JP2019068133A - Sound pick-up device, program, and method - Google Patents

Abstract

To provide a sound pick-up device that can perform area sound pick-up with higher sound quality for output destinations.SOLUTION: A sound pick-up device 100 outputs sound pick-up signals picked up by using a plurality of microphone arrays MA1, MA2 to a plurality of output destinations. In order to output, to each of the output destinations, an acoustic signal including at least a component of a target area sound as the sound pick-up signal, the sound pick-up device generates a post-mixture target area sound in which a mixed sound including at least an input signal and/or a component of an estimated noise is mixed with the target area sound according to the characteristics of the output destination, and outputs the post-mixture target area sound as the sound pick-up signal.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a sound collection device, program, and method, for example, applied to a voice communication system or a speech recognition system used in a noise environment, to emphasize sounds in a specific area in the presence of multiple sound sources. It can apply to the system which suppresses the sound of other areas.

  When using a voice communication system or a voice recognition application system in a noise environment, ambient noise that mixes simultaneously with the required target voice is a nuisance that impairs good communication and causes a reduction in voice recognition rate. Conventionally, in an environment in which a plurality of such sound sources exist, a beam former using a microphone array is a technology for obtaining unnecessary target sound by separating and collecting only sounds in a specific direction and avoiding unnecessary sound mixing. (Beam Former; hereinafter also referred to as "BF"; see Patent Documents 2 and 3). BF is a technology for forming directivity using the time difference between signals arriving at each microphone. However, if there are other sound sources around the area for sound collection in BF alone (hereinafter referred to as "target area"), the sound present in the target area (hereinafter referred to as "target area sound") It is difficult to pick up only the sound. Therefore, in the related art, an area sound collection method for collecting a target area using a plurality of microphone arrays has been proposed by Patent Document 1 and the like.

  FIG. 6 is an explanatory view showing processing of collecting a target area sound from a sound source of a target area using two microphone arrays MA1 and MA2. FIG. 6A is an explanatory view showing a configuration example of each microphone array. FIGS. 6B and 6C are graphs (image views) showing the BF output of the microphone arrays MA1 and MA2 shown in FIG. 6A, respectively, in the frequency domain. FIGS. 6 (b) and 6 (c) are graphs (image views) showing the BF output of the microphone arrays MA1 and MA2, respectively, in the frequency domain.

  In the conventional area sound collection, as shown in FIG. 6A, the directivity of the microphone arrays MA1 and MA2 is collected in an area (target area) where sound collection is desired from different directions. In the state of FIG. 6A, not only the sound (target area sound) present in the target area but also the noise (non-target area sound) in the target area direction is included in the directivity of each of the microphone arrays MA1 and MA2. There is. However, as shown in FIGS. 6 (b) and 6 (c), when the directivity of the microphone arrays MA1 and MA2 is compared in the frequency domain, the target area sound component is included in either output, but the non-target area The sound components will be different for each microphone array. In the conventional area sound pickup technology, it is possible to extract only the target area sound by suppressing components other than the components included in common in the BF output of the two microphone arrays MA1 and MA2 using such characteristics.

  The conventional area sound collection technology is a very effective method to suppress the noise generated outside the area, but if the level of the non-target area sound or background noise present in the surroundings is large, offensive noise such as musical noise etc. It may occur. As a technique for improving musical noise in area sound collection, there is a technique described in Patent Document 3. In Patent Document 3, as a method of improving the sound quality of area pickup, a method of mixing an input signal and estimation noise with an area pickup output and masking abnormal noise such as musical noise (signal mixing area pickup method) Is proposed. In the following, as described in Patent Document 3, a method of masking abnormal noise such as musical noise by mixing a predetermined sound (for example, an input signal, estimated noise, etc.) with the output of the area pickup is described as “mixing We call it "area sound collection".

JP, 2014-72708, A JP 2005-195955 A Patent No. 6187626

Asano Ta, "Sound Technology Series 16 Array signal processing of sound-Localization, tracking and separation of sound source", Japan Acoustical Society, edited by Corona, February 25, 2011

  As described above, the conventional mixing area sound collection is a method to significantly improve the sound quality of the area sound collection, but since it is a method of mixing the input signal and the estimation noise, from the viewpoint of noise suppression, its effect Is slightly weakened. Therefore, when area pickup is applied to a system equipped with both voice communication and voice recognition functions, the noise suppression level is preferable for voice communication because the noise suppression effect is not sufficient as pre-processing for voice recognition, and the recognition rate decreases. Had the problem of

  In view of the above problems, there is a demand for a sound collection device, program, and method that enables area collection of higher quality to an output destination.

