Document 3 discloses a technique related to a sound processing apparatus that can suitably reduce noise of a sound signal input from a microphone in a plurality of environments.

JP-A-6-67693 JP 2000-305594 A JP 2010-2322862 A

  The techniques disclosed in Patent Documents 1 to 3 have a problem that noise components included in an audio signal cannot be appropriately reduced when the ambient noise level is high. In addition, a wireless communication apparatus for transmitting and receiving voice is required to ensure call quality in various environments such as a noisy environment.

  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, and a noise reduction method that can appropriately reduce noise components contained in a voice signal even under various environments. That is.

  The noise reduction device according to the present invention includes a voice section determination unit that determines a voice section based on voice collected by at least one of the first and second microphones, and a sound collected by the first microphone. A voice direction detector that detects a direction of arrival of the voice based on a first collected sound signal corresponding to the second collected sound signal corresponding to a sound collected by the second microphone, and the voice section Based on the voice section information output from the determiner and the voice direction information output from the voice direction detector, noise reduction processing is performed using the first sound pickup signal and the second sound pickup signal. An audio filter, and the voice direction detector detects the direction of arrival of the voice when the voice segment determiner determines that the voice segment is a voice segment.

  The voice direction detector may detect the direction of arrival of the voice based on a phase difference between the first sound pickup signal and the second sound pickup signal.

  The adaptive filter uses the other collected sound signal to reduce a noise component included in one of the first collected sound signal and the second collected sound signal that has an earlier phase. May be.

  When the phase difference between the phase of the first sound pickup signal and the phase of the second sound pickup signal is within a predetermined range, the adaptive filter does not perform noise reduction processing and the first sound pickup signal Alternatively, the second sound collection signal may be output.

  The voice direction detector may detect the direction of arrival of the voice based on the magnitude of the first collected sound signal and the magnitude of the second collected sound signal.

  When the magnitude of the first collected sound signal is larger than the magnitude of the second collected sound signal, the adaptive filter is more preferably selected from the first collected sound signal and the second collected sound signal. You may reduce the noise component contained in any one of the large sound collection signals using the other sound collection signal.

  If the power difference, which is the difference between the magnitude of the first collected signal and the magnitude of the second collected signal, is within a predetermined range, the adaptive filter does not perform the noise reduction process. Or the second sound collection signal may be output.

  The sound direction detector is based on a phase difference between the first sound collection signal and the second sound collection signal, and a magnitude of the first sound collection signal and a magnitude of the second sound collection signal. The direction of arrival of the voice may be detected.

  When the phase of the first sound collection signal is earlier than the phase of the second sound collection signal, the speech segment determination unit may determine a speech segment based on the first sound collection signal, When the phase of the second sound collection signal is earlier than the phase of the first sound collection signal, the speech segment determination unit may determine a speech segment based on the second sound collection signal.

  The voice direction detector may be supplied with a signal having a sampling frequency of 24 kHz or more as the first and second sound pickup signals, and the adaptive filter may be supplied with the first and second sound pickup signals. A signal having a sampling frequency of 12 kHz or less may be supplied.

  The speech segment determination device may output speech segment determination information determined to be a speech segment with a higher probability than the speech segment determination information output to the adaptive filter to the speech direction detector.

  In the voice input device provided with the noise reduction device according to the present invention, the first microphone may be provided on a first surface of the voice input device, and the second microphone may be provided on the first surface. And a second surface facing each other with a predetermined distance.

  In the wireless communication device including the noise reduction device according to the present invention, the first microphone may be provided on a first surface of the wireless communication device, and the second microphone is connected to the first surface. And a second surface facing each other with a predetermined distance.

  The noise reduction method according to the present invention determines a voice section based on voice collected by at least one of the first and second microphones, and determines that the voice section is a voice section, the first microphone uses the first microphone. Detecting the direction of arrival of the voice based on a first collected signal corresponding to the collected sound and a second collected signal corresponding to the sound collected by the second microphone; The noise reduction processing is performed based on the speech section information that is the determination result of the above and the speech direction information indicating the arrival direction of the speech.

  According to the present invention, it is possible to provide a noise reduction device, a voice input device, a wireless communication device, and a noise reduction method that can appropriately reduce noise components included in a voice signal even under various environments.

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 area determination device 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 area determination device with which the noise reduction apparatus concerning Embodiment 1 is provided. It is a block diagram which shows an example of the audio | voice direction detector 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 direction detector with which the noise reduction apparatus concerning Embodiment 1 is provided. It is a block diagram which shows an example of the adaptive filter with which the noise reduction apparatus concerning Embodiment 1 is provided. 3 is a flowchart for explaining an operation of the noise reduction apparatus according to the first exemplary embodiment; It is a block diagram which shows the other example of the noise reduction apparatus concerning Embodiment 1. FIG. 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 block diagram which shows the noise reduction apparatus concerning Embodiment 2. It is a block diagram which shows an example of the signal determination part with which the noise reduction apparatus concerning Embodiment 2 is provided. 10 is a flowchart for explaining an operation of a signal determination unit provided in the noise reduction device according to the second exemplary embodiment; 10 is a flowchart for explaining an operation of a signal determination unit provided in the noise reduction device according to the second exemplary embodiment; It is a block diagram which shows an example of the adaptive filter 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; FIG. 6 is a block diagram illustrating a noise reduction device according to a third exemplary embodiment. 10 is a flowchart for explaining the operation of the noise reduction apparatus according to the third exemplary embodiment; FIG. 6 is a diagram illustrating a voice input device according to a fourth embodiment. FIG. 10 is a diagram for explaining a position of a reference sound microphone provided on the back surface of the voice input device according to the fourth exemplary embodiment; FIG. 6 illustrates a wireless communication apparatus according to a fourth embodiment.

<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. The noise reduction apparatus 1 shown in FIG. 1 includes an audio microphone 11, a reference sound microphone 12, AD converters 13 and 14, an audio section determination unit 15, an audio direction detector 16, an adaptive filter control unit 17, and an adaptive filter 18. Have.

  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 speech section determiner 15, the speech direction detector 16, and the adaptive filter 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 sound direction detector 16 and the adaptive filter 18.

In the present embodiment, an example of the frequency band of 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 segment determination unit 15 determines a voice segment based on the sound collection signal 21 output from the AD converter 13. Then, when the speech segment determination unit 15 determines that the speech segment is a speech segment, it outputs the speech segment information 23 and 24 to the speech direction detector 16 and the adaptive filter control unit 17, respectively.

  An arbitrary technique can be used for the speech segment determination processing in the speech segment determiner 15. However, when the noise reduction device is used in an environment where the noise level is high, it is necessary to determine the speech section with high accuracy. In this case, for example, the technique described in Japanese Patent Application No. 2010-260798 (hereinafter referred to as voice section determination technique A) or the technique described in Japanese Patent Application No. 2011-020659 (hereinafter referred to as voice section determination technique B) is used. Thus, it is possible to determine the voice section with high accuracy. The sound includes sounds other than human voices, but in this example, human voices are mainly detected.

  In the speech section determination 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 section determination technique A, an appropriate noise level is set for each band, a signal-to-noise level ratio with a peak of the 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.

  FIG. 2 is a block diagram illustrating an example of a speech segment determination unit 15 ′ using the speech segment determination 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 determining unit. 38.

  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). Cosine
It is recommended to use an orthogonal transformation method such as 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 first threshold value, and determines whether or not the first threshold value is exceeded. If there is a spectrum whose SNR exceeds the first threshold value, 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 segment information 23 and 24 to the speech direction detector 16 and the adaptive filter control unit 17, respectively.

  The speech section determination 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, when the spectrum whose SNR exceeds the first threshold is greater than or equal to the predetermined number, 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. it can.

  As described above, the speech segment determination unit 15 ′ using the speech segment determination technique A cuts out an input signal in units of frames having a predetermined time width, and generates a framed input signal. A spectrum generation unit 32 for converting the framing input signal from the time domain to the frequency domain to generate a spectrum pattern in which spectra for each frequency are collected, each spectrum of the spectrum pattern, and a predetermined bandwidth A peak detector 37 that determines whether or not the energy ratio of the divided frequency bands including the spectrum among the plurality of divided frequency bands that are the divided frequency bands to the energy by band exceeds a predetermined first threshold value. And a voice determination unit 38 that determines whether or not the framed input signal is voice based on the determination result of the peak detection unit, A frequency averaging unit 34 for deriving an average energy in the frequency direction of the spectrum in each divided frequency band of the spectrum pattern, and a time averaging unit 36 for deriving the energy by band that is an average of the average energy in the time direction for each divided frequency band. And comprising.

  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 first threshold is equal to or greater than a predetermined number.

  Next, the speech section determination technique B will be described. In the speech section determination 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 speech segment determination 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 set in a band where there is not much influence of noise. By specializing and extracting, it is possible to determine the speech section with high accuracy.

  FIG. 3 is a block diagram illustrating an example of a speech segment determination unit 15 ″ using the speech segment determination technique B. The speech section determiner 15 ″ includes a framing unit 41, a spectrum generating unit 42, a band dividing unit 43, an average deriving unit 44, a noise level deriving unit 45, a determination selecting unit 46, and a consonant determining 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 segment determination unit 15 ″ using the speech segment determination technique B derives the average energy for each band in the spectrum pattern and compares the energy for each band to the right in the spectrum pattern characteristic of the consonant. Detect upward trend. Therefore, the speech segment determination unit 15 ″ can accurately detect a consonant segment 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 in the high frequency band than in the adjacent low frequency band, and the counted value is a predetermined first threshold value. If it is above, the 1st judgment means which judges 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 second threshold value. 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 segment determination unit 15 ″ using the speech segment determination technique B includes the framing unit 41 that cuts out an input signal in units of a predetermined frame and generates a framed input signal, The spectrum generation unit 42 that converts the input signal from the time domain to the frequency domain and generates a spectrum pattern in which the spectrum for each frequency is collected, and the average energy for each predetermined bandwidth to be connected in the spectrum pattern The average deriving unit 44 for deriving the average energy for each band and the derived average energy for each band are compared. If the average energy for each band in the higher frequency band is higher, the framed input signal A consonant determination unit 47 that determines that a consonant is included.

  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 area determination technique A and B to the noise reduction apparatus concerning this Embodiment, a parameter can be set for every product. That is, when the speech segment determination techniques A and B are applied to a product that requires more reliable speech segment determination, a stricter threshold can be set as a parameter for speech segment determination.

  Further, in the noise reduction apparatus 1 shown in FIG. 1, it is assumed that the voice has a high probability of being picked up by the voice microphone 11, and the voice section determination unit 15 is based only on the sound pickup signal 21 of the voice microphone 11. The case where the speech section is determined 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, as in the noise reduction device 2 shown in FIG. 8, the speech section determination unit 19 determines the speech section based on the sound collection signal 21 of the sound microphone 11 and the sound collection signal 22 of the reference sound microphone 12. It may be configured.

  In this case, for example, the voice section determination unit 19 of the noise reduction device 2 illustrated in FIG. 8 includes a circuit that determines whether or not the sound collection signal 21 of the sound microphone 11 includes sound, and the sound collection of the reference sound microphone 12. And a circuit for determining whether the signal 22 includes sound. The other configuration of the noise reduction device 2 shown in FIG. 8 is the same as the configuration of the noise reduction device 1 shown in FIG.

  The voice direction detector 16 of the noise reduction apparatus 1 shown in FIG. 1 detects the voice arrival direction based on the collected sound signal 21 and the collected sound signal 22, and outputs the voice direction information 25 to the adaptive filter control unit 17. . The method for detecting the voice arrival direction is, for example, a method for detecting the voice arrival direction based on the phase difference between the sound pickup signal 21 and the sound pickup signal 22, or a sound (sound pickup signal) collected by the voice microphone 11. 21) and the difference or ratio of the sound collected by the reference sound microphone 12 (sound collection signal 22) (power difference or power ratio, these are collectively referred to as power information). And a method of detecting the direction of arrival of voice. At this time, the voice direction detector 16 detects the arrival direction of the voice when the voice segment determination unit 15 determines that the voice segment is a voice segment. That is, the voice direction detector 16 detects the voice direction in the voice section where the voice has arrived, and does not detect the voice direction when the voice section is outside the voice section.

  In addition, 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) attached to the wireless communication device. 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 voice arrival direction based on the phase difference between the sound pickup signal 21 and the sound pickup signal 22, the sampling frequency of the sound pickup signal 21 and the sound pickup signal 22 input to the sound direction detector 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 voice direction detector 16 Between them, a sampling frequency converter may be provided, and the sampling frequency of the collected sound signals 21 and 22 supplied to the sound direction detector 16 may be converted to 24 kHz or more.

  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, the AD converters 13 and 14 and the AD converters 13 and 14 are adapted. A sampling frequency converter may be provided between the filter 18 and the sampling frequency of the collected sound signals 21 and 22 supplied to the speech section determination unit 15 and the adaptive filter 18 may be converted to 8 to 12 kHz.

  First, the case where the direction of arrival of voice is detected based on the phase difference between the collected sound signal 21 and the collected sound signal 22 will be described. FIG. 4 is a block diagram illustrating an example of a voice direction detector provided in the noise reduction device 1 according to the present embodiment. The voice direction detector 16 ′ shown 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, and a phase difference information acquisition unit 56. .

  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 that is emitted at the same time with one sound source, such as when the user is transmitting, has a different transmission path to each microphone 11, 12, and therefore the phase (delay amount) detected by each microphone 11, 12 And the amplitude value (attenuation amount) is different. However, it can be said that the sound emitted by the same sound source with one sound source has a very high correlation since the phase and amplitude value of the sound components detected by the microphones 11 and 12 have a certain relationship.

