JP6436180B2 - Sound collecting apparatus, program and method - Google Patents

Description

  The present invention relates to a sound collection device, a program, and a method, and can be applied, for example, to a process of emphasizing sounds in a target area and suppressing sounds in other areas.

  As a technique for separating and collecting only sound in a specific direction in an environment where a plurality of sound sources exist, there is a beam former (Beam Former; hereinafter referred to as “BF”) using a microphone array. BF is a technique for forming directivity using the time difference between signals reaching each microphone (see Non-Patent Document 1).

  BF is roughly divided into two types, an addition type and a subtraction type. In particular, the subtraction type BF has an advantage that directivity can be formed with a smaller number of microphones than the addition type BF.

  FIG. 5 is a block diagram showing a configuration related to a conventional subtraction type BF.

  In the conventional subtraction type BF shown in FIG. 5, the number of microphones is two.

  The conventional subtractive BF first calculates the time difference between signals arriving at each microphone by sounds that are present in a target direction (hereinafter also referred to as “target sound”) by a delay device, and adds a delay to the target sound. Match the phase. In the conventional subtractor BF delay unit, the time difference is calculated by the following equation (1).

In the following formula (1), d is the distance between the microphones, c is the speed of sound, and τ _i is the delay amount. In the following equation (1), θ _L is an angle from a vertical direction to a target direction with respect to a straight line connecting the microphones.
τ _L = (dsin θ _L ) / c (1)

Here, when the blind spot exists in the direction of the first microphone with respect to the center of the first microphone and the second microphone, the delay unit in the conventional subtractive BF has the input signal x ₁ ( Delay processing is performed for t). Thereafter, the input signal x ₁ (t) subjected to the delay process is subjected to a subtraction process according to the equation (2).
A (t) = x ₂ (t) −x ₁ (t−τ _L ) (2)

The subtraction process can be performed in the same manner in the frequency domain. In that case, the expression (2) is changed as follows.

Here, when θ _L = ± π / 2, the formed directivity is cardioid unidirectional as shown in FIG. 6A, and when θ _L = 0, π, FIG. As shown in (B), the figure is bi-directional. Here, a filter that forms unidirectionality from an input signal is called a unidirectional filter, and a filter that forms bidirectionality is called a bidirectional filter.

Further, by using a spectral subtraction (hereinafter also referred to as “SS”), directivity that is strong against a blind spot of bi-directionality can be formed. The formation of directivity by SS follows equation (4). In the equation (4), the input signal X1 of the _first microphone M1 is used, but the same effect can be obtained with the input signal X2 of the _second microphone M2. In the equation (4), β is a coefficient for adjusting the strength of SS. If the value becomes negative during subtraction, flooring processing is performed in which 0 or the original value is replaced with a smaller value. This method uses a bi-directional filter to extract sound that exists outside the target direction (hereinafter also referred to as “non-target sound”), and subtracts the power spectrum of the extracted non-target sound from the power spectrum of the input signal. The target sound can be emphasized.
Y ₁ = X ₁ −βA ₁ (4)

  When it is desired to pick up only sound existing in a certain area (hereinafter referred to as “target area sound”), the sound source (hereinafter referred to as “non-target area”) around that area is simply obtained by using the subtractive BF. (Also called “sound”) may be picked up. Thus, Patent Document 1 proposes a method of collecting a target area sound by using a plurality of microphone arrays, directing directivity from different directions to the target area, and crossing the directivities in the target area.

  FIG. 7 is an explanatory diagram showing a configuration example of each microphone array in a case where the target area sound from the sound source in the target area is collected using the two microphone arrays MA1 and MA2.

  FIG. 8 is an explanatory diagram (graph) showing the BF outputs of the microphone arrays MA1 and MA2 shown in FIG. 7 in the frequency domain. FIGS. 8A and 8B are graphs (image diagrams) showing the BF outputs of the microphone arrays MA1 and MA2 in the frequency domain, respectively.

In the method described in Patent Document 1, first, the ratio of the power of the target area sound included in the BF output of each of the microphone arrays MA1 and MA2 is estimated and used as a correction coefficient. As an example, when two microphone arrays MA1 and MA2 are used, the correction coefficient for the target area sound power is calculated by the equations (5), (6) or (7), (8).

