JP2016090799A - Noise suppression device, and method and program for the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a noise suppression device which suppresses deterioration, voice distortion, and noise without necessity of a further device such as VAD (voice activity detection), and to provide a method and a program for the noise suppression device.SOLUTION: In a noise suppression device 100, an adaptive filter is successively updated by using an error signal and a second sound collection signal so as to minimize the error signal in an adaptive filter unit 110. When the absolute value of the error proportion, which is the proportion of the error signal to the second sound collection signal, is equal to or smaller than a predetermined threshold value, the adaptive filter is updated by a first update amount based on a monotonic increasing function of the error proportion. When the absolute value of the error proportion is greater than a predetermined threshold value, the adaptive filter is updated by a second update amount based on a monotonic increasing function of the error proportion, where the increment in the second update amount is smaller than the increment in the first update amount.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a noise suppression technique for suppressing a noise component included in a certain collected sound signal by using another collected sound signal. In particular, the present invention relates to a noise suppression technique that suppresses a noise component contained in one of a plurality of collected sound signals obtained from a plurality of microphones mounted on a mobile terminal.

  When sound is picked up by a microphone, it is an inevitable event that sound of the surrounding environment is picked up along with the sound. Therefore, when a sound including a target sound component and a noise component is collected by a microphone, techniques for removing or suppressing the noise component by some method have been studied so far.

  For example, conventionally, when noise suppression is performed using a microphone of a mobile terminal such as a smartphone, the spectral subtraction method has been generally used because of a small amount of calculation (see Non-Patent Document 1). In the spectral subtraction method, the noise power of a noise interval (that is, a time interval in which a voice (target sound) to be collected (target sound) is not included) is estimated from a signal collected by a close-talking microphone or the like. This is a technique for suppressing noise by subtracting, on the frequency spectrum, a noise component to be superimposed on a collected sound signal in a speech section (time section including the target sound) using the estimated noise power.

  In addition to the spectral subtraction method for monaural microphones, two microphones are mounted on the smartphone, microphone array processing is performed from the signals collected by the two microphones, and the signals collected by the sub-microphones arranged on the back are Noise suppression processing is performed in which noise suppression is performed by removing the component from the signal of the main microphone arranged so as to be positioned near the mouth during a call (see Non-Patent Document 2 and Non-Patent Document 3). . The premise for this processing method is that the characteristics of the two microphones are the same to some extent, the sub microphones only pick up noise, and the main microphone picks up both the target sound and noise.

  The noise source superimposed on the target sound exists at a position distant from the target sound source, and is more greatly affected by the transfer characteristics between the noise source and the microphone. In addition, since its characteristics are generally unknown, it is necessary to estimate it. Therefore, system identification by an adaptive filter is performed with the transmission process as an unknown system, and the target sound is extracted by subtracting the filter output obtained by multiplying the sound collection signal of the sub microphone by the adaptive filter from the sound collection signal of the main microphone.

  At this time, it is desirable that the distance between the two microphones is not too large and not too small. This is because, if the distance between the two microphones is too large, noise components having different characteristics will be collected, and the erroneous noise component will be subtracted in the simple spectral subtraction method. On the other hand, if the distance between the two microphones is too small, the correlation between the noise components of each microphone will increase, but the target sound component that should not be removed will be collected simultaneously with the noise component by the sub microphone. The premise that microphones only collect noise is broken. In other words, the spectral subtraction method using two microphones has the contradictory acoustic characteristics that the target sound must be picked up only by the main microphone while keeping the correlation while picking up the noise with the two microphones. Applied as an ideal. However, in reality, since the characteristics of the two microphones are the same, it is difficult to prevent the target sound that has been turned around from being collected by the sub microphone.