  According to a first aspect of the present invention, in a sound collection device that outputs a collected sound signal collected using a microphone array to a plurality of output destinations, (1) estimating background noise included in an input signal input from the microphone array Noise suppression means for acquiring the noise component as the estimation noise and using the acquired estimation noise to suppress the noise component of the input signal to acquire the noise-suppressed signal, and (2) a target area for the noise-suppressed signal Directivity forming means for acquiring a first non-target area sound in which directivity is formed in a direction other than the direction and a target area direction sound in which directivity is formed in the target area direction; (3) the target area direction sound To extract a second non-target area sound from the target area direction, and further using the second non-target area sound and the target area direction sound to generate a target area sound with the target area as a sound source Get An audio signal including at least the component of the target area sound at each of the output destinations of the target area sound extraction unit and (4), as the sound collection signal, according to the characteristics of the output destination An output unit capable of generating a target area sound after mixing the mixed sound including at least the component of the input signal and / or the component of the estimated noise in the target area sound and outputting it as the collected sound signal And.

  According to a second sound collecting program of the present invention, a computer mounted on a sound collection device for outputting a collected sound signal collected using a microphone array to a plurality of output destinations, (1) an input input from the microphone array Noise suppression means for estimating background noise included in a signal and acquiring it as estimation noise, and using the acquired estimation noise to suppress the noise component of the input signal to acquire a noise-suppressed signal, (2) Directivity forming means for acquiring, for the noise-suppressed signal, a first non-target area sound in which directivity is formed in a direction other than the target area direction and a target area direction sound in which directivity is formed in the target area direction; (3) A second non-target area sound from the target area direction is extracted using the target area direction sound, and further, using the second non-target area sound and the target area direction sound, the target The A target area sound extraction unit for acquiring a target area sound whose sound source is a sound source; and (4) outputting an audio signal including at least the component of the target area sound as the sound collection signal to each of the output destinations. The mixed sound including at least the component of the input signal and / or the component of the estimated noise is mixed with the target area sound according to the characteristics of the output destination to generate a target area sound after mixing, and the sound collection signal And output means capable of outputting as.

  According to a third aspect of the present invention, there is provided a sound collection method performed by a sound collection device for outputting a collected sound signal collected using a microphone array to a plurality of output destinations. (2) The noise suppression means estimates background noise contained in the input signal input from the microphone array and acquires it as estimation noise, and acquires it The noise component of the input signal is suppressed using the estimated noise to obtain a noise-suppressed signal, and (3) the directivity forming unit causes the noise-suppressed signal in a direction other than the target area. The first non-target area sound forming directivity and the target area directional sound forming directivity in the target area direction are acquired, and (4) the target area sound extraction unit uses the target area directional sound Target area direction The second non-target area sound is extracted, and further, using the second non-target area sound and the target area direction sound, the target area sound having the target area as a sound source is acquired, and (5) the above The output means outputs an acoustic signal including at least the component of the target area sound as the collected sound signal to each of the output destinations, and the target area sound is output according to the characteristics of the output destination. A mixed sound including at least a component of the input signal and / or a component of the estimation noise may be mixed to generate a target area sound after mixing and output as a sound collection signal.

  According to the present invention, it is possible to provide a sound collection device capable of collecting an area with higher sound quality to an output destination.

It is the block diagram shown about the functional composition of the sound collection device concerning a 1st embodiment. It is a block diagram showing composition of subtraction type BF (when there are two microphones) concerning a 1st embodiment. It is an explanatory view shown about an example of a directivity filter formed by subtraction type BF (when there are two microphones) concerning a 1st embodiment. It is the block diagram shown about the functional composition of the sound collection device concerning a 2nd embodiment. It is the block diagram shown about the functional composition of the sound collection device concerning a 3rd embodiment. It is explanatory drawing shown about the process which collects the target area sound from the sound source of a target area using two microphone arrays.

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

(A-1) Configuration of First Embodiment FIG. 1 is a block diagram showing a functional configuration of the sound collection device 100 of this embodiment.

  The sound collection device 100 performs target area sound collection processing for collecting a target area sound from a sound source of a target area using acoustic signals supplied from the two microphone arrays MA (MA1, MA2).

  The microphone arrays MA1 and MA2 are arranged anywhere in the space where the target area exists. For example, as shown in FIG. 6A, the positions of the microphone arrays MA1 and MA2 with respect to the target area may be anywhere as long as the directivity overlaps only in the target area, and may be arranged opposite to each other across the target area, for example. . In this embodiment, each microphone array MA comprises two or more microphones M, and each microphone M picks up an acoustic signal. As shown in FIG. 1, in this embodiment, two microphones M1 and M2 are described as being disposed in each microphone array MA. That is, each microphone array MA constitutes a 2ch microphone array. The number of microphone arrays MA is not limited to two. When there are a plurality of target areas, it is necessary to arrange the number of microphone arrays MA that can cover all the areas.

  As described above, each microphone array MA is disposed in a space where the target area is present, at a location where it can be directed to the target area. Each microphone array MA is configured by two microphones M (M1, M2). In each microphone array MA, an acoustic signal based on the sound captured by the two microphones M1 and M2 is supplied to the sound collection device 100.