Here, | Y _1k |, | Y _2k | is the power of the frequency k of the BF outputs of the microphone arrays MA1 and MA2, N is the total number of frequency bins, and α is a power correction coefficient for the BF output. Further, mode represents the mode value and median represents the median value. Thereafter, each BF output is corrected by the correction coefficient, and SS is performed to extract the non-target area sound existing in the target area direction. Further, by extracting the extracted non-target area sound from the output of each BF, the own area sound can be extracted.

  FIG. 9 is an explanatory diagram (image is a diagram) showing a change in the power spectrum of each frequency component when area sound collection processing is performed based on the BF output acquired using the microphone arrays MA1 and MA2 shown in FIG. .

First, the input signal _{X 1} of the microphone array MA1, obtain BF output _{Y 1} that suppresses the non-target area sound _{N 2} (see FIG. 9 (a)).

In order to extract the non-target area sound N ₁ (n) existing in the direction of the target area viewed from the microphone array MA1, the microphone array MA2 is extracted from the BF output Y ₂ (n) of the microphone array MA1 as shown in the equation (7). SS obtained by multiplying the BF output Y ₂ (n) by the power correction coefficient α (see FIG. 9B). After that, according to the equation (8), the non-target area sound is SS from each BF output to extract the target area sound (see FIG. 9C). γ (n) is a coefficient for changing the strength at the time of SS.
N ₁ = Y ₁ −αY ₂ (7)
Z ₁ = Y ₁ −γN ₁ (8)

  As described above, in the method described in Patent Document 1, since the non-linear processing SS is performed by the equations (4) and (8) in order to extract the target area sound, in a high noise environment, An unpleasant noise called musical noise may occur.

Therefore, in Patent Document 2, the section where the target area sound exists and the section where the target area sound does not exist are determined. In the section where the target area sound does not exist, the sound that has been subjected to the area sound collection processing is not output. It is suppressed. In order to determine whether or not the target area sound exists, first, the power spectrum ratio (hereinafter referred to as “area sound output data”) between the input signal and the output data obtained by extracting the target area sound according to the equation (9) ( Area sound output data / input signal) is calculated. When a sound source exists in the target area, the target area sound is included in both the input signal X ₁ and the area sound output data Z ₁ , so that the power spectrum ratio of the target area sound component is close to 1. Conversely, the non-target area sound component is suppressed in the area sound output data, and therefore the power spectrum ratio is a small value. In addition, in the apparatus described in Patent Document 2, other background noise components are also subjected to SS several times in the area sound collection process, so that the power spectrum ratio is reduced to some extent without performing dedicated noise suppression processing in advance. Small value. On the contrary, when the target area sound does not exist, the area sound output data includes only weak noise that is not erased compared to the input signal, and therefore, the power spectrum ratio becomes a small value in the entire region. Due to this feature, taking the average of the power spectrum ratios obtained at the respective frequencies according to the equation (10), a large difference is produced between when the target area sound exists and when it does not exist. Here, m and n are the upper limit and the lower limit of the processing band, respectively, and are set to, for example, 100 Hz to 6 kHz in which audio information is sufficiently included. And in the apparatus described in Patent Document 2, the average power spectrum ratio is determined by a preset threshold, and when it is determined that the target area sound does not exist, no sound is output without outputting the area sound output data, or Outputs sound with reduced gain of input sound.

JP 2014-072708 A JP, 2006-127457, A

Asano Tadashi, "Acoustic Technology Series 16 Sound Array Signal Processing-Sound Source Localization / Tracking and Separation-", Acoustical Society of Japan, Corona, February 25, 2011

  If the method described in Patent Document 1 is used, even if a non-target area sound exists around the target area, the target area sound can be collected. That is, if the method of Patent Document 1 is used, it is possible to extract only sound existing in the area even under high noise. However, in a situation where the SN ratio is 0 dB or less, there is a possibility that part of the target area sound component and the non-target area sound component overlap. If the target area sound is extracted by the SS in this state, the target area sound component is deleted due to the influence of the non-target area sound component overlapping the target area sound component, and as a result, the extracted target area sound is distorted. Furthermore, there is a risk that musical noise will also become stronger.

  Moreover, if the method described in Patent Document 2 is used, it is possible to suppress the influence of musical noise generated in the area sound collection processing. However, in a high noise environment such as a place where there are many people such as an event venue or a place where music is flowing in the surroundings, there is a possibility that the SN ratio deteriorates and the power spectrum of the sound output by area sound collection becomes small. is there. In such a situation, the average power spectrum ratio between the area sound collection output and the input signal also becomes small. In particular, components with originally low power, such as unvoiced consonants, have a poor target area sound determination accuracy, and a part of the target area sound may be lost. Also, when improving the sound quality by mixing the input signals, if the target area sound judgment accuracy deteriorates, it will be output even when the target area sound does not exist, and only the non-target area sound included in the input signal may be heard. is there.

