JP2015025914A - Voice signal processor and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice signal processor capable of suppressing the occurrence of musical noise even if noise components are suppressed according to a coherence filter method.SOLUTION: The present invention is related to a voice signal processor that suppresses noise components included in input voice signals by coherence filter processing. The processor includes: means for calculating a coherence filter coefficient; and means for smoothing the calculated coherence filter coefficient in a frequency area and then making it apply to the coherence filter processing.

Description

  The present invention relates to an audio signal processing apparatus and program, and handles, for example, an audio signal from a telephone or a video conference apparatus (in this specification, an audio signal such as an audio signal or an acoustic signal is called an “audio signal”). It can be applied to communication devices and communication software.

  One of the methods for suppressing the noise component contained in the acquired audio signal is a coherence filter method. As described in Patent Document 1, the coherence filter method is a method of suppressing a noise component having a large bias in the arrival direction by multiplying the cross-correlation of signals having blind spots on the left and right for each frequency.

JP 2013-061421 A

  However, the coherence filter method has an effect of suppressing a noise component, but on the other hand, there is a problem that an abnormal sound component (tone noise) called musical noise is generated and the naturalness of sound is impaired.

  Therefore, there is a demand for an audio signal processing apparatus and program that can suppress the occurrence of musical noise even if the noise component is suppressed according to the coherence filter method.

  According to a first aspect of the present invention, in a speech signal processing apparatus that suppresses a noise component contained in an input speech signal by coherence filter processing, (1) coherence filter coefficient calculation means for calculating a coherence filter coefficient, and (2) calculation Coefficient smoothing means for smoothing the above-described coherence filter coefficient on the frequency domain and then applying it to the coherence filter processing is provided.

  The audio signal processing program according to the second aspect of the present invention provides a computer mounted on an audio signal processing device that suppresses noise components contained in an input audio signal by coherence filter processing. (1) Coherence for calculating coherence filter coefficients And (2) smoothing the calculated coherence filter coefficient in the frequency domain, and then functioning as a coefficient smoothing means to be applied to the coherence filter processing.

  According to the present invention, since the obtained coherence filter coefficient is smoothed in the frequency domain and then used for the coherence filter processing, even if the noise component is suppressed according to the coherence filter method, the generation of musical noise is prevented. An audio signal processing device and a program that can be suppressed can be provided.

It is a block diagram which shows the whole structure of the audio | voice signal processing apparatus of 1st Embodiment. It is a block diagram which shows the detailed structure of the coherence filter process part in 1st Embodiment. It is explanatory drawing which shows the property of the directivity signal from the directivity formation part in 1st Embodiment. It is explanatory drawing which shows the characteristic of two directivities by the directivity formation part in 1st Embodiment. It is a flowchart which shows detailed operation | movement of the coherence filter process part in 1st Embodiment. It is a block diagram which shows the detailed structure of the coherence filter process part in 2nd Embodiment. It is explanatory drawing which shows the behavior of the coherence for every azimuth | direction. It is explanatory drawing which shows a response | compatibility (conversion table) with the coherence which the averaging parameter memory | storage part in 2nd Embodiment has memorize | stored, and an averaging parameter. It is explanatory drawing which shows the directivity of the noise signal produced | generated in 3rd Embodiment.

(A) First Embodiment Hereinafter, a first embodiment of an audio signal processing device and a program according to the present invention will be described in detail with reference to the drawings.

  When noise components are suppressed according to the coherence filter method, the cause of musical noise is that isolated points in the frequency domain are caused by a particular frequency component becoming significantly larger or smaller due to the addition of coherence filter coefficients. The inventor of the present application has recognized that this occurs.

  The audio signal processing apparatus and program according to the first embodiment suppress the occurrence of isolated points in the frequency domain by smoothing the coherence filter coefficients using the coherence filter coefficients of the adjacent frequency components, and reduce musical noise. I tried to reduce it.