  Then, the sound collection device 100 supplies the sound signal of the area collected to the voice recognition unit 10 and the speaker 11. The voice recognition unit 10 and the speaker 11 may be directly connected (locally disposed) to the sound collection device 100 or indirectly connected via a network to supply collected sound signals. May be

  The speaker 11 is a device that outputs, for example, a sound signal of an area collected at a remote place to an operator (user). In the following, it is assumed that the entire system for allowing a human to listen to a sound signal collected by the speaker 11 or the like is also referred to as a "communication system".

  The speech recognition unit 10 is, for example, a device accompanied by speech recognition processing such as converting the speech contained in the sound signal collected in an area at a remote location into a text. Hereinafter, the entire system that recognizes speech contained in an acoustic signal and performs processing based on the recognition result (for example, processing of Speech to Text, voiceprint recognition processing, etc.) will be referred to as "voice recognition system". .

  Therefore, the sound collection device 100 simultaneously collects the area of the plurality of output destinations having different characteristics (characteristics of the application), ie, the call system (speaker 11) and the speech recognition system (speech recognition unit 10). It will supply a signal. In other words, the sound collection device 100 picks up an acoustic signal (target sound emphasis signal) according to the characteristic of the output destination, and outputs each acoustic signal to the output destination of the corresponding characteristic.

  As described above, in this embodiment, the sound collection device 100 collects (generates) an audio signal for a call system and an audio signal for a voice recognition system, and an audio signal collected for the call system. Are supplied to the speaker 11, and the sound signal collected for the speech recognition system is supplied to the speech recognition unit 10. In this embodiment, as shown in FIG. 1, the output destination of the communication system is only the speaker 11, and the output destination of the voice recognition system is only the voice recognition unit 10. However, the output destination corresponding to each characteristic is It is good also as plural. Further, in this embodiment, the sound collection device 100 collects sound signals according to two characteristics of the communication system and the speech recognition system, but generates sound signals according to three or more types of characteristics, respectively. It may be output to the corresponding output destination.

  Next, the internal configuration of the sound collection device 100 will be described with reference to FIG.

  The sound collection device 100 includes a signal input unit 3, a noise suppression unit 4, a directivity formation unit 5, a target area sound extraction unit 6, a mixing level calculation unit 7, a mixing level adjustment unit 8, and a signal mixing unit 9. There is.

  For example, the sound collection device 100 may cause a computer including a processor, a memory, and the like to execute a program (including the sound collection program according to the embodiment), but even in that case, functionally, It can be shown as FIG.

(A-2) Operation of First Embodiment Next, an operation (a sound collection method according to the embodiment) of the sound collection device 100 of the first embodiment having the above-described configuration will be described.

  The signal input unit 3 converts an acoustic signal collected by each of the microphone arrays MA1 and MA2 from an analog signal into a digital signal and inputs it. Thereafter, the signal input unit 3 transforms the digital signal from the time domain to the frequency domain using, for example, a fast Fourier transform.

  The noise suppression unit 4 estimates and suppresses the component of the background noise included in the signal acquired by the signal input unit 3. The noise suppression unit 4 can perform noise suppression using, for example, a spectral subtraction (hereinafter, simply referred to as “SS”), a Wiener filtering method, or the like.

  The directivity forming unit 5 uses BF to form directivity in the direction of the target sound for an acoustic signal whose background noise has been suppressed by the noise suppressing unit 4 for each microphone array MA.

  Here, directivity formation by BF of each microphone array MA (MA1, MA2) will be described using FIG. 2 and FIG. BF is a technology for forming directivity by using the time difference between signals arriving at each microphone (see Non-Patent Document 1). Although BF is roughly divided into two types of addition and subtraction types, here, a subtraction type BF which can form directivity with a small number of microphones will be described.

  FIG. 2 is a block diagram showing a configuration according to the subtractive BF 200 when the number of microphones M is two.

  FIG. 3 is an explanatory view showing an example of a directional filter formed by the subtractive BF 200 using two microphones M. As shown in FIG.

  In the subtraction type BF 200 shown in FIG. 2, first, the delay unit 210 calculates the time difference of the signal that the sound (target sound) present in the target direction arrives at each microphone M (M1, M2) and adds a delay. Adjust the phase of the target sound by. The above-mentioned time difference can be calculated by the following equation (1).

Here, d is the distance between the microphones M1 and M2, c is the speed of sound, and τ _i is the delay amount. Further, θ _L is an angle from a perpendicular direction to a target direction with respect to a straight line connecting the microphones M (M 1, M 2).

Here, when a dead angle exists in the direction of the microphone M1 with respect to the centers of the microphones M1 and M2, the delay unit 210 performs delay processing on the input signal x ₁ (t) of the microphone M1. After that, in the subtraction type BF 200, the subtractor 220 performs processing (subtraction processing) according to the following equation (2).