  In view of the above problems, a sound collection device, a program, and a method, and a determination device, a program, and a method that can improve the determination accuracy of a target area sound in an environment with strong background noise are desired. .

The first aspect of the present invention is: (1) directivity forming means for forming directivity in the target area direction from the input signal by a beam former; and (2) in the target area direction by directivity formed by the directivity forming means. Non-target area sound extraction means for extracting existing non-target area sound; and (3) using non-target area sound existing in the target area direction extracted by the non-target area sound extraction means from the output of the beamformer. (4) power spectrum ratio calculating means for calculating a power spectrum ratio between the input signal and the extracted sound for each frequency component; 5) Determination means for determining whether or not there is a target area sound for each frequency component using the power spectrum ratio calculated by the power spectrum ratio calculation means; and (6) The determined frequency components with sound object area is present in serial determination means possess and output means for outputting the frequency components of said extracted sound, (7) and the output means, with respect to all frequency components, When the ratio of the frequency components determined by the determining means that the target area sound does not exist in the input signal exceeds a second threshold, the extracted sound is not output with all frequency components .

The sound collecting program of the second aspect of the present invention is formed by (1) directivity forming means for forming directivity in the direction of a target area by a beam former from an input signal, and (2) the directivity forming means. Non-target area sound extracting means for extracting non-target area sound existing in the target area direction due to directivity; and (3) existing in the target area direction extracted by the non-target area sound extracting means from the output of the beamformer. A target area sound extraction means for outputting an extracted sound obtained by extracting a target area sound using a non-target area sound; and (4) a power for calculating a power spectrum ratio of the input signal and the extracted sound for each frequency component. (5) Whether there is a target area sound for each frequency component using the power spectrum ratio calculated by the power spectrum ratio calculation means; Determination means for determining, (6) the determination for the determined frequency components with sound object area is present in means to function as output means for outputting the frequency components of said extracted sound, (7) the output means Outputs the extracted sound with all frequency components when the ratio of the frequency components determined by the determining means that the target area sound does not exist in the input signal exceeds the second threshold for all frequency components It is characterized by not .

According to a third aspect of the present invention, there is provided a sound collection method performed by the sound collection device. (1) Directivity forming means, non-target area sound extraction means, target area sound extraction means, power spectrum ratio calculation means, determination means, and output means (2) The directivity forming means forms directivity in the direction of the target area from the input signal by a beam former, and (3) the non-target area sound extracting means is formed by the directivity forming means. (4) The target area sound extraction means is present in the target area direction extracted by the non-target area sound extraction means from the output of the beamformer. (5) The power spectrum ratio calculation means outputs the input signal and the extracted sound for each frequency component. When the power spectrum ratio is calculated, (6) the determination means determines whether there is a target area sound for each frequency component using the power spectrum ratio calculated by the power spectrum ratio calculation means, and (7 ) The output means outputs the frequency component of the extracted sound for the frequency component determined by the determination means that the target area sound exists , and (8) the output means for all frequency components, When the ratio of the frequency components determined by the determining means that the target area sound does not exist in the input signal exceeds a second threshold, the extracted sound is not output with all frequency components .

  ADVANTAGE OF THE INVENTION According to this invention, the determination precision of the target area sound can be improved in the environment where background noise is strong.

It is the block diagram shown about the functional structure of the sound collection device which concerns on 1st Embodiment. It is the block diagram shown about the functional structure of the area sound collection process part which concerns on 1st Embodiment. It is the block diagram shown about the functional structure of the sound collection device which concerns on 2nd Embodiment. It is the block diagram shown about the functional structure of the sound collection device which concerns on 3rd Embodiment. It is a block diagram which shows the structure which concerns on the subtraction type BF in case the number of conventional microphones is two. It is a figure which shows the directional characteristic formed by the subtraction type | mold BF using the conventional two microphones. It is explanatory drawing shown about the example of a structure of each microphone array in the case of picking up the target area sound from the sound source of a target area using two conventional microphone arrays. It is explanatory drawing shown in the frequency domain about each BF output of the conventional two microphone array. It is explanatory drawing shown about the change of the power spectrum of each component at the time of area sound collection processing based on BF output acquired using the conventional two microphone arrays.