(A-1) Configuration of First Embodiment FIG. 1 is a block diagram showing a configuration of an audio signal processing device according to the first embodiment. Here, the part excluding the pair of microphones m1 and m2 can be configured by hardware, and can also be realized by software (audio signal processing program) executed by the CPU and the CPU. However, even if any realization method is adopted, it can be functionally represented in FIG.

  In FIG. 1, an audio signal processing apparatus 10 according to the first embodiment includes a pair of microphones m1 and m2, an FFT (Fast Fourier Transform) unit 11, a coherence filter processing unit 12, and an IFFT (Inverse Fast Fourier Transform) unit 13. Have.

  The pair of microphones m1 and m2 are arranged apart from each other by a predetermined distance (or an arbitrary distance), and each captures surrounding sounds. Each of the microphones m1 and m2 is omnidirectional (or has a very gentle directivity in the front direction). Audio signals (input signals) captured by the respective microphones m1 and m2 are converted into digital signals s1 (n) and s2 (n) via corresponding A / D converters (not shown) and given to the FFT unit 11. . Note that n is an index indicating the input order of samples, and is expressed as a positive integer. In the text, it is assumed that the smaller n is the older input sample, and the larger n is the newer input sample.

The FFT unit 11 receives input signal sequences s1 (n) and s2 (n) from the microphones m1 and m2, and performs fast Fourier transform (or discrete Fourier transform) on the input signals s1 and s2. Thereby, the input signals s1 and s2 can be expressed in the frequency domain. In performing the Fast Fourier Transform, analysis frames FRAME1 (K) and FRAME2 (K) composed of predetermined N samples are configured and applied from the input signals s1 (n) and s2 (n). An example of constructing the analysis frame FRAME1 (K) from the input signal s1 (n) is shown in the following equation (1), and the analysis frame FRAME2 (K) is the same.

  K is an index indicating the order of frames and is expressed by a positive integer. In the text, it is assumed that the smaller the K, the older the analysis frame, and the larger, the newer the analysis frame. In the following description, it is assumed that the index representing the latest analysis frame to be analyzed is K unless otherwise specified.

  The FFT unit 11 converts the frequency domain signals X1 (f, K) and X2 (f, K) into the frequency domain signals X1 (f, K) by performing a fast Fourier transform process for each analysis frame. And X2 (f, K) are supplied to the coherence filter processing unit 12, respectively. Note that f is an index representing a frequency. X1 (f, K) is not a single value, but is composed of spectral components of a plurality of frequencies f1 to fm, as shown in equation (2). Furthermore, X1 (f, K) is a complex number and consists of a real part and an imaginary part. The same applies to X2 (f, K) and later-described B1 (f, K) and B2 (f, K).

X1 (f, K) = {X1 (f1, K), X1 (f2, K),..., X1 (fm, K)} (2)
In the coherence filter processing unit 12 to be described later, the frequency domain signal X1 (f, K) of the frequency domain signals X1 (f, K) and X2 (f, K) is mainly used, and the frequency domain signal X2 (f, K) is used. However, the processing may be performed with the frequency domain signal X2 (f, K) as the main and the frequency domain signal X1 (f, K) as the sub (see equation (8) described later).

  The coherence filter processing unit 12 has a detailed configuration shown in FIG. 2 to be described later. The coherence filter processing unit 12 performs coherence filter processing, obtains a signal Y (f, K) in which noise components are suppressed, and supplies the signal Y (f, K) to the IFFT unit 13. is there.

  The IFFT unit 13 performs an inverse fast Fourier transform on the noise-suppressed signal Y (f, K) to obtain an output signal y (n) that is a time domain signal.

  FIG. 2 is a block diagram illustrating a detailed configuration of the coherence filter processing unit 12.

  In FIG. 2, the coherence filter processing unit 12 includes an input signal receiving unit 21, a directivity forming unit 22, a filter coefficient calculation unit 23, a filter coefficient smoothing processing unit 24, a filter processing unit 25, and a post-filter processing signal transmission unit 26. .

  In the coherence filter processing unit 12, these units 21 to 26 operate in cooperation to execute processing shown in a flowchart of FIG. 5 described later.