BOLL S. F., "Suppression of acoustic noise in speech using spectral subtraction.Acoustics", Speech and Signal Processing, 1979, IEEE Transactions on, Volume: 27, Issue: 2, pp.113-120. Jian Zhang et. Al. "A FAST TWO-MICROPHONE NOISE REDUCTION ALGORITHM BASED ON POWER LEVEL RATIO FOR MOBILE PHONE", Kowloon: Chinese Spoken Language Processing (ISCSLP), 2012, 8th International Symposium on, pp.206-209. Isao Nakanishi, “Knowledge Forest”, Group 1 (Signal / System)-Volume 9 (Digital Signal Processing) Chapter 3, Adaptive Signal Processing, [online], IEICE, “Knowledge Forest”, [2014 Search October 23], Internet <http://www.ieice-hbkb.org/files/01/01gun_09hen_03m.pdf>

  Using the two microphones attached to the smartphone and using the submicrophone pickup signal placed on the back to perform the spectral subtraction method to remove the noise component from the main microphone pickup signal, Even the target sound that has been picked up is removed from the picked-up signal of the main microphone, and there is a problem in that the target sound is deteriorated and is distorted as known as musical noise.

  Therefore, in order to suppress deterioration and distortion of speech, system identification is performed by an adaptive filter only in a noise interval that is a non-speech interval. By using the adaptive filter estimated in the noise section, it is possible to eliminate the noise while leaving the target sound.

  In the present invention, the process of promoting learning of the adaptive filter only in the non-speech period is realized by modifying the expression of the adaptive filter. An object of the present invention is to provide a noise suppression apparatus, a method and a program for suppressing noise by suppressing deterioration and distortion of a voice without requiring a new apparatus such as VAD (voice activity detection).

  In order to solve the above-described problem, according to one aspect of the present invention, the noise suppression device suppresses a noise component included in the first sound collection signal using the second sound collection signal. The noise suppression device performs filtering using an adaptive filter on the second collected sound signal and obtains a filtered signal, and a subtractor obtains a difference between the first collected signal and the filtered signal as an error signal. The adaptive filter unit sequentially updates the adaptive filter using the error signal and the second collected sound signal so that the error signal is minimized, and the ratio of the error signal to the second collected signal is When the absolute value of a certain error ratio is less than or equal to a predetermined threshold, the adaptive filter is updated with the first update amount based on a monotonically increasing function with respect to the error ratio, and when the absolute value of the error ratio is larger than the predetermined threshold, The adaptive filter is updated with the second update amount based on the monotonically increasing function whose increase amount is smaller than the first update amount with respect to the error ratio.

  In order to solve the above-described problem, according to another aspect of the present invention, a noise suppression method suppresses a noise component included in a first sound pickup signal using the second sound pickup signal. In the noise suppression method, the second collected sound signal is filtered using an adaptive filter, and an adaptive filter step for obtaining a filtered signal, and a subtracting step for obtaining a difference between the first collected signal and the filtered signal as an error signal. In the adaptive filter step, using the error signal and the second collected sound signal, the adaptive filter is sequentially updated so that the error signal is minimized, and the ratio of the error signal to the second collected signal is When the absolute value of a certain error ratio is less than or equal to a predetermined threshold, the adaptive filter is updated with the first update amount based on a monotonically increasing function with respect to the error ratio, and when the absolute value of the error ratio is larger than the predetermined threshold, The adaptive filter is updated with the second update amount based on the monotonically increasing function whose increase amount is smaller than the first update amount with respect to the error ratio.

  According to the present invention, there is an effect that it is possible to suppress deterioration and distortion of voice and suppress noise without newly requiring a device such as VAD (voice activity detection).

The functional block diagram of the noise suppression apparatus which concerns on 1st embodiment. The figure which shows the example of the processing flow of the noise suppression apparatus which concerns on 1st embodiment. FIG. 3A is a front view of the noise suppression device according to the first embodiment, and FIG. 3B is a rear view of the noise suppression device according to the first embodiment. The figure which shows the example of the limiting function f ((beta)). The figure which shows the example of the limiting function f ((beta)).

  Hereinafter, embodiments of the present invention will be described. In the drawings used for the following description, components having the same function and steps performing the same process are denoted by the same reference numerals, and redundant description is omitted. Further, the processing performed for each element of a vector or matrix is applied to all elements of the vector or matrix unless otherwise specified.