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

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

  The sound collection device 100 uses two microphone arrays MA (MA1, MA2) to perform target area sound collection processing for collecting a target area sound from a sound source in the target area.

  The microphone arrays MA1 and MA2 are arranged at any place in the air where the target area exists. For example, as shown in FIG. 7, the positions of the microphone arrays MA1 and MA2 with respect to the target area may be anywhere as long as the directivities overlap only in the target area. Each microphone array MA is composed of two or more microphones M, and an acoustic signal is collected by each microphone M. In this embodiment, description will be made assuming that two microphones M (M1, M2) that collect sound signals are arranged 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 a number of microphone arrays MA that can cover all areas.

  The sound collection device includes a data input unit 1, an area sound collection processing unit 2, a frequency-specific power ratio calculation unit 3, and a frequency-specific area sound determination unit 4.

  FIG. 2 is a block diagram showing an example of a functional configuration of the area sound collection processing unit 2.

  In the example of this embodiment, the area sound collection processing unit 2 includes a directivity forming unit 2-1, a delay correction unit 2-2, spatial coordinate data 2-3, a target area sound power correction coefficient calculation unit 2-4, and Description will be made assuming that the target area sound extraction unit 2-5 is included.

  Detailed processing of each functional block constituting the sound collection device 100 will be described later.

  The sound collection device 100 may be configured entirely by hardware (for example, a dedicated chip or the like), or may be partially or entirely configured as software (program). For example, the sound collection device 100 may be configured by installing a program (including the determination program and the sound collection program of the embodiment) in a computer having a processor and a memory.

(A-2) Operation of First Embodiment Next, the operation (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 data input unit 1 converts the acoustic signals collected by the microphone arrays MA1 and MA2 from analog signals to digital signals. And the data input part 1 performs the conversion process (For example, the process converted from a time domain to a frequency domain using a fast Fourier transform etc.) about the said digital signal.

  The area sound collection processing unit 2 forms directivity for each microphone array based on the input signal of the microphone array obtained from the data input unit 1, and extracts components included in the directivity as target area sounds. .

  In this embodiment, the area sound collection processing by the area sound collection processing unit 2 will be described as being realized by the configuration shown in FIG. 4, for example, but a configuration in which the target area sound is extracted using other methods is applied. You may make it do.

  Below, operation | movement of each component of the area sound collection process part 2 shown in FIG. 2 is demonstrated.

  The directivity forming unit 2-1 extracts, for each microphone array MA, non-target area sounds that exist in directions other than the target direction (for example, extraction by a bi-directional filter), and inputs the power spectrum of the extracted non-target area sounds. By subtracting from the power spectrum of the signal, a sound (BF output) having directivity in the direction of the target area is acquired. Specifically, the directivity forming unit 2-1 acquires, as a BF output, a sound in which directivity is formed in the direction of the target area by BF according to the equation (4) for each microphone array MA.

  The delay correction unit 2-2 calculates and corrects a delay caused by a difference in distance between the target area and each microphone array. The delay correction unit 2-2 acquires the position of the target area and the position of the microphone array from the spatial coordinate data 2-3, and calculates the difference in arrival time of the target area sound to each microphone array MA (MA1, MA2). . Then, the delay correction unit 2-2 makes the target area sound reach all the microphone arrays MA (MA1, MA2) simultaneously with reference to the microphone array MA (MA1, MA2) arranged farthest from the target area. To add delay.

  The spatial coordinate data 2-3 holds position information of all the target areas, the microphone arrays MA (MA1, MA2), and the microphones M (M1, M2) constituting the microphone arrays MA (MA1, MA2).

  The target area sound power correction coefficient calculation unit 2-4 calculates a correction coefficient for making the power of the target area sound component included in each BF output the same according to the equation (5) or (6).

  The target area sound extraction unit 2-5 performs SS on each BF output data corrected by the correction coefficient calculated by the target area sound power correction coefficient calculation unit 2-4 according to the equation (7), and removes noise present in the target area direction. Extract. Further, the target area sound extraction unit 2-5 extracts the target area sound by performing SS on the extracted noise from the output of each BF according to the equation (8).

