JP7020283B2 - Sound source direction determination device, sound source direction determination method, and sound source direction determination program - Google Patents

Description

The present invention relates to a sound source direction determination device, a sound source direction determination method, and a sound source direction determination program.

The first directional microphone is arranged to detect the sound propagating along the first direction, and the second directional microphone is arranged to detect the sound propagating along the second direction intersecting the first direction. By doing so, there is a sound source direction determination device that determines the sound source direction. In this sound source direction determination device, when the sound pressure of the sound detected by the first directional microphone is larger than the sound pressure of the sound detected by the second directional microphone, the sound is along the first direction. It is determined that the sound is propagated. On the other hand, when the sound pressure of the sound detected by the second directional microphone is larger than the sound pressure of the sound detected by the first directional microphone, the sound propagates along the second direction. Is determined.

Japanese Unexamined Patent Publication No. 2018-40982 Japanese Patent No. 5387459

Watanabe et al., "Basic study on sound source position estimation using directional microphone", [online], [Search on September 11, 2017], Internet (URL: http: //www.cit.nihon-u. ac.jp/kouendata/No.41/2_denki/2-008.pdf) Kohei Yamamoto, "Diffraction Calculation Method", Noise Control, Japan, 1997, Vol. 21, No. 3, pp. 143-147

However, since the directional microphone is larger in size and more expensive than the omnidirectional microphone, the size of the sound source direction determination device is larger and the price is higher than when the omnidirectional microphone is used. There is.

One aspect of the present invention is to improve the accuracy of sound source direction determination using an omnidirectional microphone.

In one embodiment, the microphone installation portion is provided with a first sound path and a second sound path inside. The first sound path is provided with a first opening opened on the first flat surface at one end, and sound propagates from the first opening. The second sound path is provided at one end with a second opening opened in the second flat surface intersecting with the first flat surface, and a second sound path through which sound propagates from the second opening is provided inside. .. The first microphone is installed at the other end of the first sound path, and the second microphone is installed at the other end of the second sound path. The determination unit determines the direction in which the sound source exists based on at least one of the difference in sound pressure and the difference in phase. The difference in sound pressure is the first sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the second microphone. It is a difference from sound pressure. The difference in phase is the difference between the first phase, which is the phase of the second frequency component of the sound acquired by the first microphone, and the second phase, which is the phase of the second frequency component of the sound acquired by the second microphone. Is.

As one aspect, it is possible to improve the accuracy of sound source direction determination using an omnidirectional microphone.

It is a block diagram which shows an example of the information processing terminal which concerns on 1st to 5th Embodiment. It is a conceptual diagram which shows an example of the appearance of the sound source direction determination apparatus which concerns on 1st, 2nd, 4th, and 5th Embodiment. It is a conceptual diagram which shows an example of the appearance of the sound source direction determination apparatus which concerns on 1st, 2nd, 4th, and 5th Embodiment. It is sectional drawing which follows the cutting line 3-3 of FIG. 2A which concerns on 1st, 4th and 5th Embodiment. It is a conceptual diagram for demonstrating the diffraction of the sound of the 1st, 3rd, 4th and 5th embodiments. It is a conceptual diagram for demonstrating the diffraction of the sound of the 1st, 3rd, 4th and 5th embodiments. It is a table exemplifying the sound pressure difference between the sound pressure of the 1st microphone and the sound pressure of the 2nd microphone when the area of a flat surface is different. It is a conceptual diagram for demonstrating the diffraction of the sound of the 1st, 2nd, 4th and 5th embodiments. It is a conceptual diagram for demonstrating the diffraction of the sound of the 1st, 2nd, 4th and 5th embodiments. It is a graph for demonstrating the decrease of sound pressure by diffraction along a frequency axis. It is a block diagram which illustrates the outline of the sound source direction determination processing of 1st to 3rd Embodiment. It is a table exemplifying the sound source direction determination criteria of 1st to 4th Embodiment. It is a block diagram which shows an example of the hardware of the information processing terminal which concerns on 1st to 5th Embodiment. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 1st to 3rd Embodiment. It is sectional drawing which follows the cutting line 3-3 of FIG. 2A which concerns on 2nd Embodiment. It is a conceptual diagram which shows an example of the appearance of the sound source direction determination apparatus which concerns on 3rd and 5th Embodiment. It is a conceptual diagram which shows an example of the appearance of the sound source direction determination apparatus which concerns on 3rd and 5th Embodiment. It is a conceptual diagram which shows an example of the appearance of the sound source direction determination apparatus which concerns on 3rd and 5th Embodiment. 3 is a cross-sectional view taken along the cutting line 14-14 of FIG. 13A according to the third and fifth embodiments. It is a block diagram which illustrates the outline of the sound source direction determination process of 4th Embodiment. It is a conceptual diagram for demonstrating the phase difference when a sound reaches a microphone. It is a conceptual diagram for demonstrating the phase difference when a sound reaches a microphone. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 4th Embodiment. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 4th Embodiment. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 4th Embodiment. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 4th Embodiment. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 4th Embodiment. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 4th Embodiment. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 4th Embodiment. It is a conceptual diagram which shows an example of the sound source direction determination apparatus using the directional microphone which concerns on the related technology. It is an exemplary table for comparing the size of a directional microphone with the size of an omnidirectional microphone. It is a conceptual diagram which shows an example of the sound source direction determination apparatus using the omnidirectional microphone which concerns on the related technology. It is a conceptual diagram which shows an example of the sound source direction determination apparatus using the omnidirectional microphone which concerns on the related technology. It is a table which shows an example of the comparison between the sound pressure difference in a related technique and the sound pressure difference in this embodiment. It is a conceptual diagram for demonstrating the diffraction of a sound when a space exists on the back surface of a sound source direction determination apparatus. It is a conceptual diagram for demonstrating the diffraction of a sound when there is no void in the back surface of the sound source direction determination apparatus. It is a conceptual diagram which illustrates the high region sound pressure difference in the case where a gap exists in the back surface of the sound source direction determination device, and when it does not exist. It is a conceptual diagram which illustrates the normalized phase difference in the case where a gap exists in the back surface of the sound source direction determination apparatus, and when it does not exist. It is a block diagram which illustrates the outline of the sound source direction determination process of 5th Embodiment. It is a conceptual diagram for demonstrating the adjustment of the threshold value which determines a sound source direction. It is a conceptual diagram which illustrates the relationship between the inclination of the sound source direction determination device, and the threshold value which determines a sound source direction. It is a conceptual diagram which illustrates the relationship between the inclination of the sound source direction determination device, and the threshold value which determines a sound source direction. It is a conceptual diagram which illustrates the relationship between the inclination of the sound source direction determination device, and the phase difference of the sound reaching the upper surface and the front surface of the sound source direction determination device. It is a conceptual diagram which illustrates the relationship between the inclination of the sound source direction determination device, and the phase difference of the sound reaching the upper surface and the front surface of the sound source direction determination device. It is a conceptual diagram which illustrates the relationship between the phase difference of a user's voice and the phase difference of the voice of a dialogue partner when the inclination of the sound source direction determination device is different. It is a conceptual diagram which illustrates the relationship between the phase difference of a user's voice and the phase difference of the voice of a dialogue partner when the inclination of the sound source direction determination device is the same. It is a conceptual diagram which illustrates the relationship between the phase difference of a user's voice and the phase difference of the voice of a dialogue partner when the inclination of the sound source direction determination device is the same. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 5th Embodiment. It is a flowchart which shows an example of the flow of the sound source direction determination processing which concerns on 5th Embodiment.

[First Embodiment]
Hereinafter, an example of the first embodiment will be described in detail with reference to the drawings.

FIG. 1 illustrates the main functions of the information processing terminal 1. The information processing terminal 1 includes a sound source direction determination device 10 and a speech translation device 14.

The sound source direction determination device 10 includes a first microphone (hereinafter, “microphone” is also referred to as “microphone”) 11, a second microphone 12, and a determination unit 13. The speech translation device 14 includes a first translation unit 14A, a second translation unit 14B, and a speaker 14C.

Each of the first microphone 11 and the second microphone 12 is an omnidirectional microphone and acquires omnidirectional sound. The determination unit 13 determines the direction in which the sound source of the sound acquired by the first microphone 11 and the second microphone 12 exists. The voice translation device 14 determines a language represented by a voice signal corresponding to a sound propagating from the sound source direction acquired by the first microphone 11 or the second microphone 12 based on the sound source direction determined by the determination unit 13. Translate into the language of.

Specifically, when the determination unit 13 determines that the sound source exists in, for example, the upper first direction, the first translation unit 14A first determines the language represented by the audio signal corresponding to the acquired sound. Translate into a language (eg English). When the determination unit 13 determines that the sound source exists in the second direction, which is the front, for example, the second translation unit 14B uses the second language (for example, the language represented by the voice signal corresponding to the acquired sound). , Japanese). The speaker 14C outputs the language translated by the first translation unit 14A or the second translation unit 14B by voice.

2A and 2B illustrate the appearance of the sound source direction determination device 10. The sound source direction determination device 10 is used, for example, by putting it in the chest pocket of the user's shirt and fastening it to the corresponding portion of the clothing near the user's chest with a clip or a pin, or hanging it around the user's neck with a strap. It is a device that is supposed to be. FIG. 2A illustrates the upper surface of the housing 18 of the sound source direction determination device 10. The housing 18 is an example of a microphone installation portion. The upper surface of the housing 18, which is an example of the first flat surface, is a surface facing upward when the sound source direction determination device 10 is placed in the chest pocket, that is, a surface closest to the user's mouth.

On the upper surface of the housing 18, there is an opening 11O which is an example of the first opening provided at one end of the first sound path. A first microphone 11 is installed at the other end of the first sound path. Hereinafter, in the figure, the arrow FR represents the front of the sound source direction determination device 10. A speaker 14C is also arranged on the upper surface of the housing 18. That is, in the examples of FIGS. 2A and 2B, the speech translation device 14 is included in the housing 18 of the sound source direction determination device 10. The length of the upper surface of the housing 18 in the front-rear direction is, for example, 1 [cm].