On the other hand, when the sound source exists in various places such as a noise component, the phase and amplitude value of the sound component detected by each of the microphones 11 and 12 has a different phase difference for each sound source, and the attenuation amount is also different. The nature is low. In the present embodiment, since the voice arrival direction is detected in the voice section, it can be said that the correlation between the voice components detected by the microphones 11 and 12 is very high. Therefore, the phase difference can be obtained by measuring this correlation only in the speech section, 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.

The cross-correlation function between the 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.

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 regarding the correlation between the reference signal and the comparison signal obtained by the above calculation to the phase difference information 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. There is a high possibility that the waveform is an audio signal. Therefore, the phase difference information acquisition unit 56 obtains the sound component collected by the sound microphone 11 and the reference sound microphone 12 by obtaining the phase difference between the two signal waveforms determined to have high correlation. The phase difference of the voice component can be obtained.

  Then, the phase of the sound component included in the sound collected by the sound microphone 11 (that is, the phase of the sound component of the sound collection signal 21) is included in the sound collected by the reference sound microphone 12. Is earlier than the phase (that is, the phase of the sound component of the collected sound signal 22) (that is, when the phase difference is positive), the sound source is located closer to the sound microphone 11 than the reference sound microphone 12, that is, It can be estimated that the speaker is speaking into the voice microphone 11.

  On the other hand, when the phase of the sound component included in the sound collected by the sound microphone 11 is slower than the phase of the sound component included in the sound collected by the reference sound microphone 12 (that is, the phase difference is negative). In this case, it can be estimated that the sound source is located closer to the reference sound microphone 12 than the sound microphone 11, that is, the speaker is speaking toward the reference sound microphone 12.

  Further, the phase difference between the phase of the sound component contained in the sound collected by the sound microphone 11 and the phase of the sound component contained in the sound collected by the reference sound microphone 12 is within a predetermined range. In this case (-T <phase difference <T, that is, when the absolute value of the phase difference is smaller than the predetermined value T), it is estimated that the sound source is located near the middle between the sound microphone 11 and the reference sound microphone 12. can do.

  The phase difference information acquisition unit 56 outputs the acquired phase difference information as the voice direction information 25 to the adaptive filter control unit 17.

  The voice direction detector 16 detects the voice arrival direction when the voice section determiner 15 determines that the voice section is the voice section. Therefore, even if noise is mixed, if the voice segment determination unit 15 determines that the voice segment is present, the voice component collected by the voice microphone 11 and the voice component collected by the reference sound microphone 12 Therefore, the direction of voice can be detected with high accuracy.

  Next, a description will be given of a case where the voice arrival direction is detected based on the power information of the collected sound signal 21 and the collected sound signal 22. FIG. 5 is a block diagram illustrating another example of the voice direction detector included in the noise reduction device 1 according to the present embodiment. The audio direction detector 16 '' shown in FIG. 5 includes an audio signal buffer 61, an audio signal power calculation unit 62, a reference signal buffer 63, a reference signal power calculation unit 64, a power difference calculation unit 65, and a power information acquisition unit 66. Prepare. The voice direction detector 16 '' shown in FIG. 5 can obtain power information (power difference in the case of FIG. 5) of the sound pickup signal 21 and the sound pickup signal 22 in a certain unit time.

  The audio 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 reference 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 audio signal power calculation unit 62 calculates a power value per unit time by using the collected sound signals for the unit time accumulated in the audio signal buffer 61. In addition, the reference signal power calculation unit 64 calculates a power value per unit time by using the collected sound signals for the unit time accumulated in the reference 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 obtained by the audio signal power calculation unit 62 and the power value of the sound collection signal obtained by the reference signal power calculation unit 64 to calculate the power difference. The power difference is output to the power information acquisition unit 66.

  The power information acquisition unit 66 acquires the power information of the sound collection signal 21 and the sound collection signal 22 based on the power difference output from the power difference calculation unit 65.

  For example, when the loudness of the sound collected by the sound microphone 11 is larger than the loudness of the sound collected by the reference sound microphone 12, that is, the power value of the collected sound signal 21 is that of the collected sound signal 22. If it is greater than the power value, it can be assumed that the sound source is located closer to the sound microphone 11 than the reference sound microphone 12, that is, that the speaker is speaking toward the sound microphone 11.

  On the other hand, when the loudness of the sound collected by the sound microphone 11 is smaller than the loudness of the sound collected by the reference sound microphone 12, that is, the power value of the sound collection signal 21 is equal to that of the sound collection signal 22. If it is smaller than the power value, it can be estimated that the sound source is located closer to the reference sound microphone 12 than the sound microphone 11, that is, the speaker is speaking toward the reference sound microphone 12.

  When the difference between the sound collected by the sound microphone 11 and the sound collected by the reference sound microphone 12 is within a predetermined range (−P <power difference <P, that is, the absolute power difference) When the value is smaller than the predetermined value P), it can be estimated that the sound source is located near the middle between the sound microphone 11 and the reference sound microphone 12.

  The power information acquisition unit 66 outputs the acquired power information (that is, information regarding the power difference) to the adaptive filter control unit 17 as the voice direction information 25.

  As described above, the voice direction detector 16 detects the voice arrival direction based on the phase difference between the collected sound signal 21 and the collected sound signal 22, and the power of the collected sound signal 21 and the collected sound signal 22. A method of detecting the direction of voice arrival based on information can be used. The method using the phase difference and the method using the power information may be used alone or in combination with each other. 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, if the method using the phase difference and the method using the power information are used in combination to accurately detect the voice direction, the voice direction can be detected with higher accuracy.

  The adaptive filter control unit 17 shown in FIG. 1 is a control for controlling the adaptive filter 18 based on the voice section information 24 output from the voice section determiner 15 and the voice direction information 25 output from the voice direction detector 16. A signal is generated, and the generated control signal 26 is output to the adaptive filter 18. Here, the control signal 26 includes voice section information 24 and voice direction information 25.

  The adaptive filter 18 uses the sound collection signal 21 and the sound collection signal 22 to generate a sound signal with reduced noise, and outputs the sound signal with reduced noise as an output signal 27. The adaptive filter 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 (sound signal) 21, and collects the reference sound based on the reference sound. A noise component that may be included in the sound signal 21 is generated in a pseudo manner. The adaptive filter 18 can perform noise reduction processing by subtracting the pseudo-generated noise component from the collected sound signal 21.

  Here, when a large amount of audio component is mixed in the reference sound microphone 12, the level of the audio signal may decrease, or it may become an echo component and the intelligibility of the audio signal may decrease. Therefore, for example, an allowable value in the case where an audio component is mixed in the reference sound microphone 12 is obtained in advance, and if the mixing of the audio component is within the allowable value range, noise reduction processing by the adaptive filter 18 is performed. If the mixing of audio components is outside the allowable range, the noise reduction processing in the adaptive filter 18 may be omitted and the sound collection signal (audio signal) 21 of the audio microphone 11 may be output as it is.

  In addition, when the mixing of audio components is outside the allowable range, it is also assumed that noise components are mainly collected in the audio microphone 11 and audio components are mainly collected in the reference sound microphone 12. In this case, the sound collection signal 21 and the sound collection signal 22 may be interchanged in the adaptive filter 18. That is, the noise reduction processing can be appropriately performed in the adaptive filter 18 by treating the sound collection signal 22 of the reference sound microphone 12 as a sound signal and treating the sound collection signal 21 of the sound microphone 11 as a reference signal. .

  The adaptive filter control unit 17 outputs a control signal 26 for performing the above processing to the adaptive filter 18. The speech section information 24 supplied to the adaptive filter control unit 17 is information for determining the update timing of the adaptive filter coefficient in the adaptive filter 18. For example, when the speech segment determination unit determines that it is not a speech segment (that is, a noise segment), the adaptive filter coefficient of the adaptive filter 18 may be updated in order to actively reduce the noise component. . On the other hand, for example, when it is determined that the speech section is determined to be a speech section, noise reduction processing may be performed using an existing adaptive filter coefficient.

  Further, for example, the phase of the sound component included in the sound collection signal 21 of the sound microphone 11 and the sound included in the sound collection signal 22 of the reference sound microphone 12 obtained by the sound direction detector 16 shown in FIG. The phase difference from the component phase is defined as a phase difference PD1. The predetermined value is T (positive value).

  If the relationship of phase difference PD1 ≧ T is established, the adaptive filter control unit 17 controls the adaptive filter 18 to perform normal noise reduction processing, for example. That is, in this case, since the phase of the sound pickup signal 21 of the sound microphone 11 is earlier than the phase of the sound pickup signal 22 of the reference sound microphone 12, the adaptive filter 18 is included in the sound pickup signal (voice signal) 21. An output signal 27 is generated by reducing the noise component using the collected sound signal (reference signal) 22. At this time, the voice section determination unit 15 can determine the voice section based on the sound collection signal 21 of the voice microphone 11.

If the relationship of phase difference PD1 ≦ −T is established, the adaptive filter control unit 17 may perform control so that the sound collection signal 21 and the sound collection signal 22 are switched in the adaptive filter 18, for example. That is, in this case, since the phase of the sound collection signal 22 of the reference microphone 12 is earlier than the phase of the sound collection signal 21 of the sound microphone 11, the adaptive filter control unit 17 performs the sound collection signal 22 of the reference sound microphone 12. Is treated as a sound signal, and the sound collection signal 21 of the sound microphone 11 is treated as a reference signal. Then, the adaptive filter control unit 17 generates an output signal 27 by controlling the adaptive filter 18 to reduce the noise component included in the collected sound signal (audio signal) 22 using the collected sound signal (reference signal) 21. can do. At this time, the speech section determination unit 15 may determine the speech section based on the sound collection signal 22 of the reference sound microphone 12 (in the case of the configuration in FIG. 8). When the phase of the sound pickup signal 22 of the reference microphone 12 is earlier than the phase of the sound pickup signal 21 of the sound microphone 11, the sound pickup signal 22 of the reference microphone 12 is more than the sound pickup signal 21 of the sound microphone 11. This is because it is more suitable for voice segment detection.

  The adaptive filter control unit 17 can determine that the collected sound signals 21 and 22 are unsuitable for noise reduction processing in the adaptive filter 18 when the relationship −T <phase difference PD1 <T holds. In this case, the adaptive filter 18 may output the collected sound signal 21 or the collected sound signal 22 as they are. That is, when the absolute value of the phase difference between the phase of the sound pickup signal 21 of the sound microphone 11 and the phase of the sound pickup signal 22 of the reference sound microphone 12 is smaller than a predetermined value (T), the adaptive filter 18 reduces noise. The sound collection signal 21 or the sound collection signal 22 may be output as the output signal 27 without processing.

  In this case, since the collected sound signals 21 and 22 having a small phase difference PD1 are unsuitable for noise reduction processing, the adaptive filter control unit 17 further selects the sound level in order to select a slightly better condition. For example, if the volume of the sound collected by the voice microphone 11 is larger than the volume of the sound collected by the reference microphone 12, the collected sound signal 21 is output as the output signal 27. You may control. The adaptive filter control unit 17 also outputs the collected sound signal 22 as an output signal 27 when the loudness of sound collected by the speech microphone 11 is smaller than the loudness of sound collected by the reference sound microphone 12, for example. May be output as

  When the direction of voice arrival is detected based on the power information of the sound collection signal 21 and the sound collection signal 22 (see FIG. 5), the following processing can be performed. Here, the difference between the magnitude of the collected sound signal 21 of the voice microphone 11 and the magnitude of the collected sound signal 22 of the reference sound microphone 12 is referred to as a power difference PD2. The predetermined value is P (positive value).

  When the relationship of power difference PD2 ≧ P holds, for example, the adaptive filter 18 performs normal noise reduction processing. That is, in this case, since the magnitude of the collected sound signal 21 of the voice microphone 11 is larger than the magnitude of the collected sound signal 22 of the reference sound microphone 12, the adaptive filter 18 is included in the collected sound signal (voice signal) 21. The output signal 27 is generated by reducing the noise component generated by using the collected sound signal (reference signal) 22. At this time, the voice section determination unit 15 can determine the voice section based on the sound collection signal 21 of the voice microphone 11.

  When the relationship of power difference PD2 ≦ −P is established, for example, the sound collection signal 21 and the sound collection signal 22 may be interchanged in the adaptive filter 18. That is, in this case, since the magnitude of the collected sound signal 22 of the reference microphone 12 is larger than the magnitude of the collected sound signal 21 of the sound microphone 11, the collected sound signal 22 of the reference sound microphone 12 is treated as a sound signal. The collected sound signal 21 of the sound microphone 11 is handled as a reference signal. Then, in the adaptive filter 18, the output signal 27 can be generated by reducing the noise component included in the collected sound signal (audio signal) 22 using the collected sound signal (reference signal) 21. At this time, the speech section determination unit 15 may determine the speech section based on the sound collection signal 22 of the reference sound microphone 12.

  When the relationship of -P <power difference PD2 <P is established, it is possible to determine that the collected sound signals 21 and 22 are unsuitable for noise reduction processing in the adaptive filter 18. In this case, the adaptive filter 18 may output the collected sound signal 21 or the collected sound signal 22 as they are. That is, when the absolute value of the power difference between the magnitude of the sound pickup signal 21 of the sound microphone 11 and the magnitude of the sound pickup signal 22 of the reference sound microphone 12 is smaller than a predetermined value (P), the adaptive filter 18 is noisy. The sound collection signal 21 or the sound collection signal 22 may be output as the output signal 27 without performing the reduction process.

  In this case, since the collected sound signals 21 and 22 having a small power difference PD2 are signals unsuitable for noise reduction processing, the adaptive filter control unit 17 further determines the phase in order to select a condition that is a little better. For example, when the phase of the sound collection signal 21 of the sound microphone 11 is earlier than the phase of the sound collection signal 22 of the reference sound microphone 12, the sound collection signal 21 may be output as the output signal 27. For example, when the phase of the sound collection signal 21 of the sound microphone 11 is slower than the phase of the sound collection signal 22 of the reference sound microphone 12, the sound collection signal 22 may be output as the output signal 27.