The frequency-specific power ratio calculation unit 3 uses, for each frequency, the input signal X ₁ supplied from the data input unit ₁ and the area sound output data Z ₁ supplied from the area sound collection processing unit 2 for each frequency. The power ratio | R _k | is calculated. Specifically, the power ratio calculation unit 3 for each frequency calculates the power ratio for each frequency based on the equation (11). Here, | X _1k | is the power of frequency k in the input signal X ₁ (input signal of the first microphone M1) of the microphone array MA1, and | Z _1k | is the power of frequency k in the area sound output data. is there. M is the lower limit of the frequency to be processed, and n is the upper limit of the frequency.

  The frequency-specific area sound determination unit 4 compares the power ratio | R | calculated by the frequency-specific power ratio calculation unit 3 with a preset threshold T1 for each frequency to determine an area sound component. Specifically, the frequency-specific area sound determination unit 4 compares the power ratio | R | and the threshold value T1 for each frequency, and extracts components whose power ratio | R | exceeds the threshold value T1.

  In the frequency-specific area sound determination unit 4, the threshold value T1 may be the same value for all frequencies, or a different value may be applied for each frequency. In the frequency-specific area sound determination unit 4, for example, a value that decreases with increasing frequency from low to high may be applied to T <b> 1. In addition, the frequency-specific area sound determination unit 4 may set, for example, a value larger than the low frequency (for example, a frequency higher than 100 Hz) to T1 for the low frequency (for example, 100 Hz or less).

In this embodiment, the frequency-specific area sound determination unit 4 has an area sound component for the frequency (frequency component) where the power ratio | R | exceeds the threshold T1 (| R |> T1) (input signal X ₁ and there are components of the objective area sound area sound output data Z ₁₎ and is described as being determined.

Frequency-area sound determination unit 4, for a determined frequency area sound components are present (frequency components), and directly outputs the area sound output data Z ₁ supplied from the area sound-pickup processing section 2, the area sound component the frequency is determined not to exist, the area sound output data Z ₁ outputs predetermined sound data without outputting (e.g., silence data set in advance).

Note that the frequency-specific area sound determination unit 4 outputs the frequency determined to have no area sound component by decreasing the gain of the area sound output data Z ₁ or the input signal X ₁ instead of silence. Also good.

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

  In the sound collection device 100 of this embodiment, the power ratio between the area sound output data and the input signal is obtained for each frequency (for each frequency component), and it is determined whether the sound is the target area sound component. In the sound collection device 100 of this embodiment, the power ratio of each frequency is compared with a preset threshold value T1, and a frequency whose power ratio exceeds the threshold value T1 is determined as a target area sound component. Output area sound output data. Further, in the sound collection device 100 of this embodiment, it is determined that the frequency whose power ratio is lower than the threshold value T1 is not the target area sound component, and nothing is output at that frequency, or the gain of the area sound output data is lowered. Output. In the area sound output data, the value of the main component of the target area sound becomes large. Therefore, in the sound collection device 100 of this embodiment, the component in which the target area sound exists is output as it is. Further, in the sound collecting apparatus 100 of this embodiment, a component whose value is small and determined not to be the target area sound component is not output, but there is no influence because it is not related to the target area sound. Even in a voiceless consonant having a small average power in the entire band, there is a peak in the power spectrum. However, in the sound collection device 100 of this embodiment, the power ratio is obtained for each frequency, so the main component of the voiceless consonant is large. It becomes a value and it is determined that it is the target area sound component.

  As described above, in the sound collection device 100 of this embodiment, the power ratio between the area sound output data and the input signal is obtained for each frequency component, the presence / absence of the target area sound component is determined, and the frequency determined as the target sound component By outputting only the components, it is possible to prevent the lack of the target area sound even in a high noise environment.

(B) Second Embodiment Hereinafter, a second embodiment of the 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. 3 is a block diagram showing a functional configuration of the sound collection device 100A of this embodiment. In FIG. 3, the same or corresponding parts as those in FIG.

  Hereinafter, the difference from the first embodiment will be described for the sound collection device 100A of the second embodiment.

  The sound collection device 100 </ b> A is different from the first embodiment in that an area sound determination unit 5 is added after the frequency-specific area sound determination unit 4.

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

  As described above, the sound collection device 100A is different from the first embodiment in that the area sound determination unit 5 is added. Below, the determination process of the target area sound centering on the area sound determination part 5 is demonstrated.

Area sound determination unit 5, there is interest area sound in a section if (input signal X ₁ and the area sound output data Z ₁ there is interest areas sound from the determination results of all frequencies of the frequency by area sound determination unit 4 sections determines that whether) a, when it is determined that an interval in which there is interest area sound output area sound output data Z ₁ at all frequencies, predetermined data in all frequency if it is determined non-existent and (For example, silent data) is output.