FIG. 2B illustrates the front surface of the housing 18 of the sound source direction determination device 10. The front surface, which is an example of the second flat surface, is a surface facing the dialogue partner with which the user interacts, for example, when the sound source direction determination device 10 is placed in the chest pocket.

On the front surface of the housing 18, there is an opening 12O provided at one end of the second sound path. A second microphone 12 is installed at the other end of the second sound path. Hereinafter, in the figure, the arrow UP represents the upper part of the sound source direction determination device 10. The size of the front surface of the housing 18 is, for example, about the same size as a general business card.

The sound source direction determination device 10 determines that the sound determined to have a sound source above is the voice uttered by the user, translates it into a first language, and outputs the voice from the speaker 14C. A voice signal corresponding to the sound is transmitted to the first translation unit 14A of 14. Further, the sound source direction determination device 10 determines that the sound determined to have a sound source in front is the voice uttered by the dialogue partner. The sound source direction determination device 10 transmits a voice signal corresponding to the sound to the second translation unit 14B of the voice translation device 14 so as to translate into a second language and output the sound from the speaker 14C.

FIG. 3 represents a cross-sectional view taken along the cutting line 3-3 of FIG. 2A. One end of the second sound path 12R is provided with an opening 12O opened in the front surface of the housing 18, and the second microphone 12 is installed at the other end of the second sound path.

One end of the first sound path 11R is provided with an opening 11O opened on the upper surface of the housing 18, and the first microphone 11 is installed at the other end of the first sound path 11R. The first sound path 11R has a bent portion 11K in the middle. The bent portion 11K is an example of the second diffraction portion.

FIG. 4A illustrates a case where the sound source is located in front of the sound source direction determination device 10. When the area of the front surface of the housing 18 is larger than the predetermined value which is an example of the first predetermined value, the second microphone 12 is reflected by the front surface of the housing 18 in addition to the sound directly arriving through the opening 12O. The sound diffracted by the opening 12O, which is an example of the third diffractive part, is acquired.

FIG. 4B illustrates a case where the sound source is located above the sound source direction determination device 10. The sound does not reach the second microphone 12 directly, and the second microphone 12 acquires the sound diffracted by the opening 12O. Therefore, the sound pressure of the sound acquired by the second microphone 12 is higher when the sound source is present in the front than when the sound source is located above.

FIG. 5 illustrates the sound pressure acquired by the second microphone 12 when the sound source is present in front of and above the sound source direction determination device 10. When the area of the front surface of the sound source direction determination device 10 is 2 [square cm], which is an example of a size equal to or less than a predetermined value, the sound pressure of the sound in which the sound source exists in front of the sound source direction determination device 10 is −26 [dbov. ]. Further, the sound pressure of the sound in which the sound source exists above the sound source direction determination device 10 is −29 [dBov]. Therefore, the sound pressure difference between the sound pressure of the sound from the sound source existing in front of the sound source direction determination device 10 and the sound pressure of the sound from the sound source existing above is 3 [dB].

On the other hand, when the area of the front surface of the sound source direction determination device 10 is 63 [square cm], which is an example of a size larger than a predetermined value, the sound pressure of the sound in which the sound source exists in front of the sound source direction determination device 10 is -24. [DBov]. Further, the sound pressure of the sound in which the sound source exists above the sound source direction determination device 10 is −30 [dBov]. Therefore, the sound pressure difference between the sound pressure of the sound from the sound source existing in front of the sound source direction determination device 10 and the sound pressure of the sound from the sound source existing above is 6 [dB].

That is, when the area of the front surface of the sound source direction determination device 10 is 2 [square cm], the sound pressure difference depending on the direction of the sound source is larger when the area is 63 [square cm], and the direction of the sound source can be easily determined. .. This is because when the area of the front surface is larger than a predetermined value, the sound generated in front of the sound source direction determination device 10 is sufficiently reflected.

The predetermined value may be, for example, 1000 times the cross-sectional area of the sound path. That is, the diameter of the microphone hole of the second microphone 12 is, for example, 0.5 [mm], and the diameter 1 of the second sound path 12R is twice the diameter of the microphone hole of the second microphone 12. If it has a circular cross section of [mm], it may have an area larger than about 785 [mm2]. For example, the second sound path 12R may have the same diameter from one end to the other end, or may gradually decrease in diameter from one end to the other end. Further, the second sound path may have, for example, a rectangular cross section.

The length from one end to the other end of the second sound path 12R may be, for example, 3 [mm], but may be longer or shorter than 3 [mm]. Further, the second sound path 12R may be orthogonal to the front surface of the housing 18, or the second sound path 12R and the front surface of the housing 18 may intersect at an angle other than 90 [degrees]. ..

6A and 6B show the sound pressure acquired by the first microphone 11 when the sound source is above the sound source direction determination device 10 and when the sound source is in front of the sound source direction determination device 10. FIG. 6A illustrates a case where the sound source is located above the sound source direction determination device 10.

Since the length of the upper surface of the housing 18 in the front-rear direction is short and the area of the upper surface is equal to or less than a predetermined value, even when the sound source is above the sound source direction determination device 10, the sound is reflected and diffracted as illustrated in FIG. 4A. I can't expect to get the sound. Therefore, the first sound path 11R is provided with a bent portion 11K. Since the first sound path 11R has a bent portion 11K, the sound from above does not reach the first microphone 11 directly, is diffracted by the bent portion 11K of the first sound path 11R, and is acquired by the first microphone 11. To.

FIG. 6B illustrates a case where the sound source is located in front of the sound source direction determination device 10. The sound is diffracted by the opening 11O, which is an example of the first diffracting portion, further diffracted by the bent portion 11K, and acquired by the first microphone 11.

FIG. 7 shows the sound pressure of the sound acquired by the first microphone 11 when the sound source is above the sound source direction determination device 10, and the first microphone 11 when the sound source is in front of the sound source direction determination device 10. The difference in sound pressure from the sound pressure of the sound acquired in is illustrated. The solid line represents the sound pressure [dB] of the sound acquired by the first microphone 11 when the sound source is above the sound source direction determination device 10, and the broken line represents the sound source in front of the sound source direction determination device 10. In this case, it represents the sound pressure [dB] of the sound acquired by the first microphone 11.

That is, the vertical distance between the solid line and the broken line is the sound pressure of the sound acquired by the first microphone 11 when the sound source is above the sound source direction determination device 10, and the sound source is the sound source direction determination device 10. It represents the sound pressure difference from the sound pressure of the sound acquired by the first microphone 11 when it exists in front of. The horizontal axis of the graph of FIG. 7 is the frequency [Hz], and the sound pressure difference tends to be smaller as the frequency is lower and larger as the frequency is higher. That is, there are cases where the sound source has the number of diffractions of once above the sound source direction determination device 10 and cases where the sound source having the number of diffractions twice exists in front of the sound source direction determination device 10. The sound pressure difference becomes more remarkable as the frequency becomes higher.

Ｎは、フレネル数であり、（２）式で表される。
Ｎ＝δ／（λ／２）
＝δ・ｆ／１６５ …（２）
δは、回折経路と直接経路との経路差［ｍ］であり、λは音の波長［ｍ］であり、ｆは音の周波数［Ｈｚ］であり、音速（＝λ×ｆ）を３３０［ｍ／秒］とした場合である。即ち、図７のグラフにも表されるように、周波数ｆが高いほど、回折による減音量Ｒは大きくなる傾向を有する。したがって、本実施形態では、音源の方向を判定する際に、音の高域成分の音圧差を使用する。 The volume reduction R [dB] due to diffraction is expressed by, for example, Eq. (1).

N is a Fresnel number and is expressed by the equation (2).
N = δ / (λ / 2)
= Δ ・ f / 165 ... (2)
δ is the path difference [m] between the diffraction path and the direct path, λ is the wavelength of sound [m], f is the frequency of sound [Hz], and the speed of sound (= λ × f) is 330 [. m / sec]. That is, as shown in the graph of FIG. 7, the higher the frequency f, the larger the volume reduction R due to diffraction tends to be. Therefore, in the present embodiment, the difference in sound pressure of the high frequency component of the sound is used when determining the direction of the sound source.

The first sound path 11R has a circular cross section with a diameter of 1 [mm], which is twice the diameter of the microphone hole when the diameter of the microphone hole of the first microphone 11 is 0.5 [mm]. You may be doing it. For example, the first sound path 11R may have the same diameter from one end to the other end, or may gradually decrease in diameter from one end to the other end.

The diameter of the first sound path 11R gradually decreases from one end to the bent portion 11K, and may have the same diameter from the bent portion 11K to the other end. Further, the first sound path 11R may have, for example, a rectangular cross section.

The length from one end of the first sound path 11R to the bent portion 11K and the length from the bent portion 11K to the other end may be, for example, 3 [mm], but is more than 3 [mm]. It may be long or short. Further, from one end of the first sound path 11R to the bent portion 11K may be orthogonal to the upper surface of the housing 18, and the first sound path 11R and the upper surface of the housing 18 are other than 90 [degrees]. It may intersect at an angle. Further, the bent portion 11K to the other end of the first sound path 11R may be orthogonal to the bent portion 11K from the one end portion, or may intersect at an angle other than 90 [degrees].

Further, the periphery of the first microphone 11 is surrounded by the side wall except for the portion where the other end of the first sound path 11R and the side wall are connected, and there is no gap between the other end and the side wall. Further, the periphery of the second microphone 12 is surrounded by the side wall except for the portion where the other end of the second sound path 12R and the side wall are connected, and there is no gap between the other end and the side wall. The upper surface and the front surface of the housing 18 are orthogonal to each other. However, this embodiment is not limited to the example in which the upper surface and the front surface of the housing 18 are orthogonal to each other, and the upper surface and the front surface of the housing 18 may intersect at an angle other than 90 [degrees].