The processing of the subtractive BF 200 can be similarly performed in the frequency domain, and in that case, the equation (2) is changed to the following equation (3).

Here, in the case of θ _L = ± π / 2, the directivity formed by the subtractive BF 200 becomes a cardioid unidirectivity as shown in FIG. 3A. Further, in the case of “θ _L = 0, π”, the directivity formed by the subtractive BF 200 is an eight-shaped bi-directionality as shown in FIG. 3B. Furthermore, by using the SS, the subtractive BF 200 can also form strong directivity in a bi-directional blind spot. The directivity by SS is formed at the entire frequency or a designated frequency band according to equation (4). (4) In the formula, is used to input signals X ₁ microphone M1, it is possible to obtain the same effect input signal X ₂ microphones M2. Here, β is a coefficient for adjusting the intensity of SS. In the subtraction type BF 200, when the value becomes minus at the time of subtraction, a flooring process is performed to replace the value with 0 or a smaller value of the original value.

In the processing method of the subtractive BF 200 as described above, the sound (non-target sound) existing other than the target direction is extracted according to the characteristic of bi-directionality, and the amplitude spectrum of the non-target sound extracted is subtracted from the amplitude spectrum of the input signal By doing this, the target sound can be emphasized.

  The directivity forming unit 5 can obtain the BF output of each of the microphone arrays MA (MA1, MA2) using the process of the subtraction type BF 200 as described above.

  By the way, when it is desired to pick up only the sound (target area sound) present in a specific area, the sound source (non-target area sound) existing on a line in the same direction as the area is collected only by using the subtractive BF. It sounds. Therefore, in the sound collection device 100, as proposed in Patent Document 1, by using a plurality of microphone arrays MA, directivity is directed to the target area from different directions respectively, and the directivity is crossed at the target area. A process of picking up a target area sound (area pickup process) is performed.

  The directivity forming unit 5 forms the directivity of each of the two microphone arrays of the microphone array MA1 and the microphone array MA2 by BF (subtractive BF 200), and each microphone array MA1,, as in FIG. Cross the directivity of MA2 in the area (target area) where you want to collect sound from different directions.

The target area sound extraction unit 6 sets each of the BF output data Y ₁ (n) and Y ₂ (n) of the microphone array MA1 and the microphone array MA2 formed by the directivity formation unit 5 to the following equation (5) or 6) SS according to the equation to extract non-target area sounds N ₁ (n) and N ₂ (n) present in the direction of the target area. Here, α ₁ and α ₂ are correction coefficients for correcting the difference in signal level caused by the difference in distance between the target area and each microphone array MA, and should be calculated one by one according to a predetermined process. However, for the sake of simplicity here, the distance between the target area and each microphone array MA is the same (α ₁ (n) = α ₂ (n) = 1), and equations (5) and (6) are equations (7) It demonstrates as what is applied instead of Formula (8).

Thereafter, the target area sound extraction unit 6 SS extracts non-target area sound from the BF output of the microphone arrays MA1 and MA2 according to the following equations (9) and (10) to extract a target area sound. Here, γ ₁ (n) and γ ₂ (n) are coefficients for changing the intensity at SS.

  The mixing level calculation unit 7 estimates the noise estimated by the noise suppression unit 4, the non-target area sound other than the target area direction extracted by the directivity forming unit 5, and the target area sound direction extracted by the target area sound extraction unit 6. The power of the non-target area sound is calculated, and the total volume level of the input signal to be mixed with the target area sound and the background noise is determined from the magnitude of the total value of them.

Here, it is assumed that the mixing level calculation unit 7 mainly performs area collection of the microphone array MA1 according to the equation (9). In this case, the mixing level calculation unit 7 estimates the noise B ₁ (n) estimated from the input signal of the microphone array MA ₁ and the non-target area sound M ₁ (n) other than the target area direction extracted according to the equation (3) When the sum of non-target area sounds N ₁ (n) in the direction of the target area extracted according to the equation (7) is A ₁ (n), the mixing level is δ ₁ A ₁ (n). Here, δ ₁ is a variable that is proportional to the SN ratio of the target area sound Z ₁ (n) and A ₁ (n), and is, for example, a value that makes A ₁ (n) −20 dB at an SN ratio of 0 dB.