  Below, the specific example of operation | movement of the area sound determination part 5 is demonstrated.

  The area sound determination unit 5 first calculates the ratio of the frequency determined to be the target area sound component and the frequency determined not to be the target area sound component in the determination result of the frequency-specific area sound determination unit 4.

  For example, C1 is the number of frequencies determined to be the target area sound component (frequency where the power ratio | R | exceeds the threshold T1), and the frequency determined that is not the target area sound component (the power ratio | R | is the threshold T1). When the number of the following frequencies) is C2, the frequency ratio P1 determined to be the target area sound component is P1 = C1 / (C1 + C2), and the frequency ratio P2 determined not to be the target area sound component Is P2 = C2 / (C1 + C2). Note that the sum of C1 and C2 is the number of all frequency components (for example, the number of frequency bins).

  When the frequency ratio P2 determined not to be the target area sound component exceeds the threshold T2 [%] (when P2> T2; that is, when P1 <(100 [%] − T2 [%])), The area sound determination unit 5 updates the determination that all the frequencies (all components) are not the target area sound components, and outputs silence data for all the frequencies.

  When the frequency ratio P2 determined not to be the target area sound component falls below the threshold T2 (when P2 <T2; that is, when P1> (100 [%] − T2 [%])), the area sound determination The unit 5 compares P2 with the threshold value T3. T3 is smaller than T2 (T2> T3).

When the frequency ratio P2 determined not to be the target area sound component is lower than T3 (when P2 <T3; that is, when P1> (100% −T3 [%])), the area sound determination unit 5 updates to a determination that an object area sound components in all frequency, and outputs the area sound output data Z ₁ for all frequencies.

  When P2 does not fall below T3 (when T3 ≦ P2 ≦ T2; that is, when (100% −T3 [%]) ≧ P1 ≧ (100% −T2 [%])), the area sound determination unit 5 Performs output for each frequency in accordance with the determination by the frequency-specific area sound determination unit 4. That is, in this case, the area sound determination unit 5 outputs the content supplied from the preceding stage (frequency-specific area sound determination unit 4) as it is.

  The values of T2 and T3 are not limited, but may be T2 = 80% and T3 = 20%, for example.

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

  In the sound collection device 100A of the second embodiment, after the presence / absence of the target area sound component is determined for each frequency by the frequency-specific area sound determination unit 4, the area sound determination unit 5 further performs the target area sound component at all frequencies. The final output is determined from the ratio of. Then, the area sound determination unit 5 determines again that the target area sound component does not exist at all frequencies when the frequency determined that the target area sound component does not exist exceeds a certain ratio among all frequencies. Output silence data. Thereby, in the sound collection device 100A, even when there is a frequency erroneously determined that the target area sound exists when the target area sound does not exist, the influence can be suppressed.

(C) Third Embodiment Hereinafter, a third embodiment of the 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. 4 is a block diagram showing a functional configuration of the sound collection device 100B of this embodiment. In FIG. 4, the same or corresponding parts as those in FIG.

  Below, the difference with 1st Embodiment is demonstrated about the sound collection apparatus 100B of 3rd Embodiment.

  The sound collection device 100B is different from the first embodiment in that a signal mixing unit 6 and a mixing level calculation unit 7 are added. In the sound collecting device 100B, the signal mixing unit 6 is inserted in the subsequent stage of the frequency-specific area sound determination unit 4.

(C-2) Operation of the Third Embodiment Next, the operation of the sound collection device 100B of the third embodiment having the above-described configuration (sound collection method according to the embodiment) and the first embodiment. Explain the difference.

  As described above, the sound collection device 100B is different from the first embodiment in that the signal mixing unit 6 and the mixing level calculation unit 7 are added. Below, the determination process of the target area sound centering on the signal mixing part 6 and the mixing level calculation part 7 is demonstrated.