  FIG. 6 is a block diagram illustrating an example of the adaptive filter 18. The adaptive filter 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, an output signal selection unit 76, and a selector 77.

  When 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, according to the control signal 26 (for example, the sound direction information 25) output from the adaptive filter control unit 17. And the case where the sound collection signal 21 and the sound collection signal 22 are output as the reference signal 82 and the sound signal 81, respectively. For example, when the phase of the sound collection signal 21 of the sound microphone 11 is earlier than the phase of the sound collection signal 22 of the reference sound microphone 12, the selector 77 converts the sound collection signal 21 and the sound collection signal 22 into the sound signal 81 and the sound collection signal 22, respectively. The reference signal 82 is output. On the other hand, when the phase of the sound collection signal 22 of the reference microphone 12 is earlier than the phase of the sound collection signal 21 of the sound microphone 11, the selector 77 converts the sound collection signal 21 and the sound collection signal 22 into the reference signal 82 and the sound, respectively. The signal 81 is output.

For example, when the magnitude of the sound collection signal 21 of the sound microphone 11 is larger than the sound collection signal 22 of the reference sound microphone 12, the selector 77 outputs the sound collection signal 21 and the sound collection signal 22 respectively. The signal 81 and the reference signal 82 are output. On the other hand, when the magnitude of the collected sound signal 22 of the reference microphone 12 is larger than the magnitude of the collected sound signal 21 of the sound microphone 11, the selector 77 converts the collected sound signal 21 and the collected sound signal 22 into the reference signal 82 and the collected sound signal 22, respectively. An audio signal 81 is output.

  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 according to the control signal 26 (for example, the voice direction information 25 and the voice section signal 24). That is, the adaptive coefficient adjustment unit 74 adjusts the coefficient so that the adaptive error is reduced when the speech section information 24 indicates a noise section (non-speech section). On the other hand, when the speech section information 24 indicates a speech section, the coefficient of the adaptive filter 18 is maintained or only the coefficient is finely adjusted. Furthermore, when the voice direction information 25 indicates that the voice is coming from an inappropriate direction, the adaptive coefficient adjusting unit 74 maintains the coefficient of the adaptive filter 18 or only finely adjusts the coefficient. To do. When the voice direction information 25 indicates that the voice is coming from an inappropriate direction, it is possible to suppress the cancellation of the voice component by intentionally reducing the noise reduction effect by the noise reduction processing. . Even when the voice section information 24 indicates a noise section (non-voice section) and the voice direction information 25 indicates that voice is coming from an inappropriate direction, the coefficient adjustment unit 74 Maintains the coefficients of the adaptive filter 18 or only fine tunes the coefficients. Therefore, it is possible to suppress cancellation when a voice component is input.

  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 noise-reduced signal 84 and the feedback signal 85 are the same signal.

  The output signal selection unit 76 outputs the audio signal 81 as it is as the output signal 27 according to the control signal 26 (for example, the audio direction information 25) output from the adaptive filter control unit 17, or after noise reduction processing. Whether to output the signal 84 as the output signal 27 is selected. For example, when the voice direction information 25 indicates that voice is coming from an inappropriate direction (for example, when −T <phase difference PD1 <T), the output signal selection unit 76 outputs the voice signal 81. The signal 27 is output as it is. On the other hand, when the voice direction information 25 indicates that the voice is coming from an appropriate direction (for example, when the phase difference PD1 ≧ T and the phase difference PD1 ≦ −T), the output signal selection unit 76 reduces the noise. The processed signal 84 is output as the output signal 27.

  Next, operation | movement of the noise reduction apparatus 1 concerning this Embodiment is demonstrated. FIG. 7 is a flowchart for explaining the operation of the noise reduction apparatus 1 according to the present embodiment. This flowchart is started, for example, when reception of sound is started.

  The voice direction information 25 generated by the voice direction detector 16 is updated when it is certain that it is a voice section. Therefore, the voice direction information 25 is initialized in advance and set to a predetermined initial value (step S1). Here, the initial value is a parameter that is set when, for example, a device including a noise reduction device is used in an appropriate state (when the position of the microphone is used in an appropriate state).

  Next, it is determined whether or not the sound collected by the voice microphone 11 is a voice section using the voice section determination unit 15 (step S2). At this time, the voice section can be reliably determined by tightening the conditions for determining the voice section. In the noise reduction apparatus 1 shown in FIG. 1, it is assumed that the voice has a high probability of being picked up by the voice microphone 11, and the voice section determination unit 15 is based only on the sound pickup signal 21 of the voice microphone 11. The case where the speech section is determined 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, as in the noise reduction device 2 shown in FIG. 8, the speech section determination unit 19 determines the speech section based on the sound collection signal 21 of the sound microphone 11 and the sound collection signal 22 of the reference sound microphone 12. It may be configured.

  When the voice section is detected (step S3: Yes), the voice section determination unit 15 outputs the voice section information 23 and 24 to the voice direction detector 16 and the adaptive filter control unit 17, respectively. Then, the voice direction detector 16 detects the voice arrival direction based on the collected sound signal 21 and the collected sound signal 22 (step S4). The method for detecting the voice arrival direction is, for example, a method for detecting the voice arrival direction based on the phase difference between the sound pickup signal 21 and the sound pickup signal 22, or the size of the sound pickup signal 21 of the sound microphone 11 and the reference sound. For example, there is a method of detecting the direction of voice arrival based on the power information (that is, the difference or ratio of the collected sound signals) relating to the magnitude of the collected sound signal 22 of the microphone 12.

  The voice direction detector 16 updates the voice direction information 25 with the newly obtained voice arrival direction (step S5). On the other hand, when it is determined by the speech segment determining unit 15 that the speech segment is not a speech segment (step S3: No), the speech direction detector 16 does not newly detect the voice arrival direction, so the speech direction information 25 is not updated. . In a case other than the voice section, even if the phase difference or power information between the collected sound signal 21 and the collected sound signal 22 as described above is detected, the collected sound signal 21 and the collected sound signal 22 do not include sound. This is because the possibility is high.

As described above, the voice direction information 25 generated by the voice direction detector 16 is preferably updated when it is certain that the voice section is a voice section. In the noise reduction apparatus 1 shown in FIG. 1, the voice section information 23 and the voice section information 24 are signals that are simultaneously output from one voice section determiner 15, but are output to the voice direction detector 16 as a modification. The speech segment information may be speech segment information determined with stricter conditions than the speech segment information output to the adaptive filter control unit 17.
In other words, the voice segment information output to the voice direction detector 16 may be voice segment information determined to be a voice segment with a higher probability than the voice segment information output to the adaptive filter control unit 17.

As a more specific first example, two conditions, a first condition and a second condition that is stricter than the first condition, are set in one voice section determiner 15 and two voice sections are determined simultaneously. , Respectively, to the applied filter control unit 17 and the voice direction detector 16. As a more specific second example, instead of the speech segment determining unit 15, a first speech segment determining unit (not shown) for adaptive filter control and a speech direction detecting unit different from the adaptive filter controlling unit are used. The second voice segment determiner (not shown) is provided, and the sound collection signal 21 is input from the AD converter 13 to both the first voice segment determiner and the second voice segment determiner. The first speech segment determiner performs speech segment determination under the first condition based on the collected sound signal 21, and outputs first speech segment information as a result of performing speech segment determination to the adaptive filter control unit 17. The second speech segment determination device performs speech segment determination under a second condition that is stricter than the first condition based on the collected sound signal 21, and the second speech segment information obtained as a result of the speech segment determination is used as the speech direction detector 16. Is output.

As a method of making the second condition stricter than the first condition, for example, as an example of using the speech segment determination technique A for the first speech segment determiner and the second speech segment determiner, Each SNR is acquired, and when determining whether the target spectrum has a peak that is a feature of speech, it is determined using the SNR and a predetermined first threshold. For example, it may be possible to set the SNR value to a larger value in the two speech segment determiner than in the first speech segment determiner.

According to these modified examples, in the voice section determination used for adaptive filter control, the condition for determining the voice section is relaxed (the threshold is set to be easily determined as the voice section), so that there is a lot of noise. In the environment, it is possible to prevent the voice from being erased without being able to accurately determine the voice segment, and in the voice segment determination used for voice direction detection, the condition for determining the voice segment is made strict (voice By setting a threshold value that is difficult to determine as a section), it is possible to accurately determine the position of the speaker. That is, since the positions of the microphone and the speaker are often fixed during a call, the voice direction detector needs to update the voice direction information only when the voice section is detected under severe conditions. Therefore, it is effective to make the condition of the speech section to be output to the speech direction detector strict (setting a threshold value that is difficult to determine the speech section).

  Next, the adaptive filter control unit 17 acquires the current voice direction information 25 based on the voice direction update performed in the past from the voice direction detector 16 (step S6). Then, it is determined whether or not the reference sound collected by the reference sound microphone 12 can be used to reduce noise components contained in the sound collected by the sound microphone 11 (step S7).

  When the adaptive filter control unit 17 determines that the noise reduction process can be performed using the reference sound collected by the reference sound microphone 12 (step S7: Yes), the adaptive filter control unit 17 performs the noise reduction process by the adaptive filter 18 (Step S8). On the other hand, when the adaptive filter control unit 17 determines that it is impossible to perform the noise reduction process using the reference sound collected by the reference sound microphone 12 (step S7: No), the adaptive filter 18 Noise reduction processing by is not performed.

  For example, the sound direction is detected using the phase difference between the phase of the sound component included in the sound collection signal 21 of the sound microphone 11 and the phase of the sound component included in the sound collection signal 22 of the reference sound microphone 12. If so, proceed as follows:

  When the relationship of phase difference PD1 ≧ T is established (step S7: Yes), for example, normal noise reduction processing is performed in the adaptive filter 18 (step S8). On the other hand, when the relationship of the phase difference PD1 ≦ −T is established (step S7: Yes), for example, the collected sound signal 22 of the reference sound microphone 12 is treated as a sound signal, and the collected sound signal 21 of the sound microphone 11 is used as a reference signal. deal with. Then, the adaptive filter 18 can generate the output signal 27 by reducing the noise component included in the collected sound signal 22 using the collected sound signal 21 (step S8).

  Further, when the relationship of -T <phase difference PD1 <T is established, there is a high possibility that the distance between the sound microphone 11 and the sound source and the distance between the reference sound microphone 12 and the sound source are equal to each other. , 22 can be determined to be unsuitable signals for noise reduction processing in the adaptive filter 18 (step S7: No). In this case, the adaptive filter 18 outputs the collected sound signal 21 or the collected sound signal 22 as an output signal without performing noise reduction processing.

  In this case, for example, when the magnitude of the sound collection signal 21 of the sound microphone 11 is larger than the magnitude of the sound collection signal 22 of the reference sound microphone 12, the sound collection signal 21 may be output as the output signal 27. For example, when the sound collection signal 21 of the sound microphone 11 is smaller than the sound collection signal 22 of the reference sound microphone 12, the sound collection signal 22 may be output as the output signal 27.

  For example, when the direction of sound is detected using the magnitude of the sound pickup signal 21 of the sound microphone 11 and the sound pickup signal 22 of the reference sound microphone 12, the following processing is performed.

  When the relationship of power difference PD2 ≧ P is established (step S7: Yes), for example, normal noise reduction processing is performed in the adaptive filter 18 (step S8). On the other hand, when the relationship of power difference PD2 ≦ −P is established (step S7: Yes), for example, the collected sound signal 22 of the reference sound microphone 12 is treated as a sound signal, and the collected sound signal 21 of the sound microphone 11 is used as a reference signal. deal with. Then, the adaptive filter 18 can generate the output signal 27 by reducing the noise component included in the collected sound signal 22 using the collected sound signal 21 (step S8).

  When the relationship of -P <power difference PD2 <P holds, the distance between the sound microphone 11 and the sound source and the distance between the reference sound microphone 12 and the sound source are highly likely to be equal. 22 can be determined to be a signal unsuitable for noise reduction processing in the adaptive filter 18 (step S7: No). In this case, the adaptive filter 18 outputs the collected sound signal 21 or the collected sound signal 22 as an output signal without performing noise reduction processing.

  In this case, for example, when the phase of the sound collection signal 21 of the sound microphone 11 is earlier than the phase of the sound collection signal 22 of the reference sound microphone 12, the sound collection signal 21 may be output as the output signal 27. For example, when the phase of the sound collection signal 21 of the sound microphone 11 is slower than the phase of the sound collection signal 22 of the reference sound microphone 12, the sound collection signal 22 may be output as the output signal 27.

  The noise reduction device 1 checks whether or not sound (voice or noise) is received by the voice microphone 11 or the like (step S9). And when the sound is received (step S9: Yes), the process after step S2 is repeated. On the other hand, when the sound is not received (step S9: No), the noise reduction process by the noise reduction apparatus 1 ends.

  Next, a voice input device using the noise reduction device according to the present embodiment will be described. FIG. 9 is a diagram illustrating an example of a voice input device 500 using the noise reduction device according to the present embodiment. FIG. 9A is a front view of the voice input device 500, and FIG. 9B is a rear view of the voice input device 500. As shown in FIG. 9, 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 is 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. 9A, a speaker 506 and an audio microphone 505 are provided on the front surface of the main body 501. As shown in FIG. 9B, 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.