  The input signal receiving unit 21 receives the frequency domain signals X1 (f, K) and X2 (f, K) output from the FFT unit 11.

The directivity forming unit 22 forms two types of directivity signals (first and second directivity signals) B1 (f, K) and B2 (f, K) having strong directivity in a specific direction. . As a method of forming the directivity signals B1 (f, K) and B2 (f, K), an existing method can be applied. For example, a method of obtaining by calculation according to the equations (3) and (4). Can be applied.

  Hereinafter, the meaning of the calculation formulas of the first and second directional signals B1 (f, K) and B2 (f, K) will be described with reference to FIGS. It is assumed that a sound wave arrives from the direction θ shown in FIG. 3A and is captured by a pair of microphones m1 and m2 that are separated by a distance l. At this time, there is a time difference until the sound wave reaches the pair of microphones m1 and m2. This arrival time difference τ is given by equation (5), where d = 1 × sin θ, where d is the sound path difference, and c is the sound speed.

τ = 1 × sin θ / c (5)
Incidentally, a signal s1 (t−τ) obtained by delaying the input signal s1 (n) by τ is the same signal as the input signal s2 (t). Therefore, the signal y (t) = s2 (t) −s1 (t−τ) taking the difference between them is a signal from which the sound coming from the θ direction is removed. As a result, the pair of microphones (microphone array) m1 and m2 have directivity characteristics as shown in FIG.

  In the above, the calculation in the time domain has been described, but the same can be said if it is performed in the frequency domain. The equations in this case are the above-described equations (3) and (4). As an example, it is assumed that the arrival direction θ is ± 90 degrees. That is, the first directivity signal B1 (f) has strong directivity in the right direction as shown in FIG. 4A, and the second directivity signal B2 (f) is shown in FIG. As shown in the figure, it has a strong directivity in the left direction. In the following description, it is assumed that θ = ± 90 degrees. However, θ is not limited to ± 90 degrees.

The filter coefficient calculation unit 23 calculates a coherence filter coefficient coef (f, K) according to the equation (6) based on the first and second directivity signals B1 (f, K) and B2 (f, K). Is.

  The filter coefficient smoothing processing unit 24 performs smoothing to bring the filter coefficient value of each frequency in the coherence filter coefficient coef (f, K) closer to the filter coefficient value of a nearby frequency. The filter coefficient smoothing processing unit 24 performs smoothing by, for example, a weighted averaging process as shown in Equation (7). In the equation (7), fi represents the frequency (target frequency) to be processed now, and f (i-1) is the previous frequency in the frequency domain (of the previous frequency point in the FFT). Frequency). In the equation (7), α is a weighting coefficient that satisfies 0.0 <α <1.0.

ave_coef (fi, K) = α × coef (fi, K) + (1−α) × ave_coef (f (i−1), K) (7)
The calculation of the equation (7) is performed by calculating the coefficient coef (fi, K) at the frequency of interest fi and the smoothed coherence filter coefficient ave_coef (f (i−1) of the frequency components f1 to f (ii) smaller than the frequency of interest fi. ) And K). The smoothed coherence filter coefficient ave_coef (fi, K) obtained in this way also contributes to the coherence filter coefficient at a lower frequency, so that the generation of isolated points in the frequency domain can be suppressed.

  The calculation for smoothing performed by the filter coefficient smoothing processing unit 24 is not limited to the calculation of Expression (7), and other calculation expressions for smoothing may be applied. For example, a simple average or a weighted average of the coherence filter coefficient values of a plurality of nearby frequencies including the target frequency (applying non-averaged ones in the vicinity frequency) is applied. May be.

  The filter processing unit 25 applies the smoothed coherence filter coefficient ave_coef (f, K), performs the coherence filter process on the main frequency domain signal X1 (f, K) as shown in the equation (8), and generates noise. A signal after suppression (filtered signal) Y (f, K) is obtained. In addition, (8) Formula represents each calculation (multiplication process) of each frequency.