The mixing level adjustment unit 8 determines the volume level (mixing ratio) of the input signal to be mixed with the target area sound and the estimation noise from the mixing level calculated by the mixing level calculation unit 7 and the ratio of the estimated noise to the power of the non-target area sound. Adjust. Here, it is assumed that the mixing level adjustment unit 8 mainly performs area collection of the microphone array MA1 according to the equation (9). In this case, variable lambda ₁ that determines the ratio of the input signal and the estimated noise to be mixed, the power ratio of the estimated noise _B 1 (n) and the non-target area sound _{(M 1 (n) + N} 1 (n)) (M 1 It is inversely proportional to (n) + N ₁ (n)) / B ₁ (n). For example, when (M ₁ (n) + N ₁ (n)) / B ₁ (n) = 0, λ ₁ = 1. Here, the range taken by λ ₁ is from 0 to 1. Furthermore, here, the variable μ ₁ for satisfying the mixture level δ ₁ A ₁ (n) is calculated by the following equation (11). Here, X ₁₁ (n) is an input signal acquired from the microphone M1 forming the microphone array MA1.

The signal mixing unit 9 calculates the ratio of the target area sound extracted by the target area sound extraction unit 6 to the input signal acquired by the signal input unit 3 and the noise estimated by the noise suppression unit 4 calculated by the mixing level adjustment unit 8 Mix based on For example, when the area of the microphone array MA1 is mainly picked up according to the equation (9), the final output W ₁ (n) is mixed according to the following equation (12). The data to be mixed into the target area sound may be mixed with the input sound and the estimated noise at a predetermined ratio as described above, or may be mixed with only the input sound or only the estimated noise.

  In the sound collection device 100, the sound signal (a target area emphasis sound whose sound quality on hearing is improved by mixing) output from the signal mixing unit 9 is a speaker 11 as a sound (a sound signal for a sound system) heard by a human. Supplied to On the other hand, in the sound collection device 100, an audio signal (a target area sound with a disturbance noise and background noise sufficiently suppressed; an audio signal for a voice recognition system) output from the signal mixing unit 9 is supplied to the voice recognition unit 10. Be done.

(A-3) Effects of the First Embodiment According to the first embodiment, the following effects can be achieved.

  In the sound collection device 100 according to the first embodiment, the input signal and the estimated noise are mixed with the area sound collection output for a call system (in this embodiment, the speaker 11) on the premise that a person listens to the sound. Thus, an acoustic signal in which musical noise is masked is output. In the sound signal for the call system generated by the sound collection device 100 of the first embodiment, the distortion of the target area sound is corrected, and the voice with improved ease of listening while maintaining the feeling of emphasis is obtained. . In other words, the sound signal for the communication system generated by the sound collection device 100 according to the first embodiment has a mixing area collection function of abnormal noise and distortion caused by the area collection processing when the background noise level is large. It becomes the target sound of high-quality sound reduced by.

  Further, in the sound collection device 100 according to the first embodiment, for the speech recognition system (in this embodiment, the machine speech recognition unit 10), the disturbance sound or noise that hinders speech recognition is sufficiently suppressed. The sound signal (sound in the target area is emphasized and a high recognition rate is ensured even under noise) is output. In other words, the sound signal for the speech recognition system generated by the sound collection device 100 according to the first embodiment is a target sound emphasis signal in which noise outside the target area that impedes the speech recognition performance and the disturbance sound are sufficiently suppressed. Become.

  Therefore, in the sound collection device 100 of the first embodiment, it is possible to realize higher voice recognition accuracy for the voice recognition system while providing an audio signal with improved sound quality to the call system. It is simultaneously realized to provide various sound signals. For example, it is assumed that the sound collecting apparatus 100 according to the first embodiment picks up the voice of a customer at a remote place and outputs it to a system that automatically converts the voice of the operator at the center (call center) and the customer into text. Do. In this case, the sound collection device 100 according to the first embodiment captures the voice of the customer at a remote place with the microphone arrays MA1 and MA2, and outputs an acoustic signal for a call system to the center operator (using the speaker 11) While outputting, the audio signal for the speech recognition system is output (output to the speech recognition unit 10) to a system for automatically texting the speech of the customer. The voice of the operator on the center side may be captured and input to the voice recognition unit 10 by a microphone (not shown). Further, for a customer at a remote location, the voice of the operator at the center may be captured by a microphone (not shown) and output from a speaker (not shown) at the remote location. As a result, the speech recognition unit 10 that receives the supply of the sound signal for the speech recognition system not only automatically converts the operator-customer interaction into text and saves it with higher recognition accuracy, but also quickly based on the recognition result The service can be improved by pulling out customer information and automatically searching product information. Meanwhile, the operator can talk with the customer with the improved sound quality, and the burden of work is reduced.

(B) Second Embodiment Hereinafter, a second embodiment of a sound collection device, program and method according to the present invention will be described in detail with reference to the drawings.

(B-1) Configuration of Second Embodiment FIG. 4 is a block diagram showing the entire configuration of a sound collection device 100A of the second embodiment, and the same portions as or corresponding portions to those of FIG. The code or the corresponding code is attached.

  The differences between the second embodiment and the first embodiment will be described below.