  Next, a radio communication apparatus (transceiver) 600 using the noise reduction apparatus according to this embodiment will be described. FIG. 10 is a diagram illustrating an example of a wireless communication device 600 using the noise reduction device according to the present embodiment. FIG. 10A is a front view of the wireless communication apparatus 600, and FIG. 10B is a rear view of the wireless communication apparatus 600. As shown in FIG. 10, 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. 7 is started, and the user stops pressing the PTT 608 and the sound transmission ends. In addition, the processing of the noise reduction apparatus 1 as shown in FIG.

  As described in the problem of the present invention, the techniques disclosed in Patent Documents 1 to 3 have a problem that noise components included in an audio signal cannot be appropriately reduced when the ambient noise level is high. It was.

  In other words, the conventional noise reduction device does not consider the situation where the surrounding noise level is high, and detects the voice arrival direction even in a situation where the voice cannot be sufficiently picked up. However, the noise component contained in the audio signal cannot be reduced appropriately.

  For example, a portable wireless communication device such as a transceiver is often used in a factory where there is a considerably high level of noise such as an operation sound of a work machine, a hustle and bustle, or an intersection. For this reason, in a portable wireless communication apparatus such as a transceiver, it is required to reduce noise components mixed in the microphone.

  In addition, unlike a mobile phone, a transceiver may be used to listen to sound transmitted from a speaker on the main unit side away from the ear. Therefore, the transceiver is generally carried away from the body, and there are various styles for holding the transceiver. Furthermore, a speaker microphone device (voice input device) in which a sound collection unit (microphone) and a playback unit (speaker) are separated from the transceiver main body to improve portability can provide a convenient usage pattern. For example, there are cases where the speaker talks without being conscious of heading to the microphone, such as hanging from the neck or placing it on the shoulder, or speaking from a direction closer to the back of the microphone rather than the front side of the microphone receiver. Therefore, when the speaker microphone device is used, the sound does not necessarily come from an ideal direction.

  Therefore, in order to implement noise reduction processing in transceivers and speaker microphone devices used in such an environment, it is necessary to ensure that the speech period in which speech is actually being emitted, while speech is hindered by high level noise. Therefore, it is necessary to detect the voice arrival direction only in the voice section.

  On the other hand, in the noise reduction apparatus according to the present embodiment, by using the speech section determination unit 15, it is possible to determine a section where sound is emitted even when the noise level is high. Then, when it is determined that the voice section is determined to be a voice section by the voice section determiner 15, the voice direction detector 16 detects the voice arrival direction and updates the voice direction information. Therefore, it is possible to reduce the processing amount for detecting the voice arrival direction by the voice direction detector 16. Further, since the voice direction detector 16 updates the voice direction information in the voice section, it is possible to obtain highly reliable voice direction information. Since the adaptive filter 18 can perform noise reduction processing based on highly reliable voice direction information and voice section information, the noise component included in the voice signal can be appropriately obtained even under various environments. Can be reduced.

As a more specific effect, for example, noise coming from behind the speaker can be reduced. For example, even when the sound source comes from various directions, the processing load of a predetermined adaptive filter can be handled without increasing the calculation load.
And the circuit scale, power consumption, and cost are reduced. Further, for example, even when a sound source exists at an intermediate position between the voice microphone and the reference microphone, it is possible to prevent the noise level from being lowered to a necessary voice level. Moreover, it can cope with an environment where a high noise level is mixed.

  As described above, according to the invention according to the present embodiment, a noise reduction device, a voice input device, a wireless communication device, and a noise reduction device that can appropriately reduce noise components included in a voice signal even under various environments, A noise reduction method can be provided.

<Embodiment 2>
Next, a second embodiment of the present invention will be described.
FIG. 11 is a block diagram of the noise reduction device 3 according to the second embodiment. The noise reduction device 3 according to the present embodiment includes two reference sound microphones and a signal determination unit 116, as compared to the noise reduction device 1 according to the first embodiment shown in FIG. Is different.

  The noise reduction device 3 shown in FIG. 11 includes an audio microphone 101, a reference sound microphone A (102), a reference sound microphone B (103), AD converters 104, 105, and 106, an audio section determination unit 115, and a signal determination unit. 116, an adaptive filter control unit 117, and an adaptive filter 118.

  The sound microphone 101 and the reference sound microphones 102 and 103 can pick up sounds including sound components and noise components, respectively. The sound microphone 101 collects sound mainly including sound components and converts it into an analog signal, and outputs the converted analog signal to the AD converter 104. The reference sound microphone A (102) collects a sound mainly including a noise component, converts it into an analog signal, and outputs the converted analog signal to the AD converter 112. The reference sound microphone B (103) collects a sound mainly including a noise component and converts it into an analog signal, and outputs the converted analog signal to the AD converter 106. For example, the noise component contained in the sound collected by the reference sound microphone A (102) or the reference sound microphone B (103) reduces the noise component contained in the sound collected by the sound microphone 101. Used for.

  In the noise reduction device 3 according to the present embodiment, an example in which the sound microphone 101 and the reference sound microphones 102 and 103 are connected will be described. However, in addition to the case where three microphones are connected to the noise reduction device 3, for example, four or more microphones may be provided by further adding a reference sound microphone.

  The AD converter 104 samples the analog signal output from the audio microphone 101 at a predetermined sampling rate and converts it into a digital signal, and generates a sound collection signal 111. The collected sound signal 112 generated by the AD converter 104 is output to the speech section determination unit 115, the signal determination unit 116, and the adaptive filter 118.

  The AD converter 105 samples the analog signal output from the reference sound microphone A (102) at a predetermined sampling rate and converts the sampled signal into a digital signal, thereby generating a sound pickup signal 112. The collected sound signal 112 generated by the AD converter 105 is output to the signal determination unit 116 and the adaptive filter 118.

  The AD converter 106 samples the analog signal output from the reference sound microphone B (103) at a predetermined sampling rate and converts it into a digital signal, thereby generating a sound pickup signal 113. The collected sound signal 113 generated by the AD converter 106 is output to the signal determination unit 116 and the adaptive filter 118.

  The frequency band of voice is about 100 Hz to 4000 Hz. Therefore, an analog signal including an audio component can be handled as a digital signal by setting the sampling frequency in the AD converters 104, 105, and 106 to about 8 kHz to 12 kHz.

  The voice segment determination unit 115 determines a voice segment based on the sound collection signal 111 output from the AD converter 104. When the speech segment determination unit 115 determines that the speech segment is a speech segment, it outputs the speech segment information 123 and 124 to the signal determination unit 116 and the adaptive filter control unit 117, respectively.

  An arbitrary technique can be used for the voice segment determination processing in the voice segment determiner 115. However, when the noise reduction device is used in an environment where the noise level is high, it is necessary to determine the speech section with high accuracy. In this case, for example, by using the technology described in Japanese Patent Application No. 2010-260798 (speech segment determination technology A) and the technology described in Japanese Patent Application No. 2011-020659 (speech segment determination technology B), It can be determined with high accuracy. Note that since the speech segment determination technique A and the speech segment determination technique B have been described in the first embodiment, a duplicate description is omitted.

  Further, in the noise reduction device 3 shown in FIG. 11, it is assumed that the voice has a high probability of being picked up by the voice microphone 101, and the voice section determination unit 115 is based only on the sound pickup signal 111 of the voice microphone 101. The case where the speech section is determined is shown. However, depending on how the noise reduction device is used, there may be cases where the reference sound microphone A (102) or the reference sound microphone B (103) collects more sound than the sound microphone 101. Therefore, as shown in FIG. 8, in addition to the sound collection signal 111 of the sound microphone 101, based on the sound collection signal 112 of the reference sound microphone A (102) and the sound collection signal 113 of the reference sound microphone B (103), You may comprise so that the audio | voice area determination device 115 may determine an audio | voice area.

  The signal determination unit 116 determines two sound collection signals to be used for noise reduction processing from the sound collection signal 111, the sound collection signal 112, and the sound collection signal 113, and a phase difference between the determined two sound collection signals. Get information. The signal determination unit 116 outputs the collected sound signal selection information 125 regarding the two collected sound signals used for the noise reduction processing and the phase difference information 126 of the determined two collected sound signals to the adaptive filter control unit 117.

  For the reasons described in the first embodiment, the phase difference between the collected sound signal 111 and the collected sound signal 112, the phase difference between the collected sound signal 111 and the collected sound signal 113, and the phase difference between the collected sound signal 112 and the collected sound signal 113. Is acquired, the sampling frequency of the collected sound signal 111, the collected sound signal 112, and the collected sound signal 113 input to the signal determination unit 116 may be 24 kHz or more.

  In addition, the noise reduction device 3 according to the present embodiment includes two reference sound microphones. In this case, for example, as shown in FIGS. 19B and 21B, it is preferable to arrange two reference sound microphones on the diagonal line at a predetermined distance. By arranging in this way, the path of the sound of one reference sound microphone is hindered by the influence of the hand having the voice input device shown in FIG. 19B or the wireless communication device shown in FIG. Even so, by using the other reference sound microphone, the direction of sound can be detected appropriately.

  FIG. 12 is a block diagram illustrating the signal determination unit 116 included in the noise reduction device 3 according to the present embodiment. 12 includes a cross-correlation value calculation unit 131, a power information acquisition unit 132, a phase difference information acquisition unit 133, a reference signal selection unit 134, a cross-correlation value calculation unit 135, a phase difference calculation unit 136, and A determination unit 137 is provided.

As described with reference to FIG. 11, when the collected sound signal of the voice microphone 101 is determined to be a voice section by the voice section determiner 115, the voice section determiner 115 sends the voice section information 123 to the signal determination unit 116. Output.
When the speech section information 123 is input to the signal determination unit 116 shown in FIG. 12, the cross-correlation value calculation unit 131 acquires the sound collection signal 112 of the reference sound microphone A (102) and the reference sound microphone B (103). Information regarding the correlation between the sound pickup signal 112 and the sound pickup signal 112 is acquired using the sound pickup signal 113, and the acquired information is output to the phase difference information acquisition unit 133. The phase difference information acquisition unit 133 calculates the phase difference between the two signal waveforms determined to have high correlation, thereby obtaining the phase difference between the phase of the sound component of the sound pickup signal 112 and the phase of the sound component of the sound pickup signal 113. Can be sought. Further, the phase difference information acquisition unit 133 outputs the acquired phase difference information of the collected sound signal 112 and the collected sound signal 113 to the reference signal selection unit 134 and the determination unit 137.

  Here, a method for obtaining information on the correlation between the sound collection signal 112 and the sound collection signal 112 by the cross correlation value calculation unit 131, and a phase difference between the sound collection signal 112 and the sound collection signal 113 by the phase difference calculation unit 133 are obtained. The method is the same as the method described in the voice direction detector 16 ′ in FIG. 4 (particularly, refer to the cross-correlation value calculation unit 55 and the phase difference information acquisition unit 56), and a duplicate description will be omitted.

  In the present embodiment, the signal determination unit 116 calculates the phase difference when the speech segment determination unit 115 determines that the speech segment is a speech segment. Therefore, even when noise is mixed in the collected sound signal, the phase difference can be calculated with high accuracy.

  In addition, when the speech section determination unit 115 determines that the power information acquisition unit 132 is a speech section, the power information acquisition unit 132 and the reference sound microphone B (103) and the magnitude of the collected sound signal 112 of the reference sound microphone A (102). The power information (that is, the power ratio or power difference between the collected sound signal 112 and the collected sound signal 113) is acquired based on the magnitude of the collected sound signal 113. The acquired power information is output to the reference signal selection unit 134. The method for obtaining the power information of the collected sound signal 112 and the collected sound signal 113 by the power information acquisition unit 132 is the same as the method described for the voice direction detector 16 of FIG.

  An ideal reference signal that can accurately update the filter coefficient of the adaptive filter 118 has two conditions. The first condition A is that there is little mixing of audio components. The second condition B is that it is close to the characteristics of the noise component mixed in the voice. In order to reduce the mixing of the sound component into the reference signal, it is preferable that the distance of the reference microphone is farther from the sound source. The position where the distance between the sound source and the reference microphone is long can be grasped by examining the point where the phase is most delayed. For example, when there is a reference microphone A (102) and a reference sound microphone B (103) as in the noise reduction device 3 according to the present embodiment, the sound collection signal 112 of the reference microphone A (102) is referred to. It is preferable to compare the sound pickup signal 113 of the sound microphone B (103) and select the one with the later phase as an ideal reference signal. Of course, if the distance from the sound source of the sound is far, the sound volume (sound pressure level) will also decrease, but in order to investigate whether it is close to the characteristics of the noise component mixed in the sound microphone, which is another condition It is necessary to consider the external environment where the noise reduction device 3 is used at the same time. In other words, from the viewpoint of acoustic characteristics, the influence of the shield is large, and whether the vicinity of the microphone opening is open to the outside together with the phase difference, that is, the sound pressure level of the sound input to the microphone is By observing whether it is maintained, it is possible to grasp whether it is suitable as a reference signal.

  Based on the phase difference information output from the phase difference information acquisition unit 133 and the power information output from the power information acquisition unit 132, the reference signal selection unit 134 serves as a reference signal of the sound collection signal 112 and the sound collection signal 113. Select an appropriate sound pickup signal. Thus, by using the phase difference information and the power information for selection of the reference signal, it is possible to reflect the influence of the external environment when selecting the reference signal.

  The cross-correlation value calculation unit 135 uses the sound collection signal 111 of the sound microphone 101 and the sound collection signal 138 selected by the reference signal selection unit 134 to acquire information on the correlation between these sound collection signals. The acquired information is output to the phase difference calculation unit 136. The phase difference calculation unit 136 obtains the phase difference between the two signal waveforms that are determined to have high correlation, so that the phase of the sound component of the sound collection signal 111 and the sound collection signal 138 selected by the reference signal selection unit 134 are obtained. The phase difference from the phase of the audio component can be obtained. The phase difference calculation unit 136 outputs the acquired phase difference information to the determination unit 137.