<Points of first embodiment>
The noise suppression device is equipped with a main microphone and a sub microphone. For example, the noise suppression device is a mobile terminal (small notebook personal computer, smart phone, tablet type terminal, etc.), and due to its characteristics, the housing is small and it is difficult to separate two microphones beyond a predetermined interval. Therefore, in the present embodiment, it is assumed that noise is collected with the same sound pressure (similar sound pressure) in the main microphone and the sub microphone. In the present embodiment, the sound pressure of the target sound included in the sound collection signal collected by the main microphone is larger than the sound pressure of the target sound contained in the sound collection signal collected by the sub microphone. A main microphone and a sub microphone are arranged on the mobile terminal. For example, the main microphone is arranged so that it is located in the vicinity of the mouth during a call, and the sub microphone is arranged so that the target sound is located farthest from the mouth during a call and is difficult to enter. For example, the main mylophone is arranged on the lower front surface and the bottom surface of the mobile terminal, and the sub microphone is arranged on the upper rear surface and the upper surface of the mobile terminal. In addition, the surface on which the caller touches when making a call with a mobile terminal is the front, the portion located on the mouth side when making a call is the lower portion, and the surface on the mouth side is the bottom surface. Based on this premise, the adaptive filter learning method is adjusted by using a limiting function that places different limits on the update amount of the adaptive filter in the speech and non-speech segments, and the stabilization of the adaptive filter and the target sound The point of the present invention is to suppress the deterioration of the sound and the distortion of the voice and to suppress the noise.

  In the present embodiment, noise suppression and speech enhancement are performed using an adaptive filter using two microphones attached to a mobile terminal.

If the arrangement is as described above, the relationship between the collected sound signals between the microphones is as follows. The target sound is picked up by the main microphone with a higher sound pressure than the sub microphone. In addition, noise is picked up with the same sound pressure in both the main microphone and the sub microphone. Using this property, noise is suppressed by an adaptive filter and the target sound is emphasized.
<Noise Suppression Device 100 according to First Embodiment>
FIG. 1 is a functional block diagram of a noise suppression apparatus 100 according to the first embodiment, and FIG. 2 shows a processing flow thereof. Note that the functional block diagram of FIG. 1 is for clearly showing the processing circuit, and the circuit configuration is actually built in the noise suppression apparatus 100.

  The noise suppression apparatus 100 includes a main microphone 101, a sub microphone 102, an adaptive filter unit 110, a subtraction unit 120, a filter design unit 130, and a spectrum filter unit 140.

<Main microphone 101 and sub microphone 102>
The main microphone 101 collects the target sound and noise and outputs the first sound collection signal d (n) (S101). The sub microphone 102 picks up the target sound and noise and outputs the second sound pickup signal x (n) (S102). Note that n is an index representing time. The positions of the main microphone 101 and the sub microphone 102 mounted on the noise suppression apparatus 100 are shown in FIG. 3A is a front view of the noise suppression device 100, and FIG. 3B is a rear view. For example, it is assumed that the main microphone 101 and the sub microphone 102 are omnidirectional microphones, and the frequency characteristics of the microphone sensitivity of the main microphone 101 and the sub microphone 102 are uniform. However, the present invention does not limit the characteristics of the main microphone 101 and the sub microphone 102.

  The main microphone 101 is arranged on the noise suppression device 100 so as to approach the user's mouth when the noise suppression device 100 is used as a transmission / reception device or a voice input device. The sub microphone 102 is arranged on the same housing as the main microphone 101 (that is, on the noise suppression device 100), and is highly ambient noise that is highly correlated with the ambient noise picked up by the main microphone 101 while being away from the arrangement position of the main microphone 101. It is arranged at the position to record. Further, the input hole of the sub microphone 102 is arranged so that the user does not block it with a hand. However, the user's voice is transmitted to the sub microphone 102 side through the space, transmitted through the vibration of the user's bones, muscles, and the case, or reflected by the surrounding acoustic environment. There is no denial.

<Adaptive filter unit 110>
The adaptive filter unit 110 receives the second output signal x (n) and the error signal e (n), and filters the second collected sound signal x ⁻ (n) using the adaptive filter h ⁻ (n) ( S110), the filtered signal h ^-H (n) x ^- calculated (n), and outputs. ^{However, h - (n) = [} h 0 (n), h 1 (n), ..., h M-1 (n)] T, x (n) = [x (n), x (n-1) , ..., x (n-M + 1)] ^T , ^T represents transpose, and ^H represents complex conjugate transpose. Adaptive filter h ^- (n) is for performing a convolution operation, has a length of the tap size M, the second voice collecting signal x ^- using (n) in the calculation.