  Generally, in noise suppression processing, the amount of noise suppression and the sound quality are in a trade-off relationship. That is, if the amount of suppression is increased, distortion will increase. Although the area pickup is an excellent method capable of emphasizing only the sound generated in the target area, the distortion is increased as the emphasizing effect is enhanced like general noise suppression. Therefore, in the first embodiment, the result of area collection processing with high suppression effect is output to the voice recognition system, and the result of mixing area collection processing of high sound quality is output to the speaker or communication system of the call system. Showed a configuration to output

  By the way, recent speech recognition engines have become more resistant to ambient noise, and some are able to maintain recognition performance even in a certain noise environment. For such engines, it is not always a good idea to give top priority only to the suppression amount.

  Therefore, in the sound collection device 100A of the second embodiment, different mixing amounts (mixing levels) are set for the communication system and the speech recognition system, and the mixing area sound collection output optimal for each system is set. It can be provided.

  Next, the difference between the internal configuration of the sound collection device 100A of the second embodiment and that of the first embodiment will be described.

  In the sound collection device 100A of the second embodiment, the mixing level calculation unit 7, the mixing level adjustment unit 8, and the signal mixing unit 9 respectively have a mixing level calculation unit 7A, a mixing level adjustment unit 8A, and a signal mixing unit 9A. The second embodiment differs from the first embodiment in that Details of processing contents of the mixing level calculation unit 7A, the mixing level adjustment unit 8A, and the signal mixing unit 9A will be described later.

(B-2) Operation of Second Embodiment Next, the operation (sound collection method according to the embodiment) of the sound collection device 100A in the second embodiment having the above configuration will be described with the first embodiment. Explain the difference between

  As described above, in the sound collection device 100A of the second embodiment, what differs from the first embodiment is the mixing level calculation unit 7A, the mixing level adjustment unit 8A, and the signal mixing unit 9A. The processing of these elements will be mainly described.

  The mixing level calculation unit 7A is the same as that of the first embodiment, but the mixing level (hereinafter referred to as “L1”) for the communication system (for the speaker 11) and the speech recognition system (for the speech recognition unit 10). The second embodiment differs from the first embodiment in that two types of mixing levels (hereinafter referred to as “L2”) are calculated.

As with the first embodiment, the mixing level calculator 7A applies the estimated noise B ₁ (n), the non-target area sound M ₁ (n) other than the target area direction, and the target area to the speaker 11 as in the first embodiment. When the sum of the non-target area sound N ₁ (n) in the direction is A ₁ (n), the mixing level is δ ₁ A ₁ (n). Here, δ ₁ is a variable that is proportional to the SN ratio of the target area sound Z ₁ (n) and A ₁ (n), and is, for example, a value that makes A ₁ (n) −20 dB at an SN ratio of 0 dB.

Further, the mixing level calculation unit 7A supplies, to the voice recognition unit 10, an acoustic signal in which the amount of suppression is emphasized and the distortion is suppressed as compared to the speaker 11. For example, the mixing level calculation unit 7A may set a value to set A ₁ (n) to −25 dB at an SN ratio of 0 dB as δ ₂ for the speech recognition unit 10.

  The mixing level adjustment unit 8A uses the mixing level L1 for the sound signal supplied to the communication system, and the volume level of the input signal to be mixed with the target area sound and the estimated noise (mixing ratio; hereinafter, “R1” Determine). In addition, the mixing level adjustment unit 8A uses the mixing level L2 for the sound signal supplied to the voice recognition system, and the volume level of the input signal to be mixed with the target area sound and the estimation noise (mixing ratio; To determine the

  The signal mixing unit 9A generates a mixed sound mixed based on the mixing ratio R2 for the speech recognition system and a mixed sound mixed based on the mixing ratio R1 for the speech communication system calculated by the mixing level adjustment unit 8A, It differs from the first embodiment in that each mixed sound is supplied to the corresponding system. Specifically, the signal mixing unit 9A is a voice system system for the target area sound extracted by the target area sound extraction unit 6A, the input signal acquired by the signal input unit 3, and the noise estimated by the noise suppression unit 4. The mixed sound mixed based on the direction mixing ratio R1 is supplied to the speaker 11. Further, the signal mixing unit 9A mixes the target area sound extracted by the target area sound extraction unit 6A with the input signal acquired by the signal input unit 3 and the noise estimated by the noise suppression unit 4 for the speech recognition system. The mixed sound mixed based on the ratio R2 is supplied to the speech recognition unit 10.

(B-3) Effects of Second Embodiment According to the second embodiment, the following effects can be achieved in addition to the effects of the first embodiment.

  In the sound collection device 100A of the second embodiment, since different mixing amounts (mixing ratios) can be set for the communication system and for the voice recognition system, acoustic signals of the mixing amounts suitable for the respective systems (sound collection results) Can be generated and supplied. In other words, in the sound collection device 100A of the second embodiment, the noise suppression amount and the sound quality can be given in the best input state adapted to the system even to the recent speech recognition system with improved noise resistance. .