FIG. 8 is used to illustrate an outline of the sound source direction determination process performed by the determination unit 13 of the first embodiment. The time frequency conversion unit 13A converts the sound signal corresponding to the sound acquired by the first microphone 11 installed as illustrated in FIG. 3 into time frequency. Similarly, the time frequency conversion unit 13B converts the sound signal corresponding to the sound acquired by the second microphone 12 installed as illustrated in FIG. 3 into time frequency. For example, FFT (Fast Fourier Transformation) is used for time-frequency conversion.

As described above, the sound pressure difference between the sound pressure of the sound acquired by the first microphone 11 and the sound pressure of the sound acquired by the second microphone 12 appears remarkably in the high frequency component. Therefore, the high frequency sound pressure difference calculation unit 13C calculates the average value of the sound pressure difference for each frequency band at a frequency higher than a predetermined frequency as the high frequency sound pressure difference. The sound source direction determination unit 13D determines the position of the sound source based on the high-frequency sound pressure difference calculated by the high-frequency sound pressure difference calculation unit 13C.

Specifically, the high frequency sound pressure difference calculation unit 13C calculates the spectral power pow1 [bin] of the sound signal corresponding to the sound acquired by the first microphone 11 by the equation (3), and acquires it by the second microphone 12. The spectral power pow2 [bin] of the sound signal corresponding to the sound is calculated by Eq. (4).
pow1 [bin] = re1 [bin] ² + im1 [bin] ² … (3)
pow2 [bin] = re2 [bin] ² + im2 [bin] ² … (4)
bin = 0, ..., F-1, where F is the number of frequency bands, for example 256. re1 [bin] is the real part of the frequency spectrum of the frequency band bin acquired when the sound signal of the sound acquired by the first microphone 11 is time-frequency converted. Further, im1 [bin] is an imaginary part of the frequency spectrum of the frequency band bin acquired when the sound signal of the sound acquired by the first microphone 11 is time-frequency converted.

re2 [bin] is the real part of the frequency spectrum of the frequency band bin acquired when the sound signal of the sound acquired by the second microphone 12 is time-frequency converted. Further, im2 [bin] is an imaginary part of the frequency spectrum of the frequency band bin acquired when the sound signal of the sound acquired by the second microphone 12 is time-frequency converted.

高域音圧差d_powは、音圧の相違の一例であり、スペクトルパワーpow1[i]の対数から、スペクトルパワーpow2[i]の対数を減算した値の平均値である。ｓは、高域の下限周波数帯域数であり、例えば、９６であってよい。音信号のサンプリング周波数が１６［ｋＨｚ］であり、ｓ＝９６である場合、高域とは３０００［Ｈｚ］～８［ｋＨｚ］である。 Next, the high frequency sound pressure difference d_pow is calculated by the equation (5).

The high-frequency sound pressure difference d_pow is an example of the difference in sound pressure, and is the average value obtained by subtracting the logarithm of the spectral power pow2 [i] from the logarithm of the spectral power pow1 [i]. s is the lower limit frequency band number in the high frequency range, and may be 96, for example. When the sampling frequency of the sound signal is 16 [kHz] and s = 96, the high frequency range is 3000 [Hz] to 8 [kHz].

FIG. 9 illustrates the determination criteria and determination results of the sound source direction determination unit 13D. Comparing the high-frequency sound pressure difference d_pow with the first threshold value, which is a positive value, if the high-frequency sound pressure difference d_pow is larger than the first threshold value, the sound source is located at a position facing the upper surface of the housing 18, that is, above. Is determined. Further, when the high frequency sound pressure difference d_pow is compared with the second threshold value which is a negative value and the high frequency sound pressure difference d_pow is smaller than the second threshold value, the sound source is located at a position facing the front surface of the housing 18, that is, in front. It is determined that it exists in.

Further, as illustrated in FIG. 9, when the high frequency sound pressure difference d_pow is equal to or higher than the second threshold value and equal to or lower than the first threshold value, it is determined that the sound source direction cannot be determined. The first threshold value may be, for example, 1.5 [dB], and the second threshold value may be, for example, −1.5 [dB].

In addition, when acquiring the high frequency sound pressure difference d_pow, since the spectral power of the second microphone 12 having the opening 12O on the front surface of the housing 18 is used as a reference in the equation (5), it is illustrated in FIG. The judgment result is good. However, as illustrated in the equation (6), when the high frequency sound pressure difference d_pow is acquired with reference to the spectral power of the first microphone 11 having the opening 11O on the upper surface of the housing 18, the determination result is different.

Comparing the high-frequency sound pressure difference d_pow with the first threshold value, which is a positive value, if the high-frequency sound pressure difference d_pow is larger than the first threshold value, the sound source is located at a position facing the front surface of the housing 18, that is, in front of the housing 18. Is determined. Further, when the high frequency sound pressure difference d_pow is compared with the second threshold value which is a negative value and the high frequency sound pressure difference d_pow is smaller than the second threshold value, the sound source is located at a position facing the upper surface of the housing 18, that is, above. It is determined that it exists in.

It should be noted that the equations (5) and (6) for acquiring the high frequency sound pressure difference are examples, and the present embodiment is not limited thereto. Further, regarding an example of using a high-frequency sound pressure difference, which is a difference between the sound pressure of the high-frequency component of the sound acquired by the first microphone 11 and the sound pressure of the high-frequency component of the sound acquired by the second microphone 12. As described above, the present embodiment is not limited to this example.

The difference between the sound pressure of the predetermined frequency component of the sound acquired by the first microphone 11 and the sound pressure of the predetermined frequency component of the sound acquired by the second microphone 12 is used instead of the high frequency sound pressure difference. You may. The predetermined frequency component is an example of the first frequency component and may be a high frequency component, but a frequency at which a sound pressure difference remarkably appears between the first microphone 11 and the second microphone 12 depending on the direction of the sound source. It may be an ingredient. Further, the determination criteria and determination results in FIG. 9 are also examples, and the present embodiment is not limited to this example.

FIG. 10 illustrates the hardware configuration of the information processing terminal 1. The information processing terminal 1 includes a CPU (Central Processing Unit) 51, which is an example of a processor that is hardware, a primary storage unit 52, a secondary storage unit 53, and an external interface 54. The information processing terminal 1 also includes a first microphone 11, a second microphone 12, and a speaker 14C.

The CPU 51, the primary storage unit 52, the secondary storage unit 53, the external interface 54, the first microphone 11, the second microphone 12, and the speaker 14C are connected to each other via the bus 59.

The primary storage unit 52 is, for example, a volatile memory such as a RAM (Random Access Memory).

The secondary storage unit 53 includes a program storage area 53A and a data storage area 53B. The program storage area 53A is, for example, a sound source direction determination program for causing the CPU 51 to execute the sound source direction determination process, a speech translation program for causing the CPU 51 to execute the speech translation process based on the determination result of the sound source direction determination process, and the like. I remember the program of. The data storage area 53B stores sound signals corresponding to the sounds acquired from the first microphone 11 and the second microphone 12, intermediate data temporarily generated in the sound source direction determination process and the speech translation process, and the like.

The CPU 51 reads a sound source direction determination program from the program storage area 53A and deploys it to the primary storage unit 52. The CPU 51 operates as the determination unit 13 in FIG. 1 by executing the sound source direction determination program. The CPU 51 reads a speech translation program from the program storage area 53A and deploys it in the primary storage unit 52. The CPU 51 operates as the first translation unit 14A and the second translation unit 14B in FIG. 1 by executing the speech translation program. Programs such as a sound source direction determination program and a voice translation program are stored in a non-temporary recording medium such as a DVD (Digital Versatile Disc), read via a recording medium reading device, and expanded in the primary storage unit 52. May be good.

An external device is connected to the external interface 54, and the external interface 54 controls transmission / reception of various information between the external device and the CPU 51. For example, the speaker 14C may be an external device that is not included in the sound source direction determination device 10 and is connected via the external interface 54.

Next, the outline of the operation of the sound source direction determination device 10 will be described. The outline of the operation of the sound source direction determination device 10 is illustrated in FIG. For example, when the user turns on the power of the sound source direction determination device 10, the CPU 51 reads a sound signal for one frame at least. Specifically, one frame of sound signal corresponding to the sound acquired from the first microphone 11 (hereinafter referred to as the first sound signal) and one frame of sound signal acquired from the second microphone 12 The sound signal (hereinafter referred to as the second sound signal) and the sound signal are read. One frame may be, for example, 32 [msec] when the sampling frequency is 16 [kHz].

In step 102, the CPU 51 performs time-frequency conversion on each of the sound signals read in step 101. In step 103, the CPU 51 calculates the spectral power of each of the time-frequency-converted sound signals using the equations (3) and (4), and uses the equation (5) to calculate the high frequency sound. Calculate the pressure difference d_pow.

In step 104, the CPU 51 compares the high-frequency sound pressure difference d_pow calculated in step 103 with the first threshold value, and when the high-frequency sound pressure difference d_pow is larger than the first threshold value, the sound source is above the sound source direction determination device 10. It is determined that it exists, and the process proceeds to step 105. In step 105, the CPU 51 allocates the sound signal to the process of translating the sound signal from the second language to the first language, and proceeds to step 108. The distributed sound signal is translated from the second language to the first language by the existing voice translation processing technology, and is output as voice from the speaker 14C, for example.

When it is determined in step 104 that the high frequency sound pressure difference d_pow is equal to or less than the first threshold value, the CPU 51 compares the high frequency sound pressure difference d_pow with the second threshold value in step 106, and the high frequency sound pressure difference d_pow is the first. If it is smaller than the two threshold values, it is determined that the sound source exists in front of the sound source direction determination device 10. If the determination in step 106 is affirmed, that is, if it is determined that the sound source is in front of the sound source direction determination device 10, the CPU 51 proceeds to step 107. In step 107, the CPU 51 allocates the sound signal to the process of translating the sound signal from the first language to the second language, and proceeds to step 108. The distributed sound signal is translated from the first language to the second language by the existing voice translation processing technology, and is output as voice from the speaker 14C, for example.