Further, the adaptive filter unit 110 sequentially uses the error signal e (n) and the second sound pickup signal x (n) to sequentially adapt the adaptive filter h ⁻ (n to minimize the error signal e (n). ) And the absolute value of the ratio of the error signal e (n) to the second collected sound signal x (n) (hereinafter also referred to as “error ratio”) is a predetermined threshold value (in this embodiment, the threshold value is 1). ) in the following cases, the adaptation by the first updating amount based on monotonically increasing function with respect to error rate β filter h ^- update the (n), when the absolute value of the error ratio β is larger than a predetermined threshold value, the error rate accommodated by second updating amount based on monotonically increasing function weight increase is smaller than the first update amount relative to β filter h ^- updating (n).

The adaptive filter design method will be described below. The main microphone 101 collects a sound in which target sound and noise are mixed. This noise is removed using the second sound pickup signal x (n) picked up by the sub microphone 102 and the adaptive filter h ⁻ (n). In the present embodiment, a normalized LMS (NLMS: Normalized least mean square) method is used to update the adaptive filter (see cited document 3).

Adaptive filter h ^- (n), the first voice collecting signal d (n) and the filtered signal h ^-H (n) x ^- as (n) error signal which is a difference between the e where (n) is minimized filter Do the design.
e (n) = d (n ) -h -H (n) x - (n) (1)
Incidentally, the adaptive filter h ^- (n) performs sequentially updated. In normal NLMS, the update formula is as follows.

Here, || x (n) || is the norm of the second collected sound signal x (n), and the adaptation constant μ is a step size parameter that determines the update amount of the update equation. The adaptive coefficient μ takes a constant value regardless of the error signal e (n) during the system operation, and the value range is a real number of 0 <μ <2. This update formula is decomposed as follows.

Here, β represents the ratio (ratio) of the error signal e (n) to the norm of the second collected sound signal x ⁻ (n).

Adaptive filter h ^- in the state where the learning has converged to some extent in (n), from the positional relationship between the two microphones attached to a mobile terminal, the non-speech section, comparable noise components in the main microphone 101 and the sub-microphone 102 Sound can be collected with sound pressure. Therefore, the adaptive filter h ^- by filtering (n), the first voice collecting signal d (n) and the filtered signal ^{h -H (n) x - (} n) which is the difference between the error signal e (n) is small Therefore, the ratio of the error signal e (n) to the norm of the second collected sound signal x ⁻ (n) is also reduced, and −1 <β <1.

On the other hand, since the main microphone 101 is arranged near the mouth of the speaker, the sound pressure of the sound picked up by the main microphone 101 is increased in the voice in the voice section. Then, the error signal e (n) contains many target sound components emitted by the speaker, and the absolute value of the error signal e (n) is the norm of the second collected signal x ⁻ (n) || x (n) It is larger than ||, and β <-1, 1 <β. In the present embodiment, by using a limit function f (β) that is non-linear with respect to β, the amount of filter update in the speech interval is reduced.

The limiting function f (β) is a nonlinear function that takes a small value with | β | ≧ 1 (β ≦ −1, β ≧ 1). For example, it is represented by the following formula.

For example, when L = 5, the function shown in FIG. 4 is obtained. For example, the limiting function f (β) may be a sigmoid function expressed by the following equation.

For example, when L = 5, the function shown in FIG. 5 is obtained. 4 and 5, when the absolute value of the error rate β is 1 or less, the first update amount based on the monotonically increasing function for the error rate β is used, and when the absolute value of the error rate β is greater than 1. The second update amount based on a monotonically increasing function that is smaller than the first update amount with respect to the error rate β is used. Therefore, the update amount (first update amount) when the absolute value of the error rate β is 1 (threshold) or less (first update amount), and the update amount (second update amount) when the absolute value of the error rate β is greater than 1 (threshold value). Bigger than. Then, the equation (5), the adaptive filter h on the basis of the first update amount or the second update amount ^- updates (n).