(C) Third Embodiment Hereinafter, a third embodiment of a sound collection device, program and method according to the present invention will be described in detail with reference to the drawings.

(C-1) Configuration of Third Embodiment FIG. 5 is a block diagram showing the entire configuration of a sound collection device 100B of the third embodiment, and the same portions as or corresponding portions to those of FIG. The code or the corresponding code is attached.

  The differences between the third embodiment and the second embodiment will be described below.

  Musical noise generated in the process of noise suppression and the like is artificial noise and is an offensive sound to human beings, but it has a characteristic that the influence on speech recognition is not so great. Although mixing estimated noise into the area sound output is effective for reducing musical noise, it is not always effective for improving the speech recognition accuracy in the speech recognition system.

  Therefore, in the sound collection device 100B of the third embodiment, for a speech system such as the speaker 11, both the input signal and the estimated noise are mixed with the target area sound to improve the sound quality, and the speech recognition unit 10 etc. For the speech recognition system of the above, only input signals effective for distortion reduction of the target speech are mixed.

  Next, with respect to the internal configuration of the sound collection device 100A of the third embodiment, differences from the second embodiment will be described.

  In the sound collection device 100B, the mixing level calculation unit 7A, the mixing level adjustment unit 8A, and the signal mixing unit 9A are excluded, and instead, two signal mixing units 91 and 92 are disposed. The first signal mixing unit 91 has a function of generating mixed sound for a voice system, and the second signal mixing unit 92 has a function of generating mixed sound for a speech recognition system.

(C-2) Operation of Third Embodiment Next, the operation of the sound collection device 100B of the first embodiment having the configuration as described above will be described. Hereinafter, the sound collection device 100B of the third embodiment will be described focusing on differences from the second embodiment.

  The processes from the microphone array MA1MA2 to the target area sound extraction unit 6 are the same as those in the first embodiment or the second embodiment.

  Further, in the embodiment, two signal mixing units of the signal mixing unit 91 and the signal mixing unit 92 are provided, and the mixing level (mixing ratio) in each signal mixing unit is the same as that in the first embodiment or the first embodiment. It shall be suitably calculated by the procedure similar to 2 embodiment. That is, in FIG. 5, in order to simplify the illustration, it is assumed that calculation processing of the mixing level (mixing ratio) or the like is performed in each signal mixing unit.

  In the sound collection device 100B according to the third embodiment, as shown in FIG. 5, a signal mixing unit 91 for a call system and a signal mixing unit 92 for a speech recognition system are provided following the target area sound extraction unit 6. Prepare.

  The target area sound extracted by the target area sound extraction unit 6 is mixed at the signal mixing unit 91 with the input signal from the signal input unit 3 and the estimated noise calculated by the noise suppression unit 4 at a suitable mixing level, It is sent to the system system (speaker 11). For example, in the signal mixing unit 91, the same processing (mixing level calculation unit, mixing level adjustment unit, and signal mixing unit as in the mixed sound generation processing for audio systems in the first embodiment or the second embodiment) And the same processing as described above is performed. Then, the mixed sound for the audio system generated by the signal mixing unit 91 is supplied to the speaker 11.

  Also, the target area sound extracted by the target area sound extraction unit 6 is mixed with the input signal from the signal input unit 3 at a suitable mixing level in the other signal mixing unit 92 and sent out to the voice recognition system. Ru. At this time, as described above, the component of the estimated noise calculated by the noise suppression unit 4 is not mixed with the mixed sound generated by the signal mixing unit 92. That is, the signal mixing unit 92 generates a mixed sound obtained by mixing the target area sound extracted by the target area sound extraction unit 6 and the input signal from the signal input unit 3 for the speech recognition system, and the speech recognition unit 10 Supply to

The mixing level when mixing the target area sound and the input signal in the signal mixing unit 92 may be calculated, for example, by the same process as the mixing level for the speech recognition system in the second embodiment. . In addition, the signal mixing unit 92 sets the ratio of estimated noise to be mixed with the target area sound to 0 by setting λ ₁ = 1 in the above equations (11) and (12), and only the component of the input signal The mixing ratio may be adjusted to be mixed.

(C-3) Effects of Third Embodiment According to the third embodiment, the following effects can be achieved in addition to the effects of the first embodiment.

  In the sound collection device 100B according to the third embodiment, musical noise is masked by mixing the input signal and the estimated noise in the area sound collection output for a call system on the premise of human listening. An acoustic signal (mixed sound) is output. As a result, in the sound collection device 100B of the third embodiment, the distortion of the target area sound is corrected as an audio signal for the call system, and the voice whose ease of listening is improved while the emphasis is maintained is maintained. Provided. Further, in the sound collection device 100B of the third embodiment, for the voice recognition system, an acoustic signal in which only input signals effective for reducing voice distortion are mixed is generated. Can contribute to the improvement of the accuracy of speech recognition in

(D) Other Embodiments The present invention is not limited to the above-described embodiments, and may include modified embodiments as exemplified below.