  Here, a method for acquiring information on the correlation between the sound collection signal 111 by the cross-correlation value calculation unit 135 and the sound collection signal 138 selected by the reference signal selection unit 134, and these sound collections by the phase difference calculation unit 136 The method for obtaining the signal phase difference is the same as the method described in the voice direction detector 16 ′ of FIG. 4 (particularly, refer to the cross-correlation value calculation unit 55 and the phase difference information acquisition unit 56), and thus a duplicate description. Is omitted.

  In the signal determination unit 116 shown in FIG. 12, the cross-correlation value calculation unit 131 and the cross-correlation value calculation unit 135, and the phase difference information acquisition unit 133 and the phase difference calculation unit 136 are provided separately. May be shared for the same processing.

  Based on the phase difference information output from the phase difference calculation unit 136, the determination unit 137 can use the sound collection signal 111 as an audio signal, or the sound collection signal selected by the reference signal selection unit 134 (that is, the sound collection signal). It is determined whether the sound signal 112 or 113) can be used as a reference signal. Then, the determination unit 137 determines two sound collection signals to be used for noise reduction processing, and outputs sound collection signal selection information 125 regarding the two selected sound collection signals to the adaptive filter control unit 117. Further, the determination unit 137 outputs the phase difference information 126 between the two selected sound pickup signals to the adaptive filter control unit 117.

  Next, the operation in the signal determination unit 116 will be described. FIGS. 13 and 14 are flowcharts for explaining the operation of the signal determination unit 116. FIG. 13 shows reference microphone selection processing for selecting a reference microphone. Here, the sound collection signal 111 can be used as an audio signal, and the determination unit 137 determines that the sound collection signal selected by the reference signal selection unit 134 (that is, the sound collection signal 112 or 113) can be used as a reference signal. It is assumed that

As shown in FIG. 13, the signal determining unit 116 first sets a reference sound microphone to be a reference and a reference sound microphone to be compared when comparing phase differences (step S21). For example, the reference sound microphone A (102) is used as a reference, and the reference sound microphone B (103) is a comparison target. Next, in the cross-correlation value calculation unit 131 and the phase difference information acquisition unit 133, the phase difference information between the sound collection signal 112 of the reference sound microphone A (102) and the sound collection signal 113 of the reference sound microphone B (103) is obtained. get. Further, the power information acquisition unit 132 acquires the power information (in this case, the power ratio) of the sound collection signal 112 and the sound collection signal 113 (step S22).

  Next, the reference signal selection unit 134 determines whether there is a predetermined phase difference between the collected sound signal 112 and the collected sound signal 113 (step S23). That is, it is determined whether the phase difference between the collected sound signal 112 and the collected sound signal 113 is within a predetermined range (that is, whether the condition of −T <phase difference <T is satisfied). Here, T is a predetermined reference value and can be set arbitrarily. When the condition of −T <phase difference <T is satisfied (step S23: Yes), it is determined that there is no predetermined phase difference. In this case, the reference signal selection unit 134 determines a signal to be selected based on the power ratio (A / B) of the sound collection signal 112 and the sound collection signal 113. For example, when the power ratio (A / B) between the sound collection signal 112 and the sound collection signal 113 is larger than 1 (step S24: Yes), the sound collection signal 112 (that is, the reference sound microphone A) is selected (step S24). S28). On the other hand, if the power ratio (A / B) between the sound collection signal 112 and the sound collection signal 113 is smaller than 1 (step S24: No), the sound collection signal 113 (that is, the reference sound microphone B) is selected (step S24). S29). In step S24, the power ratio criterion is 1, but this value is not limited to this and can be arbitrarily changed. For example, the phase difference reference value T in step S23 may be changed.

  If it is determined in step S23 that there is no predetermined phase difference, a more suitable reference signal can be selected by comparing the power ratio between the sound pickup signal 112 and the sound pickup signal 113 in step S24. That is, when there is no predetermined phase difference, no power difference occurs between the sound collection signal 112 and the sound collection signal 113 unless there is a factor such as a shield in the opening of the microphone. However, when the opening of the microphone is blocked by an obstacle such as a speaker's hand or clothes, the sound pressure level of the collected sound signal is lowered. Here, the shielding object affects the acoustic characteristics, and adversely affects the generation of a noise component in the adaptive filter. Therefore, a more suitable reference signal can be selected by selecting a signal that is less affected by the shield.

  When the condition −T <phase difference <T is not satisfied (step S23: No), it is determined that there is a predetermined phase difference. In this case, the reference signal selection unit 134 determines which phase is earlier. That is, it is determined whether the condition of phase difference ≧ T is satisfied (step S25). When the condition of phase difference ≧ T is satisfied (step S25: Yes), the phase of the collected sound signal 112 (that is, the reference sound microphone A) is preceded. At this time, since the reference signal candidate is a signal having a late phase, the collected sound signal 113 (that is, the reference sound microphone B) is a reference signal candidate. When the power ratio (B / A) between the collected sound signal 113 and the collected sound signal 112 is larger than the predetermined value P (step S26: Yes), the power of the collected sound signal 113 is ensured (that is, shielded). Therefore, the sound collection signal 113 (that is, the reference sound microphone B) is selected as a reference signal (step S30).

  On the other hand, when the power ratio (B / A) of the sound collection signal 113 and the sound collection signal 112 is equal to or less than the predetermined value P (step S26: No), the power of the sound collection signal 113 is ensured due to the influence of a shield or the like. It can be judged that it is not. Therefore, in this case, the sound collection signal 112 (that is, the reference sound microphone A) is selected as a reference signal (step S31). The signal power attenuates in proportion to the square of the distance to the sound source. Therefore, when there is a phase difference, the signal power of the signal having a late phase (that is, farther from the sound source) is attenuated than the signal having a fast phase. The predetermined value P of the power ratio is a threshold value obtained by adding an attenuation amount in consideration of this phase difference to an attenuation amount that cannot be ignored by the shielding object.

  Further, when the condition of phase difference ≧ T is not satisfied (step S25: No), the phase of the sound pickup signal 113 (that is, the reference sound microphone B) is preceded. At this time, since the reference signal candidate is a signal having a late phase, the collected sound signal 112 (that is, the reference sound microphone A) is a reference signal candidate. When the power ratio (A / B) between the sound collection signal 112 and the sound collection signal 113 is larger than the predetermined value P (step S27: Yes), the power of the sound collection signal 112 is ensured (that is, shielded). Therefore, the sound pickup signal 112 (that is, the reference sound microphone A) is selected as the reference signal.

  On the other hand, when the power ratio (A / B) between the sound collection signal 112 and the sound collection signal 113 is equal to or less than the predetermined value P (No in step S27), the power of the sound collection signal 112 is ensured due to the influence of a shield or the like. Since it can be determined that it is not, the collected sound signal 113 (that is, the reference sound microphone B) is selected as a reference signal (step S33).

  The reference signal selection unit 134 determines the reference sound microphone (sound collection signal) selected by the above processing as a candidate (step S34). When all the reference sound microphones have been checked (step S35: Yes), it is determined to use the reference sound microphone selected by the above processing (step S36). On the other hand, when the investigation of all the reference sound microphones has not been completed (step S35: No), the processes of steps S21 to S34 are repeated again. At this time, for example, the reference sound microphone newly selected as a target for comparison is set as a comparison target based on the reference sound microphone selected by the above processing.

  With the above processing, the microphone to be used as the reference sound microphone is determined from the reference sound microphone A (102) and the reference sound microphone B (103). That is, the collected sound signal (112 or 113) of the reference sound microphone selected from the reference sound microphone A (102) and the reference sound microphone B (103) is set as a reference signal candidate.

In the processing described above, the reference signal selection unit 134 uses the phase difference information output from the phase difference information acquisition unit 133 and the power ratio output from the power information acquisition unit 132 as a reference signal. An appropriate sound pickup signal was selected. However, the reference signal selection unit 134 may select an appropriate sound collection signal as a reference signal based only on the phase difference information output from the phase difference information acquisition unit 133. In this case, the power information acquisition unit 132 included in the signal determination unit 116 illustrated in FIG. 12 can be omitted. Further, steps S24, S26, and S27 in FIG. 13 can be omitted. Further, in step S22 in FIG. 13, only the phase difference information can be acquired, and the acquisition of the power ratio can be omitted.
At this time, if it is determined in step S23 that there is no predetermined phase difference (step S23: Yes), the sound collection signal 112 or the sound collection signal 113 can be selected as a reference signal. If it is determined in step S25 that the sound collection signal 112 is ahead (step S25: Yes), the sound collection signal 113 can be selected as a reference signal. If it is determined in step S25 that the sound collection signal 113 is ahead (step S25: No), the sound collection signal 112 can be selected as a reference signal.

  When the positional relationship between the voice microphone 101 and the speaker's mouth, which is the voice source, is good (for example, when the voice microphone is fixed to a headset or a helmet fixed to the head). The sound pickup signal 111 of the sound microphone 101 can be used as a sound signal, and the sound pickup signal (112 or 113) of the selected reference sound microphone can be used as a reference signal.

  However, for example, in a transceiver or speaker microphone device, the positional relationship between a sound source that emits sound and a sound microphone that collects sound may not be constant. For this reason, the case where the noise reduction apparatus is not used in an appropriate state is assumed, for example, when no sound is emitted toward the sound microphone or when sound is emitted toward the opening of the reference sound microphone. Therefore, it is necessary to verify whether the sound collection signal 111 of the sound microphone 101 can be used as a sound signal and the sound collection signal (112 or 113) of the selected reference sound microphone as a reference signal. By performing such a verification process, a combination of an audio signal and a reference signal estimated to have the highest noise reduction effect can be selected from the collected sound signals 111 to 113. FIG. 14 is a flowchart for explaining such verification processing.

  As shown in FIG. 14, the signal determination unit 116 first determines the reference sound microphone selected in step S36 of the reference microphone selection process shown in FIG. Step S41). Next, in the cross-correlation value calculation unit 135 and the phase difference calculation unit 136, the phase of the sound component included in the sound collection signal 111 of the sound microphone 101 and the sound collection signal 138 of the selected reference sound microphone are included. The phase difference information of the phase of the audio component is acquired (step S42).

The determination unit 137 determines whether there is a predetermined phase difference between the sound collection signal 111 and the selected sound collection signal 138 (step S43). That is, it is determined whether the phase difference between the collected sound signal 111 and the selected collected sound signal 138 is within a predetermined range (that is, whether the condition of −T <phase difference <T is satisfied). When the condition −T <phase difference <T is satisfied (step S43: Yes), it is determined that there is no predetermined phase difference. In this case, since it is presumed that the collected sound signal 111 has the same phase lag as the selected collected sound signal 138 (the collected sound signal having the most phase lag), the reference sound microphone of the earliest phase is selected. The sound collection signal (that is, the sound collection signal that has not been selected by the reference signal selection unit 134) is used as a sound signal, and the sound collection signal of the selected reference sound microphone is used as a reference signal (step S45).

  That is, since the sound collection signal 138 selected by the reference signal selection unit 134 is the sound collection signal having the most phase lag, the phase difference between the sound collection signal 111 and the selected sound collection signal 138 is within a predetermined range. That is, it can be inferred that the sound pickup signal 111 has the same phase lag as the sound pickup signal having the most phase lag. In this case, it is assumed that the voice microphone 101 does not play a role of collecting voice. Therefore, in step S45, the sound collection signal of the reference sound microphone with the earliest phase (that is, the sound collection signal not selected by the reference signal selection unit 134) is used as the sound signal, and the selected reference sound microphone is selected. The collected sound signal is used as a reference signal.

  When there are three or more reference sound microphones, a process similar to the process for detecting the collected sound signal with the maximum phase delay shown in FIG. 13 is performed, so that the reference sound microphone with the earliest phase is detected. A sound pickup signal can be determined. In the process shown in FIG. 13, the process of selecting the collected sound signal with the later phase is performed. However, when the collected sound signal with the earliest phase is determined, the collected signal with the earlier phase is selected. What is necessary is just to repeat and perform the process to perform.

  On the other hand, when the condition −T <phase difference <T is not satisfied (step S43: No), it is determined that there is a predetermined phase difference between the reference signal and the signal to be compared. In this case, the determination unit 137 determines whether or not the condition of phase difference ≧ T is satisfied (step S44). When the condition of phase difference ≧ T is satisfied (step S44: Yes), the phase of the collected sound signal 111 (that is, the voice microphone 101) is preceded. In this case, the sound collection signal 111 of the sound microphone 101 is used as a sound signal, and the sound collection signal (112 or 113) of the selected reference sound microphone is used as a reference signal (step S46).

  When the condition of phase difference ≧ T is not satisfied (step S44: No), the phase of the sound collection signal 138 of the selected reference sound microphone is preceded. In such a case, for example, it is considered that the speaker is speaking toward the reference sound microphone. Therefore, in this case, the sound collection signal 111 of the sound microphone 101 is used as a reference signal, and the sound collection signal (112 or 113) of the selected reference sound microphone is used as a sound signal (step S47).

  Based on the above processing, the determination unit 137 determines a microphone to be used for noise reduction processing in the adaptive filter 118, and determines the phase difference information (step S48). The determination unit 137 outputs information related to the two sound collection signals used for the noise reduction processing to the adaptive filter control unit 117 as the sound collection signal selection information 125.

There are two cases for the phase difference information 126. In the first case, the sound collection signal 111 of the sound microphone 101 and the sound collection signal 138 selected from the sound collection signal 112 of the reference sound microphone 102 or the sound collection signal 113 of the reference sound microphone 103 are noised. This is a case where the signal is used for reduction processing (step S46 or S47). The second case is a case where the collected sound signals 112 and 113 of the reference microphones 102 and 103 are signals for noise reduction processing (step S45).