  The function constraint condition is that the absolute value of β decreases with β = 1 and β = -1. In the noise section, both the main microphone 101 and the sub microphone 102 are observed with the same sound pressure. Therefore, if the adaptive filter does not suppress the main microphone signal at all, β is 1, and if it can be suppressed, | β | <1. Next, in the speech section, β> 1 since the error signal e (n) is observed larger than the second sound collection signal x (n). If the error signal e (n) is learned to be small in the speech section, the target speech is suppressed. In order to avoid this, the design of the limiting function is such that the update amount of the filter is reduced by taking a small value when β> 1.

In other words, the sound pressure that can be observed by the first collected signal d (n) collected by the main microphone 101 and the second collected signal x (n) collected by the sub microphone 102 in the voice section is large. Since there is a difference and the target sound component remains in the error signal e (n), the ratio of the error signal e (n) to the norm of the second collected signal x ⁻ (n) also increases. Therefore, | β |> 1. Conversely, the interval where | β |> 1 is highly likely to be a speech interval, and by limiting the interval where | β |> 1, the step size of the adaptive filter in the speech interval can be suppressed. it can. In other words, in the interval where | β |> 1, the value of f (β) can be made smaller than β. In addition, with this technique, the step size is not set to 0 at | β |> 1 (speech interval), so that the noise source moves in the vicinity of the sub-microphone 102 or the noise source moves greatly and the sub- When the transfer functions to the microphone 102 and the main microphone 101 change, it is possible to prevent the update of the filter from being stopped when | β |> 1. By using this limiting function f (β), it is possible to suppress the processing of the adaptive filter that proceeds in the direction of erasing the speech signal that is the target sound existing in the speech section. In addition, it is possible to obtain an effect of improving the stability of the filter by relaxing the learning of the filter in a speech section having a characteristic different from that of noise.

<Subtraction unit 120>
Subtraction unit 120, the first voice collecting signal d (n) and the filtered signal ^{h -H (n) x - (} n) receives and, the difference ^{d (n) -h -H (n} ) x - (n ) As an error signal e (n) (S120) and output.

<Filter design unit 130>
The filter design unit 130 receives the second collected sound signal x (n) and the error signal e (n), designs a filter G that suppresses the noise component that has not been erased by the subtraction unit 120 (S130), and outputs it.

Although there are various filter design methods, for example, the filter design is performed using a method using a noise removal technique based on PSD (power-spectrum density) estimation described in Reference Document 1.
(Reference 1) Kenta Niwa, Yusuke Hioka, Kazunori Kobayashi, Noriyoshi Kamado, “Implementation of a microphone array to improve speech recognition under noisy conditions”, Proc. Of the Acoustical Society of Japan, 2014, pp .717-718

  For example, the filter design unit 130 uses the second collected sound signal x (n) and the error signal e (n) as the frequency domain second collected signal X (ω, τ) and the frequency domain error signal. Convert to E (ω, τ). From the ratio of the error signal and the second collected signal, the frequency domain error signal E (ω, τ) when | E (ω, τ) | / | X (ω, τ) | (ω, τ) and the frequency domain error signal E (ω, τ) when | E (ω, τ) | / | X (ω, τ) | And At this time, the post filter G (ω) is designed based on the Wiener method according to the following equation.

Xs (ω) = E [| Es (ω, τ) | ² ], Xn (ω) = E [| En (ω, τ) ² |]. Here, ω is an index representing a frequency, τ is an index representing a frame, and E [] is an average value of the frame τ. For example, the spectrum may be calculated by converting a time domain signal into a frequency domain signal by fast Fourier transform (FFT).

<Spectral filter unit 140>
The spectrum filter unit 140 receives the error signal e (n) and the filter G, and filters the error signal e (n) using the filter G (S140). In order to suppress the unerased noise component included in the error signal e (n), the post filter G (ω) is multiplied.
Y (ω, τ) = G (ω) E (ω, τ) (9)
Finally, Y (ω, τ) is subjected to inverse fast Fourier transform (IFFT) to obtain an output signal y (n).