  (D-1) In the sound collection device of each of the above embodiments, the number of microphones of each microphone array MA used for sound collection is two, but for an acoustic signal collected using three or more microphones Sound based on the target area may be picked up. In each of the above-described embodiments, various existing methods can be applied to the number of microphones for each microphone array MA to be applied and the method for collecting sound in the target sound direction.

100 ... sound collecting device, 3 ... signal input unit, 4 ... noise suppression unit, 5 ... directivity forming unit, 6 ... target area sound extraction unit, 7 ... mixing level calculation unit, 8 ... mixing level adjustment unit, 9 ... signal Mixing section,
200 ... Subtractive BF, 210 ... Delay, 220 ... Subtractor.

Claims (7)

In a sound collection device that outputs a collected sound signal collected using a microphone array to a plurality of output destinations,
Background noise included in an input signal input from a microphone array is estimated and acquired as estimation noise, and the acquired estimation noise is used to suppress a noise component of the input signal to acquire a noise-suppressed signal Suppression means,
Directivity forming means for acquiring, for the noise-suppressed signal, a first non-target area sound in which directivity is formed in a direction other than the target area direction and a target area direction sound in which directivity is formed in the target area direction; ,
The target area direction sound is used to extract a second non-target area sound from the target area direction, and the second non-target area sound and the target area direction sound are used to generate a target area as a sound source. A target area sound extraction unit for acquiring a target area sound;
An audio signal including at least a component of the target area sound is output as the collected signal to each of the output destinations, and at least the input signal is output to the target area sound according to the characteristics of the output destination. Sound pickup apparatus characterized by: output means capable of generating a target area sound after mixing the mixed sound including the component of the component and / or the component of the estimated noise and outputting it as the sound collection signal .   The output means sets a sound including the component of the input signal and the component of the estimated noise as the mixed sound to the output destination of the speech system, and the mixed sound is output to the output destination of the voice recognition system. The sound pickup apparatus according to claim 1, wherein an audio signal not included is output as the sound pickup signal.   The sound pickup apparatus according to claim 1, wherein the output unit adjusts the volume level of the mixed sound in accordance with the characteristics of the output destination.   The output means sets the volume level of the mixed sound set as the output destination of the voice recognition system lower than the volume level of the mixed sound set as the output destination of the speech system. A sound pickup device according to Item 2.   The output means sets a sound including the component of the input signal and the component of the estimated noise as the mixed sound to the output destination of the speech recognition system, and the mixed sound to the output destination of a speech system The sound pickup apparatus according to claim 1, wherein the sound including the input signal and the estimated noise not including the input signal is set. A computer mounted on a sound collection device that outputs a collected sound signal collected using a microphone array to a plurality of output destinations,
Background noise included in an input signal input from a microphone array is estimated and acquired as estimation noise, and the acquired estimation noise is used to suppress a noise component of the input signal to acquire a noise-suppressed signal Suppression means,
Directivity forming means for acquiring, for the noise-suppressed signal, a first non-target area sound in which directivity is formed in a direction other than the target area direction and a target area direction sound in which directivity is formed in the target area direction; ,
The target area direction sound is used to extract a second non-target area sound from the target area direction, and the second non-target area sound and the target area direction sound are used to generate a target area as a sound source. A target area sound extraction unit for acquiring a target area sound;
An audio signal including at least a component of the target area sound is output as the collected signal to each of the output destinations, and at least the input signal is output to the target area sound according to the characteristics of the output destination. A sound pickup program characterized by comprising: an output means capable of generating a target area sound after mixing the mixed sound including the component of the component and / or the component of the estimated noise and outputting it as the sound collection signal . In a sound collection method performed by a sound collection device that outputs a collected sound signal collected using a microphone array to a plurality of output destinations,
The sound collection device includes noise suppression means, directivity formation means, target area sound extraction unit, and output means.
The noise suppression unit estimates background noise included in an input signal input from a microphone array and acquires it as estimation noise, and uses the acquired estimation noise to suppress noise components of the input signal to suppress noise Get a signal after
The directivity forming means includes, for the noise-suppressed signal, a first non-target area sound in which directivity is formed in a direction other than the target area direction and a target area direction sound in which directivity is formed in the target area direction. Acquired,
The target area sound extraction unit extracts a second non-target area sound from the target area direction using the target area direction sound, and further includes the second non-target area sound and the target area direction sound. Use the target area sound source to obtain the target area sound,
The output means outputs an audio signal including at least a component of the target area sound as the collected sound signal to each of the output destinations, and the output area includes the target area sound according to the characteristics of the output destination. It is possible to generate a target area sound after mixing a mixed sound including at least the component of the input signal and / or the component of the estimated noise, and output it as the sound collection signal. .

Images