In FIG. 12, in the first case, the determination unit 137 outputs a phase difference output such as the phase difference information 126 supplied to the adaptive filter control unit 117 from the phase difference calculation unit 136 to the adaptive filter control unit 117.
On the other hand, in the second case, the determination unit 137 outputs a phase difference output such as the phase difference information 126 supplied to the adaptive filter control unit 117 from the phase difference information acquisition unit 133 to the adaptive filter control unit 117.

The process of FIG. 14 is as outlined below. When there is one voice microphone and a plurality of reference sound microphones, the phase of a specific sound collection signal obtained from a specific reference sound microphone among the plurality of reference sound microphones (the phase of the specific sound collection signal is The phase of the collected sound signal obtained from the plurality of reference microphones is the most advanced) in some cases, the phase of the collected sound signal obtained from the sound microphone is advanced. In this case, it is preferable that the signal determination unit 116 determines the specific sound pickup signal as the first sound pickup signal in which the first noise component is reduced.

In addition, when there is one voice microphone and a plurality of reference sound microphones, the phase of a specific sound collection signal obtained from a specific reference sound microphone among the plurality of reference sound microphones (a specific sound collection signal The phase is most delayed among the phases of the collected sound signals obtained from the plurality of reference microphones), and may be delayed from the phase of the collected sound signals obtained from the sound microphone. In this case, the signal determination unit 116 determines the specific sound pickup signal as the second sound pickup signal used for reducing the noise component included in the first sound pickup signal determined as the signal whose noise is reduced. It is preferable to do.
In the process described with reference to FIG. 14, the microphone used for the noise reduction process is determined based on the phase difference information. However, the microphone used for the noise reduction process is determined in consideration of the power information in addition to the phase difference information. May be.

Specifically, in the process of FIG. 14, the signal determination unit 116 determines the sound collection signal having the most advanced phase among the plurality of sound collection signals as the first sound collection signal used for noise reduction, A sound pickup signal having a slow phase is determined as a second sound pickup signal to be used for noise component reduction by the first sound pickup signal. However, the signal determination unit 116 selects a sound collection signal having the latest phase among the plurality of sound collection signals and a level larger than a predetermined value (for example, P or more) as a noise component included in the first sound collection signal. You may determine as a 2nd sound collection signal used for reduction. Furthermore, the magnitude of the sound pickup signal having the slowest phase among the plurality of sound pickup signals may be a predetermined value or less. In this case, the signal determination unit uses the specific sound collection signal having the slowest phase next to the slowest phase among the plurality of sound collection signals to reduce the noise component included in the first sound collection signal. It is preferable to determine this as the sound pickup signal.

Further, there is a case where the phase difference is within a predetermined value except for the first sound pickup signal among the plurality of sound pickup signals (for example, −T <phase difference <T). In this case, the signal determination unit 116 uses the specific sound collection signal having the largest sound collection signal size except for the first sound collection signal to reduce the noise component included in the first sound collection signal. Preferably, it is determined as the second collected sound signal.

  The adaptive filter control unit 117 illustrated in FIG. 11 includes the speech section information 124 output from the speech section determination unit 115 and information about the two sound collection signals used for noise reduction processing output from the signal determination unit 116 (sound collection signal). A control signal for controlling the adaptive filter 118 is generated based on the selection information 125 and the phase difference information 126 of the two determined sound pickup signals, and the generated control signal 127 is output to the adaptive filter 118. Here, the control signal 127 includes voice section information 124, collected sound signal selection information 125, and phase difference information 126.

  The adaptive filter 118 uses the two collected sound signals selected from the collected sound signals 111 to 113 to generate a sound signal with reduced noise, and uses the sound signal with reduced noise as the output signal 128. Output. Here, the two sound pickup signals used for the noise reduction processing in the adaptive filter 118 are the sound pickup signals determined by the signal determination unit 116. The adaptive filter 118 artificially generates a noise component that may be included in the audio signal using the reference signal in order to reduce the noise component included in the audio signal. The adaptive filter 118 can perform the noise reduction process by subtracting the pseudo-generated noise component from the audio signal.

  The adaptive filter control unit 117 outputs a control signal 127 for the adaptive filter 118 to perform the above processing to the adaptive filter 118. The speech section information 124 supplied to the adaptive filter control unit 117 is information for determining the timing of updating the adaptive filter coefficient in the adaptive filter 118. For example, when the speech segment determining unit determines that it is not a speech segment (that is, a noise segment), the adaptive filter coefficient of the adaptive filter 118 may be updated in order to actively reduce the noise component. . On the other hand, for example, when it is determined that the speech section is determined to be a speech section, noise reduction processing may be performed using an existing adaptive filter coefficient.

  FIG. 15 is a block diagram illustrating an example of the adaptive filter 118. The adaptive filter 118 includes delay elements 171_1 to 171_n, multipliers 172_1 to 172_n + 1, adders 173_1 to 173_n, an adaptive coefficient adjustment unit 174, a subtractor 175, an output signal selection unit 176, and a selector 177.

  The selector 177 outputs two of the collected sound signals 111 to 113 as the audio signal 181 and the reference signal 182 in response to the control signal 127 output from the adaptive filter control unit 117. That is, the selector 177 selects two of the collected sound signals 111 to 113 based on the collected sound signal selection information 125 output from the signal determining unit 116, one as the audio signal 181 and the other as the reference signal 182. Output as.

  The delay elements 171_1 to 171_n, the multipliers 172_1 to 172_n + 1, and the adders 173_1 to 173_n constitute an FIR filter. The pseudo noise signal 183 is generated by processing the reference signal 182 using the delay elements 171_1 to 171_n, the multipliers 172_1 to 172_n + 1, and the adders 173_1 to 173_.

  The adaptive coefficient adjustment unit 174 adjusts the coefficients of the multipliers 172_1 to 172_n + 1 in accordance with the control signal 127 (for example, the phase difference information 126 and the voice section signal 124). That is, when the speech section information 124 indicates a noise section (non-speech section), the adaptive coefficient adjustment unit 174 adjusts the coefficient so that the adaptation error is reduced. On the other hand, when the speech section information 124 indicates a speech section, the coefficient of the adaptive filter 118 is maintained or the coefficient is finely adjusted. Furthermore, when the phase difference between the audio signal and the reference signal is within a predetermined range (that is, when there is almost no phase difference), the adaptive coefficient adjustment unit 174 maintains the coefficient of the adaptive filter 118 or sets the coefficient. Only fine adjustment. When there is almost no phase difference between the audio signal and the reference signal, it can be assumed that the audio is coming from an inappropriate direction, so the audio component is reduced by consciously reducing the noise reduction effect by the noise reduction processing. Cancellation can be suppressed.

  The subtractor 175 generates a signal 184 after noise reduction processing by subtracting the pseudo noise signal 183 from the audio signal 181, and outputs the signal 184 to the output signal selection unit 176. Also, the subtractor 175 generates a feedback signal 185 by subtracting the pseudo noise signal 183 from the audio signal 181, and outputs it to the adaptive coefficient adjustment unit 174.

  The output signal selection unit 176 outputs the audio signal 181 as it is as the output signal 128 according to the control signal 127 output from the adaptive filter control unit 117 (for example, the phase difference information 126 output from the signal determination unit 116). Or whether to output the signal 184 after the noise reduction processing as the output signal 128 is selected. For example, when there is almost no phase difference between the audio signal and the reference signal, the output signal selection unit 176 outputs the audio signal 181 as it is as the output signal 128. On the other hand, when the phase difference between the audio signal and the reference signal is greater than or equal to a predetermined value, the output signal selection unit 176 outputs the signal 184 after the noise reduction processing as the output signal 128.

  Next, operation | movement of the noise reduction apparatus 3 concerning this Embodiment is demonstrated. FIG. 16 is a flowchart for explaining the operation of the noise reduction apparatus 3 according to the present embodiment.

  The collected sound signal selection information 125 and the phase difference information 126 generated by the signal determination unit 116 are updated when it is certain that the voice section is in effect. Therefore, the sound pickup signal selection information 125 and the phase difference information 126 are initialized in advance and set to predetermined initial values (step S51). Here, the initial value is a parameter that is set when, for example, a device including a noise reduction device is used in an appropriate state (when the position of the microphone is used in an appropriate state).

  Next, it is determined whether or not the sound collected by the voice microphone 101 is a voice section using the voice section determination unit 115 (step S52). At this time, the voice section can be reliably determined by tightening the conditions for determining the voice section.

  When the speech section determination unit 115 detects a speech section (step S53: Yes), the speech section determination unit 115 outputs the speech section information 123 and 124 to the signal determination unit 116 and the adaptive filter control unit 117, respectively. Then, the signal determination unit 116 acquires the collected sound signal selection information 125 and the phase difference information 126 (step S54). The signal determination unit 116 can acquire the collected sound signal selection information 125 and the phase difference information 126 by performing the processing illustrated in FIGS. 13 and 14.

  The adaptive filter control unit 117 updates the collected sound signal selection information 125 and the phase difference information 126 included in the control signal 127 supplied to the adaptive filter 118 to the newly obtained information (step S55). On the other hand, when it is determined by the speech segment determination unit 115 that the speech segment is not a speech segment (step S53: No), the adaptive filter control unit 117 and the collected sound signal selection information 125 included in the control signal 127 supplied to the adaptive filter 118 The phase difference information 126 is not updated.

  Next, the selector 177 of the adaptive filter 118 selects a sound signal and a reference signal from the collected sound signals 111 to 113 based on the collected sound signal selection information 125 (step S56). Then, the adaptive filter 118 performs noise reduction processing using the two selected sound pickup signals (step S57).

  The noise reduction device 3 checks whether or not sound (voice or noise) is received by the voice microphone 101 or the like (step S58). And when the sound is received (step S58: Yes), the process after step S52 is repeated. On the other hand, when no sound is received (step S58: No), the noise reduction process by the noise reduction device 3 ends.

In the noise reduction device 3 according to the present embodiment, by using the speech section determination unit 115, it is possible to determine a section in which sound is emitted even when the noise level is high.
Then, when it is determined that the speech section is determined to be a speech section by the speech section determiner 115, the signal determination unit 116 determines two sound collection signals to be used for noise reduction processing from the sound collection signals 111 to 113, and the determination The phase difference information of the two collected sound signals is updated. Therefore, the information processing amount in the signal determination unit 116 can be reduced. In addition, since the signal determination unit 116 updates the collected sound signal selection information and the phase difference information in the voice section, it is possible to obtain the collected sound signal selection information and the phase difference information with high reliability. In addition, since it is possible to select two optimum sound pickup signals to be used for noise reduction processing from among a plurality of sound pickup signals, when a device using a noise reduction device is used in various states. Even if it exists, a noise reduction process can be implemented accurately.

  As described above, according to the invention according to the present embodiment, a noise reduction device, a voice input device, a wireless communication device, and a noise reduction device that can appropriately reduce noise components included in a voice signal even under various environments, A noise reduction method can be provided.

<Embodiment 3>
Next, a third embodiment of the present invention will be described.
FIG. 17 is a block diagram of the noise reduction device 4 according to the third embodiment. In the noise reduction device 4 according to the present embodiment, in addition to the sound collection signal 211 of the sound microphone 201, sound collection signals 212 and 213 of the reference sound microphones 202 and 203 are also supplied to the sound section determination unit 215. This is different from the noise reduction apparatus 3 according to the second embodiment shown in FIG. 11 in that the signal determination unit 216 supplies the collected sound signal selection information 223 to the speech section determination unit 215. Other than this, since it is the same as the noise reduction device 3 described in the second embodiment, a redundant description will be omitted as appropriate.

  The noise reduction apparatus 4 shown in FIG. 17 includes an audio microphone 201, a reference sound microphone A (202), a reference sound microphone B (203), AD converters 204, 205, and 206, an audio section determination device 215, and a signal determination unit. 216, an adaptive filter control unit 217, and an adaptive filter 218.

  The sound microphone 201, the reference sound microphones 202 and 203, and the AD converters 204, 205, and 206 included in the noise reduction device 4 according to the present embodiment are respectively noise reduction devices according to the second embodiment described with reference to FIG. 3 has the same configuration as that of the voice microphone 101, the reference sound microphones 102 and 103, and the AD converters 104, 105, and 106 provided in FIG.

  In the noise reduction apparatus according to the present embodiment, the collected sound signals 211, 212, and 213 output from the AD converters 204, 205, and 206 are supplied to the speech section determiner 215, the signal determination unit 216, and the adaptive filter 218. Is done.

  The signal determination unit 216 determines a sound collection signal used for sound section determination in the sound section determination unit 215 from the sound collection signal 211, the sound collection signal 212, and the sound collection signal 213, and collects sound used for the sound section determination. Information related to the signal is output to the speech segment determination unit 215 as the collected sound signal selection information 223. When sound is input to the noise reduction device, it can be considered that the phase of the collected sound signal including the sound is the earliest. Therefore, the signal determination unit 216 can determine, for example, a sound collection signal having the earliest phase among the sound collection signal 211, the sound collection signal 212, and the sound collection signal 213 as a sound collection signal used for sound segment determination.

  For example, the configuration of the signal determination unit 216 is the same as the configuration of the signal determination unit 116 illustrated in FIG. 12, and the operation of the signal determination unit 216 is the same as the operation illustrated in the flowcharts illustrated in FIGS. . That is, the signal determination unit 216 can determine the sound collection signal regarded as the sound signal in steps S45 to S47 in the flowchart shown in FIG. 14 as the sound collection signal used for sound section determination.

  In addition, the signal determination unit 216 determines two sound collection signals to be used for noise reduction processing from the sound collection signal 211, the sound collection signal 212, and the sound collection signal 213, and determines the two collected sound collection signals. Get phase difference information. The signal determination unit 216 outputs the collected sound signal selection information 225 regarding the two collected sound signals used for the noise reduction processing and the phase difference information 226 of the determined two collected sound signals to the adaptive filter control unit 217.