If the determination in step 106 is denied, the CPU 51 proceeds to step 108. That is, when the high frequency sound pressure difference d_pow is equal to or less than the first threshold value and is equal to or greater than the second threshold value, it is determined that the sound source position cannot be determined, and the translation from the first language to the second language is also performed. It does not translate from two languages to the first language.

In step 108, the CPU 51 determines whether or not the sound source direction determination function of the sound source direction determination device 10 is turned off by, for example, a user's operation. When the determination in step 108 is denied, that is, when the sound source direction determination function is on, the CPU 51 proceeds to step 101, reads the sound signal of the next frame, and continues the sound source direction determination process. When the determination in step 108 is denied, that is, when the sound source direction determination function is off, the CPU 51 ends the sound source direction determination process.

The microphone installation portion of the present embodiment is provided with a first sound path and a second sound path inside. The first sound path is provided with a first opening opened on the first flat surface at one end, and sound propagates from the first opening. The second sound path is provided at one end with a second opening opened in the second flat surface intersecting with the first flat surface, and a second sound path through which sound propagates from the second opening is provided inside. .. The first microphone is installed at the other end of the first sound path, and the second microphone is installed at the other end of the second sound path. The determination unit determines the direction in which the sound source exists based on the difference in sound pressure. The difference in sound pressure is the first sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the second microphone. It is a difference from sound pressure.

In the present embodiment, the above configuration makes it possible to improve the accuracy of sound source direction determination using an omnidirectional microphone.

Further, in the present embodiment, the first flat surface and the second flat surface are orthogonal to each other, the area of the first flat surface is equal to or less than a predetermined value, and the area of the second flat surface is larger than the predetermined value. The first sound path has a first diffracting portion that diffracts sound in the first opening, and has a second diffracting portion that is a bending portion that diffracts sound in the middle, and the second sound path has a second diffracting portion. The second opening has a third diffracting part that diffracts sound.

In the present embodiment, according to the above configuration, even when the area of the flat surface provided with the opening of the sound path is equal to or less than a predetermined value capable of sufficiently reflecting the sound, the sound source direction determination using the omnidirectional microphone is performed. It is possible to improve the accuracy.

In this embodiment, the case where the area of the upper surface of the housing is equal to or less than the predetermined value and the area of the front surface of the housing is larger than the predetermined value is illustrated, but the area of the upper surface is larger than the predetermined value and the area of the front surface is large. It may be less than or equal to a predetermined value. In this case, the first sound path having an opening on the upper surface does not have a diffraction portion which is a bending portion, and the second sound path having an opening on the front surface has a diffraction portion which is a bending portion.

Although the case where the voice translation device 14 is included in the housing 18 of the sound source direction determination device 10, the present embodiment is not limited to this. For example, the voice translation device 14 may exist outside the housing 18 of the sound source direction determination device 10 and may be connected to the sound source direction determination device 10 via a wired connection or a wireless connection.

[Second Embodiment]
Next, an example of the second embodiment will be described. The description of the same configuration and operation as in the first embodiment will be omitted.

FIG. 12 illustrates a cross-sectional view taken along the cutting line 3-3 of FIG. 2A. In the second embodiment, as in the first embodiment, the area of the upper surface of the housing 18A of the sound source direction determination device 10A is a predetermined value or less, and the area of the front surface of the housing 18A of the sound source direction determination device 10A is a predetermined value. Greater.

In the second embodiment, the first sound path 11AR has a diffracting portion which is an example of the first diffracting portion for diffracting the sound in the opening 11AO, and is a bending portion 11AK for diffracting the sound in the middle. It has a diffractive part which is an example of two diffractive parts. Further, the second sound path 12AR has a diffracting portion which is an example of a third diffracting portion that diffracts the sound in the second opening 12AO, and a bending portion 12AK that diffracts the sound in the middle of the fourth diffracting portion. It has a diffractive part, which is an example.

The front surface of the housing 18A of the sound source direction determination device 10A has an area larger than a predetermined value as in the first embodiment, but unlike the first embodiment, the second sound path 12AR is a diffractive portion in the middle. It has a bent portion 12AK.

In the present embodiment, the above configuration makes it possible to improve the accuracy of sound source direction determination using an omnidirectional microphone by utilizing the sound reduction of a predetermined frequency component (for example, a high frequency component) by diffraction. ..

[Third Embodiment]
Next, an example of the third embodiment will be described. The description of the same configuration and operation as those of the first embodiment and the second embodiment will be omitted.

13A to 13C illustrate the appearance of the sound source direction determination device 10C of the third embodiment. 13A is the right side surface of the housing 18C which is an example of the first flat surface, FIG. 13B is the front surface of the housing 18C which is an example of the second flat surface, and FIG. 13C shows the sound source direction determination device 10C as the housing 18C. It is the figure which looked at the side connecting the front surface and the right side surface from the front. The arrow R in the figure indicates the right-hand side when the sound source direction determination device 10C is viewed from the front.

FIG. 14 illustrates a cross-sectional view taken along the cutting line 14-14 of FIG. 13A. In the third embodiment, the first sound path 11CR is provided with a first opening 11CO opened on the right side surface of the housing 18C at one end, and the second sound path 12CR is a second opening opened on the front surface of the housing 18C. An opening 12CO is provided at one end. The first microphone 11C is installed at the other end of the first sound path 11CR, and the second microphone 12C is installed at the other end of the second sound path 12CR.

Unlike the first embodiment and the second embodiment, neither the first sound path 11CR nor the second sound path 12CR has a bending portion which is a diffractive portion in the middle. This is because, in the third embodiment, both the front surface and the right side surface of the housing 18C have an area larger than a predetermined value capable of sufficiently reflecting sound. In the third embodiment, the first sound path 11CR has a diffraction section which is an example of the first diffraction section that diffracts the sound into the first opening 11CO, and the second sound path 12CR has the second opening 12CO. It has a diffractometer, which is an example of a second diffractometer that diffracts sound.

In the present embodiment, the above configuration makes it possible to improve the accuracy of sound source direction determination using an omnidirectional microphone by utilizing the sound reflected on the flat surface of the housing.

In the first to third embodiments, the sound source direction determination device may further have a third flat surface that intersects at least one of the first flat surface and the second flat surface. Further, a third opening opened on the third flat surface is provided at one end, a third sound path through which sound propagates from the third opening is provided inside the housing, and an omnidirectional third microphone is provided. It may be installed at the other end of the three sound paths.

The third sound path has a diffraction portion which is a bending portion in the middle when the area of the third flat surface is equal to or less than a predetermined value, and bends in the middle when the area of the third flat surface is larger than the predetermined value. It may or may not have a diffractive part which is a part. In this case, the sound pressure of a predetermined frequency component of the sound acquired by the microphone installed at the other end of the sound path having an opening in the flat surface intersecting the third flat surface and the sound pressure acquired by the third microphone. The direction in which the sound source exists is determined based on the difference from the sound pressure of a predetermined frequency component of the sound.

In this embodiment, an example is described in which the sound signal whose sound source direction is determined is translated from the first language to the second language or from the second language to the first language by the voice translation device 14 depending on the sound source direction. However, the present embodiment is not limited to this. The speech translation device 14 may include, for example, only one of the first translation unit 14A and the second translation unit 14B.

Further, the information processing terminal 1 may include a conference support device instead of the speech translation device 14. The conference support device switches, for example, a camera, a microphone, a display, and the like based on the determined sound source direction and sound signal. Further, the information processing terminal 1 may include a drive support device instead of the speech translation device 14. If the determined sound source direction is the driver's seat side, the drive support device provides driving support based on, for example, a sound signal, and if the determined sound source direction is the passenger seat side, for example, based on the sound signal. To provide entertainment such as playing music or video.

The information processing terminal including the sound source direction determination device may be a dedicated terminal for determining the sound source direction, but the sound source direction determination device may be incorporated in an existing terminal by hardware and software. Existing terminals are, for example, smartphones, tablets, wearable devices, or navigation systems. Further, at least a part of the hardware or software of the sound source direction determination device may be incorporated in the existing terminal, and the sound source direction determination device may be connected to the existing terminal as an external device.

The order of processing in the flowchart in FIG. 11 is an example, and the present embodiment is not limited to the order of processing.

[Fourth Embodiment]
Next, an example of the fourth embodiment will be described. The description of the same configuration and operation as those of the first to third embodiments will be omitted.

In the fourth embodiment, the sound source direction determination device 10D includes a determination unit 13'instead of the determination unit 13 of the sound source direction determination device 10 of FIG. FIG. 15 exemplifies an outline of the sound source direction determination process performed by the determination unit 13'of the fourth embodiment.

The determination unit 13'of FIG. 15 is different from the determination unit 13 of FIG. 8 in that it further includes the phase difference calculation unit 13C'. That is, the fourth embodiment is different from the first embodiment in that the normalized phase difference is used in addition to the high frequency sound pressure difference.

As illustrated in FIG. 16A, the distance until the sound US1 from above reaches the first microphone 11D is shorter than the distance until the sound US2 from above reaches the second microphone 12D. For reference, compare the arrow USD1 from the reference line RL1 to the first microphone 11D until the sound US1 reaches, and the arrow USD2 from the reference line RL1 to the second microphone 12D until the sound US2 reaches. Then it is clear.

That is, the time until the sound from above reaches the first microphone 11D and the time until the sound from above reaches the microphone 12D are different. Therefore, the phase when the sound from above reaches the first microphone 11D and the phase when the sound from above reaches the second microphone 12D are different.

Further, as illustrated in FIG. 16B, the distance until the sound FS1 from the front reaches the first microphone 11D is longer than the distance until the sound FS2 from the front reaches the second microphone 12D. It is clear from the arrow FSD1 until the sound FS1 reaches the first microphone 11D from the reference line RL2 described for reference.

That is, the time until the sound from the front reaches the first microphone 11D and the time until the sound from the front reaches the microphone 12D are different. Therefore, the phase when the sound from the front reaches the first microphone 11D and the phase when the sound from the front reaches the second microphone 12D are different. In the fourth embodiment, the phase difference is used to determine the sound source direction.