Y (f, K) = X1 (f, K) × ave_coef (f, K) (8)
Here, the physical meaning of the coherence filter process will be supplemented. The coherence filter coefficient coef (f, K) (same as the smoothed coherence filter coefficient ave_coef (f, K)) is a cross-correlation of signal components having blind spots on the left and right. It is a voice component arriving from the front with no bias, and when the correlation is small, the arrival azimuth is a component that is biased to the right or left. Therefore, multiplication by the coherence filter coefficient coef (f, K) can be said to be processing for suppressing a noise component coming from the side.

  The post-filter processing signal transmission unit 26 supplies the post-noise suppression signal Y (f, K) to the IFFT unit 13 at the subsequent stage. Further, the post-filter processing signal transmission unit 26 increases K by 1 and starts processing of the next frame.

(A-2) Operation of the First Embodiment Next, the operation of the audio signal processing device 10 of the first embodiment will be described in the order of overall operation and detailed operation in the coherence filter processing unit 12 with reference to the drawings. To do.

  Signals s1 (n) and s2 (n) input from the pair of microphones m1 and m2 are respectively converted from time domain to frequency domain signals X1 (f, K) and X2 (f, K) by the FFT unit 11. Is then provided to the coherence filter processing unit 12. Thus, the coherence filter processing unit 12 performs coherence filter processing, and the obtained noise-suppressed signal Y (f, K) is provided to the IFFT unit 13. In IFFT section 13, noise-suppressed signal Y (f, K), which is a frequency domain signal, is converted into time domain signal y (n) by inverse fast Fourier transform, and this time domain signal y (n) is output. Is done.

  Next, the detailed operation in the coherence filter processing unit 12 will be described with reference to the flowchart of FIG. FIG. 5 shows the processing of a certain frame, and the processing shown in FIG. 5 is repeated for each frame.

  When it becomes a new frame and the frequency domain signals X1 (f, K) and X2 (f, K) of the new frame (current frame K) are given from the FFT unit 11, according to the equations (3) and (4) , First and second directional signals B1 (f, K) and B2 (f, K) are calculated (step S1), and these directional signals B1 (f, K) and B2 (f, K) are calculated. ), The coherence filter coefficient coef (f, K) is calculated according to the equation (6) (step S2).

  Then, for each frequency (frequency component) fi of the coherence filter coefficient coef (f, K), a smoothing process of the coherence filter coefficient coef (f, K) as shown in the equation (7) is executed, and the post-smoothing coherence filter A coefficient ave_coef (f, K) is obtained (step S3).

  The obtained smoothed coherence filter coefficient ave_coef (f, K) is applied, and the coherence filter process is executed on the main frequency domain signal X1 (f, K) as shown in the equation (8). The noise-suppressed signal (filtered signal) Y (f, K) is supplied to the IFFT unit 13 and the frame variable K is incremented by 1 (step S4), and the process proceeds to the next frame.

(A-3) Effect of First Embodiment According to the first embodiment, in the coherence filter processing, the smoothed coherence filter coefficient obtained by smoothing the coherence filter coefficient is applied instead of the coherence filter coefficient. Since it did in this way, generation | occurrence | production of the isolated point on the frequency domain which arises by the multiplication of a coherence filter coefficient can be prevented, and the musical noise which arises by a coherence filter process can be reduced.

  As a result, it is possible to expect improvement in call sound quality in a communication device such as a video conference device or a mobile phone to which the audio signal processing device or program of the first embodiment is applied.

(B) Second Embodiment Next, a second embodiment of the audio signal processing apparatus and program according to the present invention will be described in detail with reference to the drawings.

  The degree of occurrence of musical noise in the noise-suppressed signal varies depending on from which direction the noise comes. Therefore, in the second embodiment, the contribution of other frequency components in the smoothing process as shown in the equation (7) is controlled in accordance with the arrival direction of noise.

(B-1) Configuration of Second Embodiment The overall configuration of the audio signal processing apparatus according to the second embodiment can also be represented by the above-described FIG. 1 used in the description of the first embodiment. However, the internal configuration of the coherence filter processing unit (hereinafter referred to as reference numeral 12A) is different from that of the first embodiment.