  The voice section determination unit 215 uses the collected sound signal selected according to the signal selection information 223 output from the signal determination unit 216 among the collected sound signal 211, the collected sound signal 212, and the collected sound signal 213. Determine the interval. Then, the speech segment determining unit 215 outputs the speech segment information 224 to the adaptive filter control unit 217 when it is determined to be a speech segment.

  Any technique can be used for the speech segment determination processing in the speech segment determiner 215. However, when the noise reduction device is used in an environment where the noise level is high, it is necessary to determine the speech section with high accuracy. In this case, for example, by using the technology described in Japanese Patent Application No. 2010-260798 (speech segment determination technology A) and the technology described in Japanese Patent Application No. 2011-020659 (speech segment determination technology B), It can be determined with high accuracy. Note that since the speech segment determination technique A and the speech segment determination technique B have been described in the first embodiment, a duplicate description is omitted.

  The adaptive filter control unit 217 determines the collected sound signal selection information 225 and the phase difference information 226 used for the control of the adaptive filter 218 according to the voice section information 224 output from the voice section determiner 215. That is, the collected sound signal selection information 225 and the phase difference information 226 output from the signal determination unit 216 are supplied to the adaptive filter control unit 217 at every predetermined timing. However, this includes sound collection signal selection information 225 and phase difference information 226 acquired at a timing other than the voice interval. The collected sound signal selection information 225 and the phase difference information 226 acquired at timings other than the voice section are information with low accuracy.

  On the other hand, the collected sound signal selection information 225 and the phase difference information 226 when the speech segment determining unit 215 determines that the speech segment is a highly accurate information. Therefore, the adaptive filter control unit 217 uses the collected sound signal selection information 225 and the phase difference information 226 at the timing determined as the speech section by the speech section determiner 215, and the collected sound signal selection information 225 used for controlling the adaptive filter 218 and The phase difference information 226 is determined. As described above, by using the collected sound signal selection information 225 and the phase difference information 226 in the voice section for the control of the adaptive filter 218, the adaptive filter 218 can reduce noise with high accuracy.

  Here, the speech segment information 224 is output to the adaptive filter control unit 217 after the speech segment determination process in the speech segment determiner 215. Therefore, the sound collection signal selection information 225 and the phase difference information 226 at the predetermined timing are supplied to the adaptive filter control unit 217, and the voice section information 224 corresponding to the predetermined timing is supplied to the adaptive filter control unit 217. It is earlier than the timing. Therefore, the adaptive filter control unit 217 may include a buffer that can temporarily hold the supplied sound pickup signal selection information 225 and phase difference information 226 in order to adjust these timings. As described above, the adaptive filter control unit 217 temporarily holds the collected sound signal selection information 225 and the phase difference information 226, thereby selecting the collected sound signal selection information 225 and the phase difference information 226 corresponding to the speech section information 224. be able to.

  In addition, the adaptive filter control unit 217 outputs the voice section information 224 output from the voice section determiner 215, the collected sound signal selection information (information on two sound collection signals used for noise reduction processing) 225, and the determined 2 Based on the phase difference information 226 of the two collected sound signals, a control signal 227 for controlling the adaptive filter 218 is generated, and the generated control signal 227 is output to the adaptive filter 218. Here, the control signal 227 includes voice section information 224, collected sound signal selection information 225, and phase difference information 226.

  The adaptive filter 218 uses the two collected sound signals selected from the collected sound signals 211 to 213 to generate a sound signal with reduced noise, and uses the sound signal with reduced noise as the output signal 228. Output. Here, the two sound pickup signals used for the noise reduction processing in the adaptive filter 218 are the sound pickup signals determined by the signal determination unit 216. The adaptive filter 218 artificially generates a noise component that may be included in the audio signal using the reference signal in order to reduce the noise component included in the audio signal. The adaptive filter 218 can perform noise reduction processing by subtracting the pseudo-generated noise component from the audio signal.

  The adaptive filter control unit 217 included in the noise reduction device 4 according to the present embodiment has the same configuration as the adaptive filter control unit 117 included in the noise reduction device 3 according to the second embodiment described in FIG. The explanations made are omitted. The adaptive filter 218 included in the noise reduction device 4 according to the present embodiment has the same configuration as the adaptive filter 118 included in the noise reduction device 3 according to the second embodiment described with reference to FIGS. 11 and 15. A duplicate description is omitted.

  Next, operation | movement of the noise reduction apparatus 4 concerning this Embodiment is demonstrated. FIG. 18 is a flowchart for explaining the operation of the noise reduction apparatus 4 according to the present embodiment.

  The collected sound signal selection information 225 and the phase difference information 226 generated by the signal determination unit 216 are updated when it is certain that the voice section is in effect. Therefore, the signal determination unit 216 initializes the collected sound signal selection information 225 and the phase difference information 226 in advance and sets them to predetermined initial values (step S61). Here, the initial value is a parameter that is set when, for example, a device including a noise reduction device is used in an appropriate state (when the position of the microphone is used in an appropriate state).

  Next, the signal determination unit 216 acquires the collected sound signal selection information 223 and 225 and the phase difference information 226 using the collected sound signals 211 to 213 (step S62). Then, the signal determination unit 216 outputs the collected sound signal selection information 223 related to the collected sound signal used for the speech section determination to the speech section determiner 215. In addition, the signal determination unit 216 outputs the collected sound signal selection information 225 regarding the two collected sound signals used for the noise reduction processing and the phase difference information 226 of the determined two collected sound signals to the adaptive filter control unit 217. .

  Next, the speech segment determination unit 215 determines a speech segment using a sound collection signal corresponding to the sound collection signal selection information 223 (step S63). When the speech segment determination unit 215 detects a speech segment (step S64: Yes), the speech segment determination unit 215 outputs the speech segment information 224 to the adaptive filter control unit 217. Then, the adaptive filter control unit 217 updates the collected sound signal selection information and the phase difference information to the collected sound signal selection information 225 and the phase difference information 226 at the timing determined as the speech section by the speech section determination unit 215 (step S65). On the other hand, when it is determined by the speech segment determination unit 215 that the speech segment is not a speech segment (step S64: No), the adaptive filter control unit 217 does not update the collected sound signal selection information and the phase difference information.

  Next, in the selector of the adaptive filter 218 (corresponding to the selector 177 in FIG. 15), an audio signal and a reference signal are selected from the collected sound signals 211 to 213 based on the collected sound signal selection information 225 (step S66). Then, the adaptive filter 218 performs noise reduction processing using the selected two collected sound signals (step S67).

  The noise reduction device 4 checks whether or not sound (voice or noise) is received by the voice microphone 201 or the like (step S68). And when the sound is received (step S68: Yes), the process after step S62 is repeated. On the other hand, when the sound is not received (step S68: No), the noise reduction process by the noise reduction device 4 ends.

  In the noise reduction device 3 according to the second embodiment shown in FIG. 11, the collected sound signal 111 collected by the sound microphone 101 is used for sound section determination in the sound section determiner 115. In this case, it is preferable that the collected sound signal 111 collected by the sound microphone 101 mainly includes sound. For example, the sound microphone 101 and the speaker's mouth are in a stable state with a certain distance therebetween. It assumes the state used in. In this usage, the speech segment determination unit 115 may perform speech segment determination on the collected sound signal 111 collected by the speech microphone 101. In addition, the signal determination unit 116 only has to acquire the collected sound signal selection information 125 and the phase difference information 126 only when it is determined that the speech section is determined, and there is an advantage that the load of signal processing can be reduced. .

  Thus, in the noise reduction apparatus 3 according to the second embodiment shown in FIG. 11, it is assumed that the voice microphone 101 and the speaker's mouth are being used in a stable state with a certain distance therebetween. ing. However, in some devices using the noise reduction device, for example, depending on the use situation of the speaker, the distance between the voice microphone and the speaker's mouth may not be constant and may be used in an unstable state. is there. In this case, the reference sound microphone may be able to pick up more sound than the sound microphone.

  In the noise reduction device 4 according to the present embodiment, the signal determination unit 216 determines a sound collection signal to be used for sound segment determination in the sound segment determination unit 215 from the sound collection signals 211 to 213. Then, the speech segment determination unit 215 determines a speech segment using the collected sound signal determined by the signal determination unit 216. Further, the adaptive filter control unit 217 controls the adaptive filter 218 using the collected sound signal selection information 225 and the phase difference information 226 at the timing determined by the speech segment determination unit 215 as a speech segment. Therefore, even when the noise level is high, it is possible to accurately determine the section where the voice is emitted. In addition, since it is possible to select two optimum sound pickup signals used for noise reduction processing from a plurality of sound pickup signals, this is the case when a device using the noise reduction device is used in various states. However, the noise reduction process can be performed with high accuracy.

  As described above, according to the invention according to the present embodiment, a noise reduction device, a voice input device, a wireless communication device, and a noise reduction device that can appropriately reduce noise components included in a voice signal even under various environments, A noise reduction method can be provided.

<Embodiment 4>
Next, a fourth embodiment of the present invention will be described.
Below, the case where the noise reduction apparatus provided with at least three microphones is applied to a voice input device or a wireless communication device will be described. As the noise reduction device including at least three microphones, for example, the noise reduction device according to the second or third embodiment can be used.

  FIG. 19 is a diagram illustrating an example of a voice input device 700 using a noise reduction device including at least three microphones. FIG. 19A is a front view of the voice input device 700, and FIG. 19B is a rear view of the voice input device 700. As shown in FIG. 19, the voice input device 700 is configured to be connectable to a wireless communication device 710 via a connector 703. The wireless communication device 710 is 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 710 via the voice input device 700.

  The voice input device 700 includes a main body 701, a cord 702, and a connector 703. The main body 701 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. 19A, a speaker 706 and an audio microphone 705 are provided on the front surface of the main body 701. As shown in FIG. 19B, reference sound microphones 711 and 712 and a belt clip 707 are provided on the back surface of the main body 701. An LED 709 is provided on the top surface of the main body 701. A PTT (Push To Talk) 704 is provided on a side surface of the main body 701. The LED 709 notifies the speaker of the detection state of the speaker's voice by the voice input device 700. The PTT 704 is a switch for setting the wireless communication device 710 to a voice transmission state, and detects that the protruding portion is pushed into the housing.

  For example, when the noise reduction device 3 according to the second exemplary embodiment illustrated in FIG. 11 is applied to the voice input device 700, the voice microphone 101 included in the noise reduction device 3 corresponds to the voice microphone 705 of the voice input device 700. The two reference sound microphones 102 and 103 included in the noise reduction device correspond to the reference sound microphones 711 and 712 of the sound input device 700. The output signal 128 output from the noise reduction device 3 is supplied to the wireless communication device 710 via the code 702 of the voice input device 700. That is, the voice input device 700 supplies the output signal 128 after the noise reduction processing by the noise reduction device 3 to the wireless communication device 710. Therefore, the sound transmitted from the wireless communication apparatus 710 to another wireless communication apparatus is a sound subjected to noise reduction processing. Note that the same applies to the case where the noise reduction device 4 according to the third embodiment shown in FIG.

  In the audio input device 700 according to the present embodiment, an audio microphone (first microphone) 705 is provided on the surface (first surface). FIG. 20 is a diagram for explaining the details of the positions of the reference sound microphones 711 and 712 provided on the back surface of the voice input device 700 according to the present embodiment. As shown in FIG. 20, in the audio input device 700 according to the present embodiment, the reference sound microphones (second and third microphones) 711 and 712 are separated from the surface (first surface) by a predetermined distance. Are provided so as to be asymmetric with respect to the center line 721 of the back surface. At this time, the reference sound microphones 711 and 712 are separated from each other by a distance d1. For example, d1 can be about 3 to 7 cm. The distance between the front surface and the back surface can be about 2 to 4 cm. These numerical values are examples, and the present invention is not limited to these numerical values.

  As described above, in the voice input device 700 according to the present embodiment, the reference sound microphones 711 and 712 are disposed so as to be asymmetric with respect to the center line 721 on the back surface. It is possible to prevent both of the reference sound microphones 711 and 712 from being blocked when the 700 is held. Therefore, at least one of the reference sound microphones 711 and 712 can be used for noise reduction processing with high probability. Therefore, noise can be reduced with high accuracy using the noise reduction device.

  At this time, the reference sound microphones 711 and 712 may be provided such that a line segment 722 connecting the reference sound microphones 711 and 712 and the center line 721 intersect at a predetermined angle α. The predetermined angle α is, for example, a tan α when a maximum rectangle that can enter the back surface is drawn on the back surface of the audio input device 700 in which the reference sound microphones 711 and 712 are arranged, and the side of the rectangle is a × b. = You may set to the value which satisfies a / b. That is, if the shape of the back surface of the voice input device 700 is a regular square, the predetermined angle α is about 45 degrees. The predetermined angle α is made smaller as the shape of the back surface of the voice input device 700 is longer.

The reference sound microphones 711 and 712 include two line segments 731 and 732 perpendicular to the center line 721, and two line segments 733 that are parallel to the center line 721 and symmetrical with respect to the center line 721. , 734 may be provided at diagonal positions of a rectangle 735. Thus, by arranging the reference sound microphones 711 and 712 diagonally, it is possible to select a reference sound signal that works well against noise sources from various directions.

  Next, a wireless communication device (transceiver) 800 using a noise reduction device including at least three microphones will be described with reference to FIG. FIG. 21A is a front view of the wireless communication apparatus 800, and FIG. 21B is a rear view of the wireless communication apparatus 800. As shown in FIG. 21, the wireless communication apparatus 800 includes an input button 801, a display unit 802, a speaker 803, an audio microphone 804, a PTT (Push To Talk) 805, a switch 806, an antenna 807, a lid 809, and a reference sound. Microphones 811 and 812 are provided.