<Effect>
With such a configuration, it is possible to suppress deterioration and distortion of voice and suppress noise without requiring a new device such as VAD (voice activity detection). In the present embodiment, when noise suppression is performed using two microphones mounted on a smartphone, the speed of update of the adaptive filter is changed by a limiting function for each voice section and noise section. As a result, it is possible to create a filter that suppresses filter learning in the wrong direction in the speech section and eliminates only noise collected by two microphones with equivalent sound pressure. In addition, by relaxing the filter learning in the speech section, it is possible to prevent speech suppression and stabilize the filter.

<Modification>
In this embodiment, the filter update amount is limited when β <−1 and β> 1, but the filter update amount may be limited when β <−a and β> a. The value of a is a> 0. For example, equation (6) may be replaced with the following equation.

  In the present embodiment, the processing of the adaptive filter unit 110 and the subtraction unit 120 is performed in the time domain, but the processing may be performed in the frequency domain. For example, a frequency domain conversion unit (not shown) is provided, and the first sound collection signal d (n) and the second sound collection signal x (n) are frequency domain first sound collection signals D (ω, τ ) And the frequency domain second collected signal X (ω, τ).

  The adaptive filter unit 110 receives the frequency domain second output signal X (ω, τ) and the frequency domain error signal E (ω, τ), and applies the adaptive filter H ( Filtering is performed using (ω, τ) (S110), and a filtered signal H (ω, τ) X (ω, τ) is obtained and output.

  The adaptive filter unit 110 updates the filter according to the following equation.

When the frequency domain error signal E (ω, τ) contains many target sound components, the absolute value of the frequency domain error signal E (ω, τ) is the norm of the frequency domain second collected signal X (ω, τ) | A value larger than | X (ω, τ) ||, β <-1, 1 <β. Even in this modification, the adaptive filter unit 110 reduces the filter update amount in the speech section by using a non-linear limit function f (β) with respect to the error rate β, as in the first embodiment. be able to. In this description, the same limiting function f (β) is used for the non-linear limiting function f (β) in all frequency bands, but a different limiting function f (β, ω) is used for each frequency domain ω. You may comprise as follows.

  The subtractor 120 receives the frequency domain first collected signal D (ω, τ) and the filtered signal H (ω, τ) X (ω, τ), and the difference D (ω, τ) −H (ω , τ) X (ω, τ) is obtained as a frequency domain error signal E (ω, τ) (S120) and output. If the subsequent stage (filter design unit 130 and spectral filter unit 140) performs processing in the frequency domain, the frequency domain second output signal X (ω, τ) and the frequency domain error signal E (ω, τ) can be used as they are. If a time domain signal is used, it may be converted into a frequency domain signal and output to the subsequent stage.

  Also, the point of the present embodiment is to use a restriction function that places different restrictions on the update amount of the adaptive filter between the speech section and the non-speech section. Therefore, the noise suppression apparatus 100 does not necessarily include the main microphone 101, the sub microphone 102, the filter design unit 130, and the spectrum filter unit 140.

<Other variations>
The present invention is not limited to the above-described embodiments and modifications. For example, the various processes described above are not only executed in time series according to the description, but may also be executed in parallel or individually as required by the processing capability of the apparatus that executes the processes. In addition, it can change suitably in the range which does not deviate from the meaning of this invention.

<Program and recording medium>
In addition, various processing functions in each device described in the above embodiments and modifications may be realized by a computer. In that case, the processing contents of the functions that each device should have are described by a program. Then, by executing this program on a computer, various processing functions in each of the above devices are realized on the computer.

  The program describing the processing contents can be recorded on a computer-readable recording medium. As the computer-readable recording medium, for example, any recording medium such as a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory may be used.

  The program is distributed by selling, transferring, or lending a portable recording medium such as a DVD or CD-ROM in which the program is recorded. Further, the program may be distributed by storing the program in a storage device of the server computer and transferring the program from the server computer to another computer via a network.

  A computer that executes such a program first stores, for example, a program recorded on a portable recording medium or a program transferred from a server computer in its storage unit. When executing the process, this computer reads the program stored in its own storage unit and executes the process according to the read program. As another embodiment of this program, a computer may read a program directly from a portable recording medium and execute processing according to the program. Further, each time a program is transferred from the server computer to the computer, processing according to the received program may be executed sequentially. Also, the program is not transferred from the server computer to the computer, and the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes the processing function only by the execution instruction and result acquisition. It is good. Note that the program includes information provided for processing by the electronic computer and equivalent to the program (data that is not a direct command to the computer but has a property that defines the processing of the computer).