  FIG. 6 is a block diagram showing a detailed configuration of the coherence filter processing unit 12A of the second embodiment, and the same reference numerals are given to the same and corresponding parts as in FIG.

  In FIG. 6, the coherence filter processing unit 12A of the second embodiment includes an input signal receiving unit 21, a directivity forming unit 22, a filter coefficient calculation unit 23, a filter coefficient smoothing processing unit 24, a filter processing unit 25, and a post-filter processing. In addition to the signal transmission unit 26, an arrival direction estimation unit 27, an averaging parameter determination unit 28, and an averaging parameter storage unit 29 are included.

  The input signal receiving unit 21, directivity forming unit 22, filter coefficient calculation unit 23, filter coefficient smoothing processing unit 24, filter processing unit 25, and post-filter signal transmission unit 26 are the same as those in the first embodiment. The functional description is omitted. Note that the filter coefficient smoothing processing unit 24 of the second embodiment is given from the averaging parameter determining unit 28 instead of applying the fixed averaging parameter α when performing the calculation of the equation (7). The point that the averaging parameter α (K) is applied is different from the filter coefficient smoothing processing unit of the first embodiment.

The arrival direction estimation unit 27 obtains an index value that can estimate the arrival direction of noise and gives it to the averaging parameter determination unit 28. Here, the arrival direction estimation unit 27 calculates coherence COH (K) as an index value that can estimate the arrival direction of noise. Coherence COH (K) is a value obtained by arithmetically averaging the coherence filter coefficients coef (f, K) at all frequencies, as shown in the equation (9).

  FIG. 7 is an explanatory diagram showing the behavior of coherence. As shown in FIG. 7, it can be seen that the range taken by the coherence value changes according to the arrival direction of noise. By using this property, the arrival direction of noise can be estimated.

  The averaging parameter determination unit 28 refers to the averaging parameter storage unit 29 based on the coherence COH (K) calculated by the arrival direction estimation unit 27 and uses the averaging parameter α (K) used in the filter coefficient smoothing processing unit 24. Is to determine.

  Since the isolated points in the frequency domain of the coherence filter coefficient tend to increase as the arrival direction of noise (such as disturbing speech) approaches the front, it is desirable to perform smoothing processing with more nearby frequency components. Therefore, when the arrival direction is close to the front (in other words, when the coherence COH (K) is large), the averaging parameter α is decreased to increase the contribution of other frequency components. In this case (in other words, when the coherence COH (K) is small), the control is performed to increase the averaging parameter α to reduce the contribution of other frequency components.

  The averaging parameter determination unit 28 determines an averaging parameter α (K) that can execute such control. The averaging parameter determination unit 28 may be of any specific configuration as long as it can determine the averaging parameter α (K) that can execute such control. For example, the averaging parameter determination unit 28 may determine the averaging parameter α (K) using the conversion table, and executes the calculation of the conversion function to determine the averaging parameter α (K). It may be what you do. FIG. 6 shows a configuration in the former case, and an averaging parameter storage unit 29 is provided.

  As shown in FIG. 8, the averaging parameter storage unit 29 applies the averaging parameter α applied when the range of the coherence COH (K) and the value of the coherence COH (K) calculated within the range belong. The correspondence (conversion table) with (K) is stored.

  The averaging parameter determination unit 28 has a given coherence COH (K) in any range A to B, B to C, C to D in the conversion table, (A <B <C <D <. ) And the values β, γ, δ,... (Where β> γ> δ>...) Associated with the range to which they belong are used as the averaging parameter α (K). To give. For example, if the coherence COH (K) is a value in the range of B and less than C, the averaging parameter determination unit 28 gives the averaging parameter α (K) whose value is γ to the filter coefficient smoothing processing unit 24.

(B-2) Operation of Second Embodiment Next, the operation of the audio signal processing apparatus of the second embodiment will be described. Since the overall operation is the same as that of the first embodiment, the operation of the coherence filter processing unit 12A of the second embodiment will be described below.