The phase difference calculation unit 13C'in FIG. 15 calculates the difference between the first phase, which is the phase of the sound acquired by the first microphone 11D, and the second phase, which is the phase of the sound acquired by the second microphone 12D. do. Specifically, the phase difference calculation unit 13C'calculates the normalized phase difference a_phase, which is an example of the phase difference, by the equation (7).

The normalized phase difference a_phase is the average value of the values obtained by normalizing the phase difference phase [j] of the jth frequency band with the normalization coefficient C_n [j]. j = ss, ..., ee, ss is the lower limit frequency band number for normalization phase difference calculation, ee is the upper limit frequency band number for normalization phase difference calculation, and ss and ee are included in the above bin. It is a numerical value (bin = 0, ..., ss, ..., ee, ..., F-1).

The phase difference phase [j] is calculated by Eq. (8).
phase [j] = atan (phase_ im [j] / phase_ re [j])… (8)
phase_re [j] = re1 [j] * re2 [j] + im1 [j] * im 2 [j], phase_im [j] = im1 [j] * re2 [j]-re1 [j] * im2 [ j], where atan stands for arctangent.

The normalization coefficient C_n [j] is calculated by Eq. (9).
C_n [j] = λ [j] / λ_c… (9)
λ [j] = C / f_j, λ [j] is the wavelength corresponding to the number of frequency bands j, C is the sound velocity, f_j is the frequency corresponding to the number of frequency bands j, and λ_c is the reference frequency. The wavelength of the sound. The reference frequency may be, for example, 8 [kHz], which is the upper limit frequency, when the sampling frequency is 16 [kHz].

The frequency corresponding to the upper limit frequency band number ee for normalization phase difference calculation may be, for example, C / 2L. L is the distance between the first microphone 11 and the second microphone 12. The frequency corresponding to the lower limit frequency band number ss for normalized phase difference calculation may be, for example, 100 Hz.

The upper limit frequency band number ee and the lower limit frequency band number ss for normalized phase difference calculation may be set to such an extent that the influence of noise does not increase and appropriate detection of the phase change is possible. Since the power of sound decreases as the frequency increases, the signal-to-noise ratio decreases as the frequency increases, and the influence of noise increases. In addition, if the frequency is set to a low frequency so that the influence of noise does not become large, the phase change of the low frequency sound is slower than that of the high frequency sound, and it becomes difficult to properly detect the phase change in a short time. ..

The normalized phase difference a_phase calculated by the above equation (7) becomes a positive value when the sound source is located above, that is, when the first microphone 11D is closer to the sound source than the second microphone 12D. On the other hand, when the sound source is present in front, that is, when the first microphone 11D is farther from the sound source than the second microphone 12D, a negative value is obtained. The sign of the normalized phase difference differs depending on whether the first microphone 11D or the second microphone 12D is used as a reference. Further, the method for obtaining the normalized phase difference is not limited to the above equation (7).

Next, the outline of the operation of the sound source direction determination device 10D will be described. The outline of the operation of the sound source direction determination device 10D is illustrated in FIG. 17A. The difference between FIGS. 11 and 17A is that steps 103, 104 and 106 in FIG. 11 are replaced with steps 103, 103B, 104, 104B and 106 in FIG. 17A.

That is, in FIG. 17A, the CPU 51 calculates the high frequency sound pressure difference as described above in step 103, and calculates the normalized phase difference a_phase using the equation (7) in step 103B. The CPU 51 determines in step 104 whether or not the high frequency sound pressure difference is larger than the positive first threshold value, and if the determination in step 104 is affirmed, in step 104B, the normalized phase difference is a positive third threshold value. Determine if it is greater than. If the determination in step 104B is affirmed, it is determined that the sound source exists above, and the process proceeds to step 105.

When the determination in step 104 is denied, that is, when the high frequency sound pressure difference is equal to or less than the positive first threshold value, it is determined that the sound source does not exist above, and the CPU 51 determines in step 106 that the high frequency sound pressure difference is negative. It is determined whether or not it is smaller than the second threshold value of. If the determination in step 106 is affirmed, or if the determination in step 104B is denied, that is, if the normalized phase difference is equal to or less than the positive third threshold value, it is determined that the sound source exists in front, and the CPU 51 determines. , Step 107.

If the determination in step 106 is denied, that is, if the high frequency sound pressure difference is equal to or greater than the negative second threshold value, it is determined that the determination of the sound source direction is impossible, and the CPU 51 proceeds to step 108. The positive third threshold may be, for example, 3.0 [rad].

Note that this embodiment is not limited to determining the sound source direction in steps 104, 104B, and 106 of FIG. 17A. As illustrated in FIGS. 17B to 17F, the sound source direction may be determined by combining the determination of the high frequency sound pressure difference and the determination of the normalized phase difference, or normalized as illustrated in FIG. 17G. The sound source direction may be determined by determining the phase difference.

The difference between FIGS. 11 and 17B is that steps 103, 104 and 106 in FIG. 11 are replaced with steps 103, 103B, 104, 104B, 106 and 106B in FIG. 17B.

That is, in FIG. 17B, the CPU 51 determines in step 104 whether or not the high frequency sound pressure difference is larger than the positive first threshold value, and if the determination in step 104 is affirmed, the normalized phase difference is determined in step 104B. It is determined whether or not it is larger than the positive third threshold value. If the determination in step 104B is affirmed, it is determined that the sound source exists above, and the process proceeds to step 105.

When the determination in step 104 is denied, that is, when the high frequency sound pressure difference is equal to or less than the positive first threshold value, it is determined that the sound source does not exist above, and the CPU 51 determines in step 106 that the high frequency sound pressure difference is negative. It is determined whether or not it is smaller than the second threshold value of. If the determination in step 106 is affirmed, or if the determination in step 104B is denied, the CPU 51 determines in step 106B whether the normalized phase difference is smaller than the negative fourth threshold value. If the determination in step 106B is affirmed, it is determined that the sound source exists in front, and the process proceeds to step 107.

If the determination in step 106 or step 106B is denied, that is, if the high frequency sound pressure difference is equal to or greater than the negative second threshold value, or if the normalized phase difference is equal to or greater than the negative fourth threshold value, the sound source direction It is determined that the determination is not possible, and the process proceeds to step 108.

The difference between FIGS. 11 and 17C is that steps 103, 104 and 106 in FIG. 11 are replaced with steps 103, 103B, 104, 106 and 106B in FIG. 17C.

That is, in FIG. 17C, the CPU 51 determines in step 104 whether or not the high frequency sound pressure difference is larger than the positive first threshold value, and if the determination in step 104 is affirmed, determines that the sound source exists above. Proceed to step 105.

When the determination in step 104 is denied, that is, when the high frequency sound pressure difference is equal to or less than the positive first threshold value, it is determined that the sound source does not exist above, and the CPU 51 determines in step 106 that the high frequency sound pressure difference is negative. It is determined whether or not it is smaller than the second threshold value of. If the determination in step 106 is affirmed, the CPU 51 determines in step 106B whether the normalized phase difference is smaller than the negative fourth threshold value. If the determination in step 106B is affirmed, it is determined that the sound source exists in front, and the process proceeds to step 107.

If the determination in step 106 or step 106B is denied, that is, if the high frequency sound pressure difference is equal to or greater than the negative second threshold value, or if the normalized phase difference is equal to or greater than the negative fourth threshold value, the sound source direction It is determined that the determination is not possible, and the process proceeds to step 108.

The difference between FIGS. 11 and 17D is that steps 103, 104 and 106 in FIG. 11 are replaced with steps 103, 103B, 104B, 104 and 106B in FIG. 17D.

That is, in FIG. 17D, the CPU 51 determines in step 104B whether or not the normalized phase difference is larger than the positive third threshold value. If the determination in step 104B is affirmed, that is, if the normalized phase difference is greater than the positive third threshold, the CPU 51 determines in step 104 whether the high frequency sound pressure difference is greater than the positive first threshold. .. If the determination in step 104 is affirmed, it is determined that the sound source exists above, and the CPU 51 proceeds to step 105.

When the determination in step 104B is denied, that is, when the normalized phase difference is equal to or less than the positive third threshold value, it is determined that the sound source does not exist above, and the CPU 51 determines in step 106B that the normalized phase difference is negative. It is determined whether or not it is smaller than the fourth threshold value of. When the determination in step 106B is affirmed, or when the determination in step 104 is denied, that is, when the normalized phase difference is equal to or greater than the negative fourth threshold value, or when the high frequency sound pressure difference is positive first. If it is equal to or less than the threshold value, it is determined that the sound source exists in front, and the process proceeds to step 107.

If the determination in step 106B is denied, that is, if the normalized phase difference is equal to or greater than the negative fourth threshold value, it is determined that the determination of the sound source direction is impossible, and the process proceeds to step 108.

The difference between FIGS. 11 and 17E is that steps 103, 104 and 106 in FIG. 11 are replaced with steps 103, 103B, 104B, 104, 106B and 106 in FIG. 17E.

That is, in FIG. 17E, the CPU 51 determines in step 104B whether or not the normalized phase difference is larger than the positive third threshold value. If the determination in step 104B is affirmed, that is, if the normalized phase difference is greater than the positive third threshold, the CPU 51 determines in step 104 whether the high frequency sound pressure difference is greater than the positive first threshold. .. If the determination in step 104 is affirmed, it is determined that the sound source exists above, and the CPU 51 proceeds to step 105.

When the determination in step 104B is denied, that is, when the normalized phase difference is equal to or less than the positive third threshold value, it is determined that the sound source does not exist above, and the CPU 51 determines in step 106B that the normalized phase difference is negative. It is determined whether or not it is smaller than the fourth threshold value of. When the determination in step 106B is affirmed, or when the determination in step 104 is denied, that is, when the normalized phase difference is smaller than the negative fourth threshold value, or when the high frequency sound pressure difference is positive first threshold value. If the following is true, the CPU 51 proceeds to step 106. In step 106, the CPU 51 determines whether or not the high frequency sound pressure difference is smaller than the negative second threshold value. If the determination in step 106 is affirmative, that is, if the high frequency sound pressure difference is smaller than the negative second threshold value, it is determined that the sound source exists in front, and the process proceeds to step 107.