The mixing level calculation unit 7 uses the ratio of the area sound output data Z ₁ and the non-target area sound N ₁ (hereinafter referred to as “SN ratio”) to mix the input signal X ₁ with the target area sound to be output (output data). Determine the volume level. The power spectrum O ₁ of the non-target area sound N ₁ may be extracted by, for example, SSing the area sound output data Z ₁ from the input signal X ₁ according to the equation (3). That is, O ₁ can be expressed as shown in Equation (12). The mixing level coefficient δ ₁ for adjusting the mixing volume level of the input signal X ₁ is a variable proportional to the SN ratio Z ₁ / O ₁ of the area sound output data Z ₁ and the non-target area sound N ₁ , for example, the SN ratio 0 dB. in the value of the _{X 1} to -20 dB. Due to δ ₁ , the mixed sound volume level becomes δ ₁ X ₁ . Further, δ ₁ may be set to δ ₁ Φ ₁ by weighting for each frequency instead of a constant value for all frequencies. Here, Φ ₁ may be set to a value that decreases, for example, from the low range to the high range. In that case, the mixed sound volume level is δ ₁ Φ ₁ X ₁ .
O ₁ = X ₁ −Z ₁ (12)

The signal mixing unit 6 uses the frequency acquired by the data input unit 1 as the area sound output data extracted by the area sound collection processing unit 2 at the frequency determined by the frequency-specific area sound determination unit 4 as the target area sound component. The signals are mixed based on the level calculated by the mixing level calculation unit 7. The final output | W _1k | is assumed to be mixed according to the following equation (13). Here, k is a frequency determined by the frequency-specific area sound determination unit 4 to be a target area sound component.
| W _1k | = | Z _1k | + δ ₁ | X _1k | (13)

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

  In the sound collection device 100B of the third embodiment, after determining whether or not there is a target area sound component for each frequency by the frequency-specific area sound determination unit 4, the signal mixing unit 6 and the mixing level calculation unit 7 perform the target area sound component. The gain of the input signal is adjusted and added only to the frequency determined to be output. Thereby, in the sound collection device 100B, since the input signal is added only to the target area sound collection component, mixing of the non-target area sound can be prevented and distortion of the target area sound can be corrected.

  In other words, in the sound collection device 100B of the third embodiment, when mixing the input signal for correcting distortion of the target area sound, the input signal is added only to the frequency determined as the target area sound component. Even if the non-target area sound exists in the input signal, the probability that the target area sound component and the non-target area sound component overlap at each frequency is low. Therefore, in the sound collection device 100B of the third embodiment, the non-target area sound component is not added to the output (sound collection result), and only the target area sound component is finally output.

  In addition, when mixing input signals to correct distortion of the target area sound, the sound collection device 100B of the third embodiment outputs only the target area sound component even if there is a non-target area sound. The sound quality can be improved while maintaining the performance of area pickup.

(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 area sound determination unit 5 of the first embodiment, and when there is a component whose power ratio is larger than the threshold value T1 by a certain amount or more, the subsequent few seconds are related to the value of the power ratio for the component. A function for determining that the target area sound component is present (hangover function) may be added.

  (D-2) In the sound collection device 100A (area sound determination unit 5) of the second embodiment, not all frequencies (all bands) but all frequencies (all bands) are divided into a plurality of bands (hereinafter, this division). And the power ratio is calculated for each component of each divided band, the presence / absence of the target area sound is determined for each divided band, and the presence / absence of output for each divided band (in the sound collection result) It may be determined whether or not to add.

  Specifically, for example, when the ratio of the frequency (frequency at which the power ratio exceeds the threshold T1) determined as not being the target area sound component in a certain divided band exceeds the threshold T2, the area sound determination unit 5 It may be updated to the determination that the entire divided band is not the target area sound component, and silence data may be output.

  For example, when the ratio of the frequency determined not to be the target area sound component in a certain divided band falls below the threshold T2, the area sound determination unit 5 compares the ratio with the threshold T3 (T2> T3). . When the ratio of the divided band falls below T3, the area sound determination unit 5 updates the determination that the target area sound component exists in the entire divided band, and outputs area sound output data of the entire divided band. May be output.

  In addition, when the ratio of the divided band does not fall below T3, the area sound determination unit 5 outputs according to the determination result (determination result for each frequency) of the frequency-specific area sound determination unit 4 (frequency for the divided band). The output result of the separate area sound determination unit 4 may be output as it is).

  Furthermore, when there is a frequency determined to be the target area sound component in one of the divided bands, the area sound determination unit 5 determines that the target area sound component exists in the entire divided band. The area sound output data may be output for all frequencies of the divided band.

  (D-3) It is good also as a structure which combined 2nd Embodiment and 3rd Embodiment. Specifically, the signal mixing unit 6 and the mixing level calculation unit 7 may be added to the sound collection device 100A. In this case, the signal mixing unit 6 may be inserted after the area sound determination unit 5.