  For example, when the noise reduction device 3 according to the second exemplary embodiment illustrated in FIG. 11 is applied to the wireless communication device 800, the voice microphone 101 included in the noise reduction device 3 corresponds to the voice microphone 804 of the wireless communication device 800. The reference sound microphones 102 and 103 included in the noise reduction device 3 correspond to the reference sound microphones 811 and 812 of the wireless communication device 800. The output signal 128 output from the noise reduction device 3 is subjected to high-frequency processing in an internal circuit of the wireless communication device 800 and is wirelessly transmitted from the antenna 807 to another wireless communication device. Here, since the output signal 128 output from the noise reduction device 3 is a signal on which noise reduction processing has been performed, the sound transmitted to the other wireless communication device is the sound subjected to noise reduction processing. The same applies to the case where the noise reduction device 4 according to the third exemplary embodiment illustrated in FIG.

In radio communication apparatus 800 according to the present embodiment, voice microphone (first microphone) 804 is provided on the front surface (first surface). Further, the reference sound microphones (second and third microphones) 811 and 812 are centered on the back surface (second surface) facing the front surface (first surface) with a predetermined distance therebetween. They are provided so as to be asymmetric with respect to the line. At this time, the reference sound microphones 811 and 812 are separated from each other by a distance d2. For example, d2 can be about 3 to 7 cm.
The distance between the front surface and the back surface can be about 2 to 4 cm. These numerical values are examples, and the present invention is not limited to these numerical values. The arrangement of the reference sound microphones 811 and 812 is the same as that of the reference sound microphones 711 and 712 of the voice input device shown in FIG.

  As described above, in the wireless communication apparatus 800 according to the present embodiment, the reference sound microphones 811 and 812 are arranged so as to be asymmetric with respect to the center line on the back surface. It is possible to prevent both of the reference sound microphones 811 and 812 from being blocked when the hand is held. Therefore, at least one of the reference sound microphones 811 and 812 can be used for noise reduction processing with high probability. Therefore, noise can be reduced with high accuracy using the noise reduction device.

  At this time, the reference sound microphones 811 and 812 may be provided such that a line segment connecting the reference sound microphones 811 and 812 and the center line intersect at a predetermined angle. In addition, the reference sound microphones 811 and 812 have a rectangular shape formed by two line segments perpendicular to the center line and two line segments arranged parallel to the center line and symmetrically with respect to the center line. It may be provided at a diagonal position.

  In the voice input device including the typical noise reduction device according to the above-described embodiment, the first microphone may be provided on a first surface of the voice input device, and the second microphone is The second surface facing the first surface with a predetermined distance may be provided.

  In the wireless communication device including the typical noise reduction device according to the above-described embodiment, the first microphone may be provided on the first surface of the wireless communication device, and the second microphone is The second surface facing the first surface with a predetermined distance may be provided.

  A typical noise reduction method according to the above-described embodiment determines a voice section based on voice collected by at least one of the first and second microphones, and determines that the voice section is a voice section. The direction of arrival of the sound is determined based on 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. Detection is performed, and noise reduction processing is performed based on the voice section information that is the determination result of the voice section and the voice direction information indicating the arrival direction of the voice.

In addition, another typical noise reduction device according to the above-described embodiment is based on phase difference information of a plurality of sound collection signals corresponding to sounds collected by a plurality of microphones, respectively. A first sound collection signal and a second sound collection signal used to reduce a noise component included in the first sound collection signal, and a signal determination unit And an adaptive filter that reduces a noise component contained in the first collected sound signal using the second collected sound signal.

  The noise reduction apparatus may further include a speech section determiner that determines a speech section based on one of the plurality of collected sound signals, and the signal determination unit includes the speech section determiner. May determine the first sound collection signal and the second sound collection signal from the plurality of sound collection signals.

  The noise reduction device may further include a speech segment determination unit that determines a speech segment using the first collected sound signal determined by the signal determination unit, and the adaptive filter includes the speech segment determination The noise component included in the first sound pickup signal determined by the signal determination unit when the device determines that it is a voice section may be reduced using the second sound pickup signal.

  The signal determination unit determines a sound pickup signal with the earliest phase among the plurality of sound pickup signals as the first sound pickup signal, and determines a sound pickup signal with the latest phase as the second sound pickup signal. May be.

  The signal determination unit may determine, as the second sound collection signal, a sound collection signal having a late phase and a power of the sound collection signal larger than a predetermined value among the plurality of sound collection signals.

  When the power of the collected sound signal having the slowest phase among the plurality of collected sound signals is equal to or less than a predetermined value, the signal determining unit is next delayed in phase and the power of the collected sound signal is larger than the predetermined value. A sound collection signal may be determined as the second sound collection signal.

  When the phase difference of each of the collected sound signals other than the first collected sound signal is within a predetermined range, the signal determination unit includes the collected sound signal among the collected sound signals other than the first collected sound signal. The collected sound signal having the largest power may be determined as the second collected sound signal.

  The plurality of microphones includes one voice microphone and a plurality of reference sound microphones, and the phase of the collected sound signal having the earliest phase among the plurality of sound collection signals respectively corresponding to the plurality of reference sound microphones. When the phase of the collected sound signal corresponding to the microphone for sound is earlier than the phase of the collected sound signal, the signal determining unit determines the collected sound signal having the earliest phase corresponding to the reference sound microphone as the first collected sound signal. May be.

  When the phase of the collected sound signal having the latest phase among the plurality of collected sound signals respectively corresponding to the plurality of reference sound microphones is slower than the phase of the collected sound signal corresponding to the sound microphone, the signal determining unit May determine the sound pickup signal with the latest phase corresponding to the reference sound microphone as the second sound pickup signal.

  The signal determination unit may be supplied with a signal having a sampling frequency of 24 kHz or more as the plurality of sound collection signals, and the adaptive filter is supplied with a signal having a sampling frequency of 12 kHz or less as the plurality of sound collection signals. May be.

  In the voice input device including the other typical noise reduction device according to the above-described embodiment, the first microphone of the plurality of microphones is provided on the first surface of the voice input device. The second and third microphones of the plurality of microphones may be arranged on a second surface facing the first surface with a predetermined distance from a center line of the second surface. It may be provided so as to be asymmetrical.

  In the wireless communication device including the other typical noise reduction device according to the above-described embodiment, the first microphone of the plurality of microphones may be provided on a first surface of the wireless communication device. The second and third microphones of the plurality of microphones are arranged on a second surface facing the first surface at a predetermined distance with respect to a center line of the second surface. It may be provided so as to be asymmetric.

  Another typical noise reduction method according to the above-described embodiment is based on phase difference information of a plurality of collected sound signals corresponding to sounds collected by a plurality of microphones. A first sound collection signal and a second sound collection signal to be used for noise reduction processing are determined, and a noise component included in the determined first sound collection signal is determined using the second sound collection signal. To reduce.

  Another typical audio input device according to the above-described embodiment includes a noise reduction device, and the noise reduction device mainly collects the first microphone for collecting the audio component and the noise component. And the first microphone is provided on a first surface of the voice input device, and the second and third microphones are connected to the first surface and a predetermined surface. The second surfaces facing each other at a distance are provided so as to be asymmetric with respect to the center line of the second surface.

  In the other typical audio input device according to the above-described embodiment, in the second and third microphones, a line segment connecting the second and third microphones and the center line intersect at a predetermined angle. It may be provided as follows.

  In the other typical audio input device according to the above-described embodiment, the second and third microphones include two line segments perpendicular to the center line, and parallel to the center line and the center line. May be provided at a diagonal position of a rectangle formed by two line segments arranged symmetrically with respect to each other.

  Another typical wireless communication apparatus according to the above-described embodiment includes a noise reduction device, and the noise reduction device mainly collects a noise component and a first microphone for collecting the voice component. And the first microphone is provided on a first surface of the wireless communication device, and the second and third microphones are connected to the first surface and a predetermined surface. The second surfaces facing each other at a distance are provided so as to be asymmetric with respect to the center line of the second surface.

  In another typical wireless communication apparatus according to the above-described embodiment, in the second and third microphones, a line segment connecting the second and third microphones intersects with the center line at a predetermined angle. It may be provided as follows.

  In another typical wireless communication apparatus according to the above-described embodiment, the second and third microphones include two line segments perpendicular to the center line, parallel to the center line, and the center line. May be provided at a diagonal position of a rectangle formed by two line segments arranged symmetrically with respect to each other.

  According to the above-described embodiments, it is possible to provide a noise reduction device, a voice input device, a wireless communication device, and a noise reduction method that can appropriately reduce noise components included in a voice signal even under various environments. Is possible.

  The present invention has been described with reference to the above embodiment, but 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 present application. It goes without saying that various modifications, corrections, and combinations are included.

1, 2, 3, 4 Noise reduction device 11 Audio microphone 12 Reference sound microphone 13, 14 AD converter 15 Audio interval determination device 16 Audio direction detector 17 Adaptive filter control unit 18 Adaptive filter 21, 22 Sound collection signal 23, 24 Voice section information 25 Voice direction information 26 Control signal 27 Output signal 101 Voice microphones 102, 103 Reference sound microphones 104, 105, 106 AD converter 115 Voice section decision unit 116 Signal determination unit 117 Adaptive filter control unit 118 Adaptive filter 111 , 112, 113 Collected sound signal 123, 124 Sound section information 125 Collected signal selection information 126 Phase difference information 127 Control signal 128 Output signal 201 Sound microphone 202, 203 Reference sound microphone 204, 205, 206 AD control 215 Voice segment determination unit 216 Signal decision unit 217 Adaptive filter control unit 218 Adaptive filters 211, 212, 213 Sound collection signal 223 Sound collection signal selection information 224 Voice segment information 225 Sound collection signal selection information 226 Phase difference information 227 Control signal 228 output signal

Claims (13)

A speech segment determination unit that determines a first speech segment and a second speech segment based on speech collected by at least one of the first and second microphones;
The direction of arrival of the sound based on the first sound collection signal corresponding to the sound collected by the first microphone and the second sound collection signal corresponding to the sound collected by the second microphone A voice direction detector for detecting
Based on the first voice section information output from the voice section determiner and the voice direction information output from the voice direction detector, the first sound pickup signal and the second sound pickup signal are used. An adaptive filter that performs noise reduction processing,
The speech segment determiner outputs second speech segment information determined to be a speech segment with a higher probability than the first speech segment to the speech direction detector,
The voice direction detector detects the direction of arrival of the voice when the voice period determiner determines the second voice period.
Noise reduction device.   The noise reduction device according to claim 1, wherein the voice direction detector detects the direction of arrival of the voice based on a phase difference between the first sound pickup signal and the second sound pickup signal. The adaptive filter uses the other collected sound signal to reduce a noise component included in one of the first collected sound signal and the second collected sound signal that has an earlier phase. To
The noise reduction device according to claim 2. When the phase difference between the phase of the first sound pickup signal and the phase of the second sound pickup signal is within a predetermined range,
The adaptive filter outputs the first collected sound signal or the second collected sound signal without performing noise reduction processing;
The noise reduction device according to claim 2.   The noise reduction device according to claim 1, wherein the voice direction detector detects the direction of arrival of the voice based on a magnitude of the first collected sound signal and a magnitude of the second collected sound signal. When the magnitude of the first sound collection signal is larger than the magnitude of the second sound collection signal,
The adaptive filter uses the other collected sound signal as a noise component included in one of the first collected sound signal and the second collected sound signal having a larger magnitude. To reduce,
The noise reduction device according to claim 5. When the power difference, which is the difference between the magnitude of the first collected signal and the magnitude of the second collected signal, is within a predetermined range,
The adaptive filter outputs the first collected sound signal or the second collected sound signal without performing noise reduction processing;
The noise reduction device according to claim 5.   The sound direction detector is based on a phase difference between the first sound collection signal and the second sound collection signal, and a magnitude of the first sound collection signal and a magnitude of the second sound collection signal. The noise reduction device according to claim 1, wherein an arrival direction of the voice is detected. When the phase of the first sound collection signal is earlier than the phase of the second sound collection signal, the speech segment determination unit determines a speech segment based on the first sound collection signal,
If the phase of the second collected sound signal is earlier than the phase of the first collected sound signal, the speech segment determiner determines a speech segment based on the second collected signal;
The noise reduction device according to any one of claims 1 to 8. The voice direction detector is supplied with a signal having a sampling frequency of 24 kHz or more as the first and second collected sound signals.
A signal having a sampling frequency of 12 kHz or less is supplied to the adaptive filter as the first and second collected sound signals.
The noise reduction device according to any one of claims 1 to 9. A voice input device having a noise reduction apparatus according to any one of claims 1 to 10,
The first microphone is provided on a first surface of the voice input device;
The second microphone is provided on a second surface facing the first surface at a predetermined distance,
Voice input device. A wireless communication apparatus having a noise reduction apparatus according to any one of claims 1 to 10,
The first microphone is provided on a first surface of the wireless communication device;
The second microphone is provided on a second surface facing the first surface at a predetermined distance,
Wireless communication device. A first voice section determination step of determining the first speech section based on sound collected by at least one of the first and second microphones,
A second speech segment determination step of determining a second speech segment that is a speech segment with a higher probability than the first speech segment;
When it is determined that it is the second voice section, the first sound collection signal corresponding to the sound collected by the first microphone and the sound collected by the second microphone are used. A voice direction detecting step of detecting a direction of arrival of the voice based on a second collected sound signal;
Based on the first voice segment information which is the determination result of the first voice segment and the voice direction information indicating the direction of arrival of the voice , the first sound pickup signal and the second sound pickup signal are used. A noise reduction step for performing noise reduction processing ;
Including a noise reduction method.