When the determination in step 106B is denied, or when the determination in step 106 is denied, that is, when the normalized phase difference is equal to or greater than the negative fourth threshold value, or when the high frequency sound pressure difference is negative second. If it is equal to or higher than the threshold value, it is determined that the sound source direction cannot be determined. If it is determined that the sound source direction cannot be determined, the CPU 51 proceeds to step 108.

The difference between FIGS. 11 and 17F is that steps 103, 104 and 106 in FIG. 11 are replaced with steps 103, 103B, 104B, 106B and 106 in FIG. 17F.

That is, in FIG. 17F, the CPU 51 determines in step 104B whether or not the normalized phase difference is larger than the positive third threshold value. If the determination in step 104B is affirmative, that is, if the normalized phase difference is larger than the positive third threshold value, it is determined that the sound source exists above, and the process proceeds to step 105.

When the determination in step 104B is denied, that is, when the normalized phase difference is equal to or less than the positive third threshold value, it is determined that the sound source does not exist above, and the CPU 51 determines in step 106B that the normalized phase difference is negative. It is determined whether or not it is smaller than the fourth threshold value of. If the determination in step 106B is affirmed, that is, if the normalized phase difference is smaller than the negative fourth threshold value, the CPU 51 determines in step 106 whether the high frequency sound pressure difference is smaller than the negative second threshold value. .. If the determination in step 106 is affirmative, that is, if the high frequency sound pressure difference is smaller than the negative second threshold value, it is determined that the sound source exists in front, and the process proceeds to step 107.

When the determination in step 106B is denied, or when the determination in step 106 is denied, that is, when the normalized phase difference is equal to or greater than the negative fourth threshold value, or when the high frequency sound pressure difference is negative second. If it is equal to or higher than the threshold value, it is determined that the sound source direction cannot be determined. If it is determined that the sound source direction cannot be determined, the CPU 51 proceeds to step 108.

The difference between FIGS. 11 and 17G is that steps 103, 104 and 106 in FIG. 11 are replaced with steps 103B, 104B and 106B in FIG. 17G.

That is, in FIG. 21G, the CPU 51 calculates the normalized phase difference in step 103B. In step 104B, the CPU 51 determines whether or not the normalized phase difference is larger than the positive third threshold value. If the determination in step 104B is affirmative, that is, if the normalized phase difference is larger than the positive third threshold value, it is determined that the sound source exists above, and the process proceeds to step 105.

When the determination in step 104B is denied, that is, when the normalized phase difference is equal to or less than the positive third threshold value, it is determined that the sound source does not exist above, and the CPU 51 determines in step 106B that the normalized phase difference is negative. It is determined whether or not it is smaller than the fourth threshold value of. If the determination in step 106B is affirmative, that is, if the normalized phase difference is smaller than the negative fourth threshold value, it is determined that the sound source exists in front, and the process proceeds to step 107.

If the determination in step 106B is denied, that is, if the normalized phase difference is equal to or greater than the negative fourth threshold value, it is determined that the determination of the sound source direction is impossible, and the process proceeds to step 108. The order of processing of the flowcharts in FIGS. 17A to 17G is an example, and the present embodiment is not limited to the order of the processing.

In the fourth embodiment, since the first sound path 11DR has the bent portion 11DK, the distance between the first microphone 11D and the second microphone 12D can be set from the case where the sound path does not have the bent portion. Can also be lengthened. As a result, the difference in the moving distance of the sound with respect to the wavelength of the sound having a predetermined frequency can be lengthened, and the fluctuation of the phase difference can be easily detected.

Although examples of the first sound path 11DR having the bent portion 11DK are shown in FIGS. 16A and 16B, the present embodiment is not limited to this. In this embodiment, when each of the two sound paths has a bending portion as in the second embodiment, neither of the two sound paths includes a bending portion as in the third embodiment. It is applicable even in the case.

The sound source direction determination device of the present embodiment includes a microphone installation unit, a first microphone, and a second microphone. The microphone installation portion is provided with a first opening opened on the first flat surface at one end, and is provided on a first sound path through which sound propagates from the first opening and on a second flat surface intersecting the first flat surface. A second opening is provided at one end, and a second sound path through which sound propagates from the second opening is provided inside. The first microphone is an omnidirectional microphone installed at the other end of the first sound path, and the second microphone is an omnidirectional microphone installed at the other end of the second sound path.

The determination unit of the sound source direction determination device of the present embodiment is based on at least one of the difference in sound pressure between the first sound pressure and the second sound pressure and the difference in phase between the first phase and the second phase. Determine the direction in which the sound source exists. The first sound pressure is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure is the sound pressure of the first frequency component of the sound acquired by the second microphone. The first phase is the phase of the second frequency component of the sound acquired by the first microphone, and the second phase is the phase of the second frequency component of the sound acquired by the second microphone.

In the present embodiment, this makes it possible to appropriately determine the sound source direction even when it is difficult to determine the sound source direction only due to the difference in sound pressure.

(Explanation of Fourth Embodiment)
In FIG. 22A, there is a gap on the back surface of the sound source direction determination device 10D, that is, a gap between an object BO such as clothes of a user wearing the sound source direction determination device 10D and the back surface of the sound source direction determination device 10D. Illustrates the case where is present. When the sound source is in front, the sound pressure of the sound acquired by the first microphone 11D is smaller than the sound pressure of the sound acquired by the second microphone 12D. This is because the sound pressure of the first microphone 11D is attenuated by diffraction, and the sound that is not diffracted by the first opening 11DO is diffracted at the entrance of the gap and passes through the gap, so that the sound does not reach the first microphone 11D.

In FIG. 22B, there is no gap on the back surface of the sound source direction determination device 10D, that is, a gap between the object BO such as the clothes of the user wearing the sound source direction determination device 10D and the back surface of the sound source direction determination device 10D. Illustrate the case where is not present. When the sound source is in front, the sound pressure of the sound acquired by the first microphone 11D is higher than the sound pressure of the sound acquired by the second microphone 12D. When the sound source is in front, it is difficult to determine the sound source direction even if the sound pressure of the sound acquired by the first microphone 11D is smaller than the sound pressure of the sound acquired by the second microphone 12D. , The sound pressure of the sound acquired by the first microphone 11D and the sound pressure of the sound acquired by the second microphone 12D are close to each other. This is because the sound passing through the gap in FIG. 22A is diffracted by the first opening 11DO in FIG. 22B and reaches the first microphone 11D.

FIG. 23A illustrates the high frequency sound pressure difference between the first microphone 11D and the second microphone 12D. The first block UGN from the left shows the first sound pressure difference when the sound source exists above and the void does not exist. The second block UG from the left shows the second sound pressure difference when the sound source is above and the void is present. The second sound pressure difference is smaller than the first sound pressure difference because there is sound passing through the void.

The fourth block FG from the left shows the fourth sound pressure difference when the sound source is in front and the gap is present. Since the sound pressure of the sound acquired by the second microphone 12D is higher than the sound pressure of the sound acquired by the first microphone 11D, the fourth sound pressure difference becomes a negative value.

On the other hand, the third block FGN from the left shows the third sound pressure difference when the sound source exists in front and the void does not exist. Since there is no gap, the sound passing through the gap also reaches the first microphone 11D when the gap exists, so that the sound pressure of the sound acquired by the first microphone 11D is the sound pressure of the sound acquired by the second microphone 12D. Will be greater than and will be a positive value. Even if the sound pressure of the sound acquired by the first microphone 11D is smaller than the sound pressure of the sound acquired by the second microphone 12D, the sound pressure of the sound acquired by the first microphone and the sound acquired by the second microphone The difference in sound pressure becomes so small that it is difficult to determine the direction of the sound source because it is close to the sound pressure of. The first sound pressure difference is, for example, 4.8 [dB], the second sound pressure difference is, for example, 1.8 [dB], and the third sound pressure difference is, for example, 1.2 [dB]. , The fourth sound pressure difference is, for example, −0.9 [dB].

Therefore, if there is no gap on the back surface of the sound source direction determination device 10D, it may be difficult to determine the sound source direction based on the high frequency sound pressure difference. That is, it may be difficult to set an appropriate threshold value for determining the sound source direction. For example, if the value of the positive first threshold value for determining whether or not the sound source exists above is set large, the sound source has a high-frequency sound pressure difference in front when the sound source represented by the block UG exists above. There is a risk of determining that it is a high-frequency sound pressure difference. On the other hand, if the value of the positive first threshold value is set small, there is a possibility that the high-frequency sound pressure difference when the sound source represented by the block FGN exists in front is determined to be the high-frequency sound pressure difference in which the sound source exists above. Occurs.

FIG. 23B illustrates a normalized phase difference between the phase of the sound acquired by the first microphone 11D and the phase of the sound acquired by the second microphone 12D. The first block UG from the left shows the first phase difference when the sound source exists above and the void does not exist. The second block UGN from the left shows the second phase difference when the sound source is above and the void is present.

The third block FG from the left shows the third phase difference when the sound source is in front and the void is present. The fourth block FGN from the left shows the phase difference when the sound source exists in front and the void does not exist. That is, when the sound source is present above, regardless of the presence or absence of a gap on the back surface of the sound source direction determination device 10, the phase difference shows a positive value. Also, when the sound source is in front, the phase difference shows a negative value. The first phase difference is, for example, 6.1 [rad], the second phase difference is, for example, 6.0 [rad], and the third phase difference is, for example, -2.5 [rad]. Yes, the fourth phase difference is, for example, −1.4 [rad]. Therefore, regardless of whether or not there is a gap on the back surface of the sound source direction determination device 10, it is relatively easy to set an appropriate threshold value for determining the sound source direction.