  When it becomes a new frame and the frequency domain signals X1 (f, K) and X2 (f, K) of the new frame (current frame K) are given from the FFT unit 11, according to the equations (3) and (4) , First and second directional signals B1 (f, K) and B2 (f, K) are calculated, and based on these directional signals B1 (f, K) and B2 (f, K), The coherence filter coefficient coef (f, K) is calculated according to the equation (6).

  Thereafter, according to the equation (9), coherence COH (K) obtained by arithmetically averaging the coherence filter coefficients coef (f, K) at all frequencies is calculated, and an averaging parameter α ( K) is taken from the conversion table.

  Then, by applying the extracted averaging parameter α (K), for each frequency (frequency component) fi of the coherence filter coefficient coef (f, K), a coherence filter coefficient coef ( The smoothing process of f, K) is executed, and the post-smoothing coherence filter coefficient ave_coef (f, K) is obtained.

  The obtained smoothed coherence filter coefficient ave_coef (f, K) is applied, and the coherence filter process is executed on the main frequency domain signal X1 (f, K) as shown in the equation (8). A noise-suppressed signal (filtered signal) Y (f, K) is output from the coherence filter processing unit 12A.

(B-3) Effect of Second Embodiment According to the second embodiment, the averaging parameter to be applied is determined according to the noise arrival direction, and the smoothing process of the coherence filter coefficient is performed. It is possible to obtain a musical noise reduction effect that does not depend on the noise arrival direction.

  As a result, by applying the present invention to a communication device such as a video conference system or a mobile phone, it is possible to expect improvement in call sound quality.

  As a result, it is possible to expect improvement in call sound quality in a communication device such as a video conference device or a mobile phone to which the audio signal processing device or program of the second embodiment is applied.

(C) Third Embodiment Next, a third embodiment of the audio signal processing apparatus and program according to the present invention will be described.

  In the second embodiment, coherence COH is applied as an index value representing the arrival direction of noise (such as disturbing speech). In the third embodiment, the SN ratio SNR (K) is applied instead of the coherence COH (K) as an index value representing the arrival direction of noise.

  The overall configuration of the audio signal processing apparatus according to the third embodiment can also be represented by FIG. 1 used in the description of the first embodiment. The detailed configuration of the coherence filter processing unit (12A) of the third embodiment can also be represented by FIG. 6 used in the description of the second embodiment.

  However, as described above, the arrival direction estimation unit 27 calculates the SN ratio SNR (K) instead of the coherence COH (K), unlike the second embodiment. The averaging parameter determination unit 29 determines an averaging parameter based on the calculated SN ratio SNR (K).

  Hereinafter, the SN ratio SNR (K) calculation method executed by the arrival direction estimation unit 27 of the third embodiment will be described.

  The arrival direction estimating unit 27 calculates the noise signal N (f, K) according to the equation (10) based on the frequency domain signals X1 (f, K) and X2 (f, K). The calculation of equation (10) corresponds to a process of forming directivity having a blind spot on the front as shown in FIG. Therefore, only components coming from the left and right can be obtained. Since the target direction is assumed to be the front direction (for example, the target speaker is assumed to be in front), it can be said that the component coming from the side is noise.

N (f, K) = X1 (f, K) -X2 (f, K) (10)
Next, the arrival direction estimating unit 27, based on the main frequency domain signal X1 (f, K) and the noise signal N (f, K), the SN ratio SNR (K) in the current frame K according to the equation (11). The denominator of equation (11) is the level of the noise signal, and the numerator is the level of the target sound signal, assuming that the target sound comes from the front and the noise comes from the side (left and right). Therefore, the S / N ratio can be estimated by the equation (11), where η is a parameter that takes a value in the range of 0 <η <1.

  The SN ratio SNR (K) calculated as described above is applied as an index value representing the arrival direction of noise, and an averaging parameter corresponding to the arrival direction of noise is set in the same manner as in the second embodiment described above. decide.