  DESCRIPTION OF SYMBOLS 100 ... Sound collection apparatus, 1 ... Data input part, 2 ... Area sound collection process part, 3 ... Power ratio calculation part according to frequency, 4 ... Area sound determination part according to frequency, 5 ... Area sound determination part, 2-1 ... Direction Sex formation unit, 2-2 ... delay correction unit, 2-3 ... spatial coordinate data, 2-4 ... target area sound power correction coefficient calculation unit, 2-5 ... target area sound extraction unit.

Claims (6)

Directivity forming means for forming directivity in the direction of the target area by a beamformer from an input signal;
Non-target area sound extracting means for extracting non-target area sound existing in the target area direction due to directivity formed by the directivity forming means;
A target area sound extraction means for outputting a sound extracted as a result of extracting a target area sound using a non-target area sound existing in a target area direction extracted by the non-target area sound extraction means from the output of the beam former; ,
Power spectrum ratio calculating means for calculating a power spectrum ratio of the input signal and the extracted sound for each frequency component;
Determination means for determining whether or not a target area sound exists for each frequency component, using the power spectrum ratio calculated by the power spectrum ratio calculation means;
An output means for outputting the frequency component of the extracted sound for the frequency component determined by the determination means that the target area sound exists ;
The output means extracts all frequency components when the ratio of the frequency components determined by the determining means that the target area sound does not exist in the input signal exceeds a second threshold with respect to all frequency components. A sound collecting device characterized by not outputting sound.   The determination means determines, for each frequency component, whether or not a target area sound exists based on a comparison result between the power spectrum ratio calculated by the power spectrum ratio calculation means and a first threshold value. The sound collection device according to claim 1. The output means has a ratio of frequency components that the determination means determines that no target area sound is present in the input signal for all frequency components below a third threshold value that is smaller than the second threshold value. If, pickup device according to claim 1, characterized in that it outputs the extracted sound in all frequency components. Mixing level calculation means for calculating a volume level of the input signal to be mixed with an output sound based on a ratio between the input signal and a non-target area sound extracted based on the extracted sound and the extracted sound. ,
The output means mixes and outputs the input signal gain-adjusted based on the volume level calculated by the mixing level calculation means for the frequency component determined by the determination means that the target area sound exists. sound collection device according to any one of claims 1 to 3,. Computer
Directivity forming means for forming directivity in the direction of the target area by a beamformer from an input signal;
Non-target area sound extracting means for extracting non-target area sound existing in the target area direction due to directivity formed by the directivity forming means;
A target area sound extraction means for outputting a sound extracted as a result of extracting a target area sound using a non-target area sound existing in a target area direction extracted by the non-target area sound extraction means from the output of the beam former; ,
Power spectrum ratio calculating means for calculating a power spectrum ratio of the input signal and the extracted sound for each frequency component;
Determination means for determining whether or not a target area sound exists for each frequency component, using the power spectrum ratio calculated by the power spectrum ratio calculation means;
To function as an output unit for outputting the frequency components of said extracted sound for the determined frequency components with sound object area is present in the determination unit,
The output means extracts all frequency components when the ratio of the frequency components determined by the determining means that the target area sound does not exist in the input signal exceeds a second threshold with respect to all frequency components. Sound collection program characterized by not outputting sound. In the sound collection method performed by the sound collection device,
Directivity forming means, non-target area sound extraction means, target area sound extraction means, power spectrum ratio calculation means, determination means, and output means,
The directivity forming means forms directivity in the direction of the target area by a beamformer from an input signal,
The non-target area sound extracting means extracts non-target area sound existing in the target area direction due to the directivity formed by the directivity forming means,
The target area sound extraction means extracts the extracted sound as a result of extracting the target area sound from the output of the beamformer using the non-target area sound existing in the target area direction extracted by the non-target area sound extraction means. Output,
The power spectrum ratio calculating means calculates the power spectrum ratio of the input signal and the extracted sound for each frequency component,
The determination means determines whether or not there is a target area sound for each frequency component, using the power spectrum ratio calculated by the power spectrum ratio calculation means,
The output means outputs the frequency component of the extracted sound for the frequency component determined by the determining means that the target area sound exists ,
The output means extracts all frequency components when the ratio of the frequency components determined by the determining means that the target area sound does not exist in the input signal exceeds a second threshold with respect to all frequency components. A sound collection method characterized by not outputting sound.

Images