  In addition, although each device is configured by executing a predetermined program on a computer, at least a part of these processing contents may be realized by hardware.

Claims (7)

A noise suppression device that suppresses a noise component included in a first collected sound signal using a second collected sound signal,
Filtering the second collected sound signal using an adaptive filter and obtaining a filtered signal; and
A subtracting unit for obtaining a difference between the first collected sound signal and the filtered signal as an error signal;
The adaptive filter unit sequentially updates the adaptive filter using the error signal and the second sound pickup signal so as to minimize the error signal, and the error signal relative to the second sound pickup signal is updated. When the absolute value of the error ratio, which is a ratio, is less than or equal to a predetermined threshold, the adaptive filter is updated with a first update amount based on a monotonically increasing function with respect to the error ratio, and the absolute value of the error ratio is less than the predetermined threshold If larger, the adaptive filter is updated with a second update amount based on a monotonically increasing function whose increase amount is smaller than the first update amount with respect to the error ratio.
Noise suppression device. The noise suppression device of claim 1,
The first sound pickup signal is a signal picked up by a first microphone arranged to pick up a target sound,
The second sound pickup signal is a signal picked up by a second microphone arranged to pick up ambient noise correlated with the ambient noise included in the first sound pickup signal.
Noise suppression device. The noise suppression device of claim 1,
The first sound pickup signal is a signal picked up by a first microphone arranged at the mouth of the speaker, and is a signal picked up the target sound and ambient noise that are the utterance of the speaker,
The second sound collecting signal has a sound pressure of the target sound included in the second sound collecting signal that is lower than a sound pressure of the target sound included in the first sound collecting signal, and the second sound collecting signal. The sound collected by the second microphone arranged so that the sound pressure of the noise included in the signal is approximately the same as the sound pressure of the noise included in the first sound collection signal,
Noise suppression device. The noise suppression device according to any one of claims 1 to 3,
The predetermined threshold value is a> 0, the error rate is β, the first update amount or the second update amount is f (β), an index representing time is n, and the filter coefficient of the adaptive filter The filter length is M, the filter coefficient at time n is h ⁻ (n) = [h (n), h (n−1),..., H (n−M + 1)], and the adaptation constant is μ. , the second voice collecting signal and x (n) at time ^{n, x - (n) =} [x (n), x (n-1), ..., x (n-M + 1)] and the second collected signals x ^- the norm of (n) || x ^- a (n) ||, the error signal at time n and e (n), and the L and real number larger than 1, the update formula of the adaptive filter Is

And

Or

Is,
Noise suppression device. A noise suppression method for suppressing a noise component included in a first sound pickup signal using a second sound pickup signal,
An adaptive filter unit that filters the second collected sound signal using an adaptive filter and obtains a filtered signal; and
The subtracting unit includes a subtracting step for obtaining a difference between the first collected sound signal and the filtered signal as an error signal,
In the adaptive filter step, using the error signal and the second sound pickup signal, the adaptive filter is sequentially updated so that the error signal is minimized, and the error signal relative to the second sound pickup signal is updated. When the absolute value of the error ratio, which is a ratio, is less than or equal to a predetermined threshold, the adaptive filter is updated with a first update amount based on a monotonically increasing function with respect to the error ratio, and the absolute value of the error ratio is less than the predetermined threshold If larger, the adaptive filter is updated with a second update amount based on a monotonically increasing function whose increase amount is smaller than the first update amount with respect to the error ratio.
Noise suppression method. The noise suppression method according to claim 5, comprising:
The first sound pickup signal is a signal picked up by a first microphone arranged to pick up a target sound,
The second sound pickup signal is a signal picked up by a second microphone arranged to pick up ambient noise correlated with the ambient noise included in the first sound pickup signal.
Noise suppression method.   The program for functioning a computer as a noise suppression apparatus in any one of Claims 1-4.

Images