  Also in the third embodiment, since the averaging parameter determined in accordance with the noise arrival direction is applied and the coherence filter coefficient is smoothed, the same effects as in the second embodiment can be obtained. Can do.

(D) Other Embodiments In the description of each of the above-described embodiments, various modified embodiments have been referred to. However, modified embodiments exemplified below can be given.

  In the second embodiment, coherence COH (K) is applied as an index value representing the arrival direction of disturbing speech, and in the third embodiment, the SN ratio SNR (K) is an index value representing the arrival direction of disturbing speech. However, as long as it represents the arrival direction of disturbing speech, other index values may be applied, or a plurality of index values may be applied simultaneously. For example, the averaging parameter α (K) may be controlled according to the combination of the range to which the coherence COH (K) belongs and the range to which the SN ratio SNR (K) belongs.

  The number of coherence COH (K) ranges in the conversion table referred to in the description of the second embodiment may be two or more, and is not limited to a predetermined number.

  In each of the above embodiments, the processing that was processed with the frequency domain signal may be performed with the time domain signal if possible, and conversely, the processing that was processed with the time domain signal is possible. In this case, processing may be performed using a frequency domain signal.

  In each of the above-described embodiments, the noise suppression technique is shown by applying the coherence filter method alone, but other noise suppression techniques (see Patent Document 1), for example, the voice switch method, the Wiener filter method, the frequency subtraction method, and the like. You may make it use together.

  In each of the above-described embodiments, the audio signal processing apparatus and the program that immediately process the signal captured by the pair of microphones are shown, but the audio signal to be processed of the present invention is not limited to this. For example, the present invention can be applied to processing a pair of audio signals read from a recording medium, and the present invention can also be applied to processing a pair of audio signals transmitted from the opposite device. Can be applied.

  DESCRIPTION OF SYMBOLS 10 ... Audio | voice signal processing apparatus, 11 ... FFT part, 12, 12A ... Coherence filter processing part, 13 ... IFFT part, m1, m2 ... Microphone, 21 ... Input signal receiving part, 22 ... Directivity formation part, 23 ... Filter coefficient Calculation unit, 24 ... Filter coefficient smoothing processing unit, 25 ... Filter processing unit, 26 ... Filtered signal transmission unit, 27 ... Arrival direction estimation unit, 28 ... Average parameter determination unit, 29 ... Average parameter storage unit

Claims (6)

In an audio signal processing apparatus that suppresses noise components included in an input audio signal by coherence filter processing,
Coherence filter coefficient calculating means for calculating a coherence filter coefficient;
An audio signal processing apparatus comprising: coefficient smoothing means for smoothing the calculated coherence filter coefficient on a frequency domain and applying the smoothed coefficient to a coherence filter process.   2. The audio signal processing according to claim 1, wherein the coefficient smoothing unit includes an average processing unit that smoothes a frequency component in the coherence filter coefficient by averaging the frequency component with an adjacent frequency component. apparatus.   The noise direction reflection value calculation unit for calculating the value reflecting the arrival direction of the noise component in the input voice signal and the averaging parameter indicating the reflection degree of the adjacent frequency component in the averaging reflected the calculated arrival direction. The audio signal processing apparatus according to claim 2, further comprising an averaging parameter determination unit that determines the value according to a value.   The audio signal according to claim 3, wherein the noise direction reflection value calculation unit calculates a coherence that is an arithmetic average value of all the frequencies of the coherence filter coefficient as a value reflecting the arrival direction of the noise component. Processing equipment.   The audio signal processing apparatus according to claim 3, wherein the noise direction reflection value calculation unit calculates an SN ratio of the input audio signal as a value reflecting the arrival direction of the noise component. A computer mounted on an audio signal processing device that suppresses noise components contained in the input audio signal by coherence filter processing,
Coherence filter coefficient calculating means for calculating a coherence filter coefficient;
An audio signal processing program that functions as a coefficient smoothing unit that smoothes the calculated coherence filter coefficient in a frequency domain and then applies the coefficient to a coherence filter process.

Images