When the sound source is above the sound source direction determination device 10D, the sound reaches the first microphone 11D before reaching the second microphone 12D. Further, when the sound source is in front of the sound source direction determination device 10D, the sound reaches the second microphone 12D before reaching the first microphone 11D. Therefore, the phase difference can be used to determine the sound source direction. Further, since the phase difference is not so affected by the absolute sound pressure, it is possible to obtain an appropriate phase difference even if the absolute sound pressure fluctuates depending on the presence or absence of a gap on the back surface of the sound source determination device 10D.

[Fifth Embodiment]
Next, an example of the fifth embodiment will be described. The description of the same configuration and operation as those of the first to fourth embodiments will be omitted. In the fifth embodiment, the threshold value for determining the direction of the sound source is adjusted based on the sound signal corresponding to the sound spoken by the user and the dialogue partner.

FIG. 24 illustrates an outline of the sound source direction determination process of the fifth embodiment performed by the determination unit 13 ”instead of the determination unit 13 of the sound source direction determination device 10 of FIG. 1. The time-frequency conversion unit 85A1 is the first. 1 The sound signal corresponding to the sound acquired by the microphone 11 is time-frequency-converted, and the time-frequency conversion unit 85A2 converts the sound signal corresponding to the sound acquired by the second microphone 12 into time-frequency.

The utterance section detection unit 85B1 detects the utterance section of the sound signal corresponding to the sound acquired by the first microphone 11, and the utterance section detection unit 85B2 detects the sound signal corresponding to the sound acquired by the second microphone 12. Detects the speech section. Existing methods can be applied to detect the utterance section.

The phase calculation unit 85C1 calculates the phase of the sound signal corresponding to the sound acquired by the first microphone 11 by using the detected sound signal in the utterance section. The phase calculation unit 85C2 calculates the phase of the sound signal corresponding to the sound acquired by the second microphone 12 by using the detected sound signal in the utterance section. The average phase difference calculation unit 85D calculates the phase difference using the calculated phase, and calculates the phase difference average value which is the average value of the phase difference in the utterance section.

The past utterance phase difference storage unit 85E stores the calculated phase difference average value for use as a future past utterance phase difference. The phase difference comparison unit 85F compares the phase difference average value with the previously stored past utterance phase difference.

When there is a difference between the phase difference average value and the past utterance phase difference exceeding a predetermined value which is an example of the third predetermined value, the threshold value adjusting unit 85G adjusts the threshold value for determining the sound source direction. The difference is the absolute value of the value obtained by subtracting the past utterance phase difference from the phase difference average value.

For example, the front surface of the housing 18 of the sound source direction determination device 10 is tilted at a plurality of different angles with respect to the vertical direction, and the phase difference average value of the user's voice and the phase difference average value of the voice of the dialogue partner are set. Get the difference at each angle. Of the plurality of acquired absolute values of the difference, the minimum value can be used as the third predetermined value. The third predetermined value may be, for example, 4.1 [rad]. If there is no past utterance phase difference exceeding the third predetermined value, the threshold value is not adjusted.

When there are a plurality of past utterance phase differences having a difference exceeding a predetermined value, the threshold value adjusting unit 85G adjusts the threshold value for determining the sound source direction by using the latest past utterance phase difference. Specifically, for example, the average value (that is, an intermediate value) of the phase difference average value of the current utterance section and the past utterance phase difference is set as the threshold value for determining the sound source direction. The sound source direction determination unit 85H determines the sound source direction using the adjusted threshold value, and outputs the determination result.

FIG. 25 will be used to describe the adjustment of the threshold value for determining the sound source direction. The vertical axis of FIG. 25 represents the phase difference [rad], and the horizontal axis represents time, that is, the frame number. The broken line 86P represents the phase difference between the sound signal corresponding to the sound acquired by the first microphone 11 and the sound signal corresponding to the sound acquired by the second microphone 12 for each frame.

As described above, the phase difference average value of the utterance section 86H1 which is the previous utterance section is stored as the past utterance phase difference in the data storage area 53B of the secondary storage unit 53, for example. There is a difference 86D exceeding a predetermined value between the phase difference average value of the utterance section 86H2, which is the current utterance section, and the past utterance phase difference corresponding to the utterance section 86H1.

The threshold value adjusting unit 85F sets, for example, an average value of the past utterance phase difference corresponding to the utterance section 86H1 and the phase difference average value of the utterance section 86H2 as the threshold value 86T. The set threshold value is used to determine the sound source direction of the sound signal in the utterance section 86H2.

As illustrated in FIG. 26A, the sound source direction determination device 10 is assumed to be mounted on the user so that the front surface of the housing 18 is substantially parallel to the vertical direction. In FIG. 24A, with the predetermined phase difference threshold value 81T as a boundary, the sound source direction of the voice in the region 81U is determined to be upward, that is, the utterance of the user, and the sound source direction of the voice in the region 81F is forward, that is, the dialogue partner. It is determined to be an utterance.

However, the sound source direction determination device 10 may be tilted as illustrated in FIG. 26B depending on the body shape of the user who is the wearer of the sound source direction determination device 10, the wearing method, or the like. For example, when the user is a woman, the front surface of the housing 18 of the sound source direction determination device 10 is tilted diagonally upward as illustrated in FIG. 26B due to the influence of the tilt of the chest. In this case, as illustrated by the phase difference threshold value 82T, the boundary of the determination is also inclined, that is, rotated.

In FIG. 26B, with the phase difference threshold value 82T as a boundary, the sound source direction of the voice in the region 82U is determined to be upward, that is, the utterance of the user who is the wearer, and the sound source direction of the voice in the region 82F is forward, that is, dialogue. It is determined that the speech is from the other party. Therefore, there is a possibility that the utterance of the user exemplified by the arrow 82V is determined to be the utterance of the dialogue partner.

FIG. 27A illustrates a case where the front surface of the housing 18 of the sound source direction determination device 10 is substantially parallel in the vertical direction, and FIG. 27B shows a case where the front surface of the housing 18 is inclined so as to face diagonally upward. Illustrate. The phase difference 83D exemplified in FIG. 27A and the phase difference 84D exemplified in FIG. 27B are substantially equal to each other. The phase difference 83D represents the phase difference between the arrow 83F1 indicating the arrival at the upper surface of the voice of the dialogue partner and the arrow 83F2 indicating the arrival at the front surface. The phase difference 84D represents the phase difference between the arrow 84U1 indicating the arrival at the upper surface of the user's voice and the arrow 84U2 indicating the arrival at the front surface.

FIG. 28 illustrates the phase difference 91A corresponding to the phase difference 83D of FIG. 27A and the phase difference 91B corresponding to the phase difference 84D of FIG. 27B. With the phase difference threshold value 91T, it is difficult to distinguish between the phase difference 91A and the phase difference 91B, and even if the threshold value is adjusted, it is difficult to distinguish between the phase difference 91A and the phase difference 91B.

On the other hand, there is a difference in the phase difference between the voice of the user who is the wearer and the voice of the dialogue partner, even if the sound source direction determination device 10 is tilted, if the tilt is the same, the difference exceeds a predetermined value. Therefore, by adjusting the phase difference threshold value based on the utterance of the user and the utterance of the dialogue partner, the sound source direction can be appropriately determined even if the sound source direction determination device 10 is tilted.

FIG. 29A illustrates a phase difference 92A of the user's voice and a phase difference 92B of the voice of the dialogue partner when the front surface of the housing 18 of the sound source direction determination device 10 is substantially parallel in the vertical direction. By adjusting the phase difference threshold value 92T to the average value of the phase difference 92A and the phase difference 92B, the phase difference 92A and the phase difference 92B can be distinguished from each other. That is, the sound source direction can be appropriately determined.

FIG. 29B illustrates a phase difference 93A of the voice of the user and a phase difference 93B of the voice of the dialogue partner when the front surface of the housing 18 of the sound source direction determination device 10 is tilted so as to face diagonally upward. By adjusting the phase difference threshold value 93T to the average value of the phase difference 93A and the phase difference 93B, the phase difference 93A and the phase difference 93B can be distinguished. That is, the sound source direction can be appropriately determined.

FIG. 30A shows an example of the flow of the sound source determination process. The CPU 51 sets the variable NP to 0 in step 201. The variable NP is a variable for summing the normalized phase differences of the utterance sections.

In step 202, the CPU 51 reads the sound signal corresponding to the sound acquired by the first microphone 11 and the second microphone 12 for one frame, and in step 203, the time frequency is converted. The CPU 51 determines in step 204 whether or not the utterance section has started.

If the determination in step 204 is denied, the CPU 51 returns to step 202. If the determination in step 204 is affirmed, the CPU 51 calculates the normalized phase difference in step 205, and adds the normalized phase difference to the variable NP in step 206.

In step 207, the CPU 51 reads the sound signal corresponding to the sound acquired by the first microphone 11 and the second microphone 12 for one frame, and in step 208, the time frequency is converted. The CPU 51 determines in step 209 whether or not the utterance section has ended.

If the determination in step 209 is denied, the CPU 51 returns to step 205. If the determination in step 209 is affirmed, the CPU 51 divides the value of the variable NP by the number of frames of the sound signal read in step 207 in step 210, so that the average normalization position is an example of the phase difference average value. Calculate the phase difference. The CPU 51 stores the calculated average normalized phase difference as the past utterance phase difference in the data storage area 53B of the secondary storage unit 53 for future use in step 211.

In step 212, the CPU 51 compares the past utterance phase difference stored in the previous process with the average normalized phase difference. If the determination in step 212 is affirmed, and if there is a difference exceeding a predetermined value between the past utterance phase difference and the average normalized phase difference, the CPU 51 adjusts the threshold value in step 213 and proceeds to step 214. Specifically, in step 213, the CPU 51 adjusts the threshold value by setting the average value of the past utterance phase difference and the average normalized phase difference as a threshold value which is an example of the sixth threshold value.

If the determination in step 212 is denied, the CPU 51 does not adjust the threshold value and proceeds to step 214. In step 214, the CPU 51 determines whether or not the sound source direction of the sound signal read in step 207 is upward. Specifically, it is determined whether or not the average normalized phase difference exceeds the threshold value.

If the determination in step 214 is affirmed, the CPU 51 is set to translate the sound signal read in step 207 into the first language in step 215. If the determination in step 214 is denied, the CPU 51 determines in step 216 whether or not the sound source direction of the sound signal read in step 207 is forward. Specifically, it is determined whether or not the average normalized phase difference is equal to or less than the threshold value.

If the determination in step 216 is affirmed, the CPU 51 is set to translate the sound signal read in step 207 into a second language in step 217. In step 218, the CPU 51 determines whether or not an operation instructing the user to end the sound source direction determination process, such as pressing a predetermined button, has been performed.

If the determination in step 218 is denied, the CPU 51 returns to step 201, and if the determination in step 218 is affirmed, the CPU 51 ends the sound source direction determination process.

FIG. 30B shows an example of the flow of the sound source direction determination process. The sound source direction determination process of FIG. 30B adjusts the threshold value based on the sound pressure difference between the voice of the user and the voice of the dialogue partner.

In step 231 the CPU 51 sets 0 to the variable HV for calculating the total of the high frequency sound pressure differences. Steps 232 to 234 are the same as steps 202 to 204 in FIG. 30A.

The CPU 51 calculates the high-frequency sound pressure difference in step 235, and adds the high-frequency sound pressure difference calculated in step 236 to the value of the variable HV. Steps 237 to 239 are similar to steps 207 to 209 of FIG. 30A.

In step 240, the CPU 51 calculates the average high-frequency sound pressure difference, which is an example of the sound pressure difference average value, by dividing the value of the variable HV by the number of frames of the sound signal read in step 237. The CPU 51 stores the calculated average high-frequency sound pressure difference as the past utterance sound pressure difference in the data storage area 53B of the secondary storage unit 53 for future use in step 241.

The CPU 51 compares the past utterance sound pressure difference stored in the previous process with the average high frequency sound pressure difference. If the determination in step 242 is affirmed, the CPU 51 adjusts the threshold value by setting the average value of the past utterance sound pressure difference and the average high frequency sound pressure difference as a threshold value as an example of the fifth threshold value in step 243. Proceed to step 244. The determination in step 242 is affirmed when there is a difference between the past utterance sound pressure difference and the average high frequency sound pressure difference exceeding a predetermined value which is an example of the second predetermined value.

For example, the front surface of the housing 18 of the sound source direction determination device 10 is tilted at a plurality of different angles with respect to the vertical direction, and the average value of the sound pressure difference of the user's voice and the average value of the sound pressure difference of the voice of the dialogue partner are set. Get the difference at each angle. Of the plurality of acquired absolute values of the difference, the minimum value can be used as the second predetermined value. The second predetermined value may be, for example, 3.0 [dB]. If there is no past speech pressure difference exceeding the second predetermined value, the threshold value is not adjusted.

If the determination in step 242 is denied, the CPU 51 does not adjust the threshold value and proceeds to step 244. In step 244, the CPU 51 determines whether or not the sound source direction of the sound signal read in step 237 is upward. Specifically, it is determined whether or not the average high-frequency sound pressure difference exceeds the threshold value.

If the determination in step 244 is affirmed, the CPU 51 is set to translate the sound signal read in step 237 into the first language in step 245. If the determination in step 244 is denied, the CPU 51 determines in step 246 whether or not the sound source direction of the sound signal read in step 207 is forward. Specifically, it is determined whether or not the average high-frequency sound pressure difference is equal to or less than the threshold value. Step 248 is similar to step 218 of FIG. 30A.

FIG. 30A is an example in which the fifth embodiment is applied to the sound source direction determination process of FIG. 17G of the fourth embodiment, and FIG. 30B shows the fifth embodiment to the sound source direction determination process of FIG. 11 of the third embodiment. This is an applied example. However, the fifth embodiment may be applied to the sound source direction determination process of FIGS. 17A to 17F of the fourth embodiment. That is, both the threshold value for determining the sound pressure difference and the threshold value for determining the phase difference may be adjusted.

If there are a plurality of past utterance phase differences whose difference from the phase difference average value exceeds a predetermined value, the latest past utterance phase difference may be used, or the difference among the past utterance phase differences within the predetermined time may be used. The maximum past utterance phase difference may be used. Further, the average value of the past utterance phase differences within a predetermined time may be used.

When there are multiple past utterance sound pressure differences whose difference from the average sound pressure difference exceeds a predetermined value, the latest past utterance sound pressure difference may be used, or the difference among the past utterance sound pressure differences within the predetermined time is the largest. The past utterance sound pressure difference may be used. Further, the average value of the past utterance sound pressure difference within a predetermined time may be used.

Although an example of calculating the phase difference average value or the sound pressure difference average value of a plurality of frames in the utterance section has been described, the phase difference average value and the sound pressure difference average value of a part of the utterance section may be calculated. good. Further, when the utterance section extends for a long time, the utterance section may be divided into a plurality of parts, and the phase difference average value or the sound pressure difference average value may be calculated for each of the plurality of divided subsections.

An example of adjusting the threshold value for determining the sound source direction naturally during a dialogue between the user and the dialogue partner has been described, but at the beginning of the dialogue, the user and the dialogue partner alternately utter a phrase exceeding a predetermined time length. The voice of the utterance may be used to adjust the threshold. The phrase may be, for example, a default greeting (eg, "hello").

In the above example, step 216 of FIG. 30A can be omitted, but for example, the threshold value for determining the sound source direction in step 214 and the threshold value for determining the sound source direction in step 216 are set to be different values. May be good. Specifically, for example, the threshold value used in step 216 may be reduced by a predetermined amount from the threshold value used in step 214.

This makes it possible to reduce the risk of erroneous determination of sound that is difficult to determine in the direction of the sound source, that is, sound that can be determined to be sound from any sound source direction. The same applies to step 246 of FIG. 30B. Further, the threshold value used in step 214 or step 244 may be increased by a predetermined amount.

When the signal-to-noise ratio of the sound signal is calculated and the signal-to-noise ratio is smaller than the predetermined value which is an example of the fourth predetermined value, the threshold value for determining the sound source direction is set to the predetermined value which is an example of the fifth predetermined value. You may try to lower it by a minute. This is because the smaller the signal-to-noise ratio, the smaller the difference in phase difference and sound pressure difference depending on the sound source direction.

The fourth predetermined value may be, for example, a stationary noise ratio, and the fifth predetermined value may be, for example, 0.5 [dB] in the case of a threshold value for distinguishing the average sound pressure difference value, and the phase difference. In the case of the threshold value for distinguishing the average value, it may be, for example, 0.5 [rad]. The steady-state noise ratio can be calculated by existing methods.

Although the example applied to the sound source direction determination device 10 exemplified in FIGS. 2A and 2B has been described, the present embodiment may be applied to the sound source direction determination device 10C exemplified in FIGS. 13A to 13C. According to the present embodiment, even when the user exists at a position deviated from the position facing the right side surface and the front surface of the housing 18C and speaks, the sound source direction, that is, the speaker is appropriately determined. be able to.

The order of processing of the flowcharts in FIGS. 30A and 30B is an example, and the present embodiment is not limited to the order of the processing.

In the present embodiment, by adjusting the threshold value for determining the sound source direction based on the voice of the user and the voice of the dialogue partner, the sound source direction can be appropriately determined even when the sound source determination device is tilted. Can be done.

(Related technology)
Next, the related technology will be described. In the related art, as illustrated in FIG. 18, two directional microphones are arranged so as to intersect the directional 11XOR of the directional microphone 11X and the directional 12XOR of the directional microphone 12X. For example, the oriented 11XOR is directed upwards and the directed 12XOR is directed forward.

With this configuration, it is possible to determine the direction of the sound source by using the sound pressure difference of the sound acquired by the directional microphone 11X and the directional microphone 12X. That is, when the sound pressure of the sound acquired by the directional microphone 11X is larger than the sound pressure of the sound acquired by the directional microphone 12X, the sound source is located above and the sound pressure of the sound acquired by the directional microphone 12X is directional. If it is higher than the sound pressure of the sound acquired by the microphone 11X, the sound source is in front.

However, since the directional microphone is larger than the omnidirectional microphone as illustrated in FIG. 19, it is difficult to miniaturize the sound source direction determination device when the directional microphone is used. In the example of FIG. 19, the volume of the directional microphone is 226 [cubi mm], and the volume of the omnidirectional microphone is 11 [cubi mm]. That is, the volume of the directional microphone is about 20 times the volume of the omnidirectional microphone. Further, since the directional microphone is more expensive than the omnidirectional microphone, it is difficult to reduce the price of the sound source direction determination device when the directional microphone is used.

However, it is difficult to realize a sound source direction determination device capable of accurately determining the sound source direction by simply replacing the directional microphone of the sound source direction determination device illustrated in FIG. 18 with an omnidirectional microphone. .. As illustrated in FIG. 20A, the range 11YOR in which the omnidirectional microphone 11Y can acquire sound and the range 12YOR in which the omnidirectional microphone 12Y can acquire sound substantially overlap. Therefore, there is no significant difference in the sound pressure difference of the sounds acquired by the omnidirectional microphones 11Y and 12Y to the extent that the sound source direction can be accurately determined.

In FIG. 20B, the width in the front-rear direction is about 1 [cm] and the front surface is similar to the first embodiment in which the first microphone 11Y is installed on the upper surface of the housing 18Y and the second microphone 12Y is installed on the front surface. Illustrates the sound source direction determination device 10Y of the related technology, which is about the size of a business card. The first microphone 11Y and the second microphone 12Y are omnidirectional microphones. FIG. 21 illustrates the sound pressure difference of the sound source direction determination device 10Y of the related technology and the sound pressure difference of the sound source direction determination device 10 of the first embodiment. When the sound source is above the sound source direction determination device, the sound pressure difference between the sound pressure of the sound acquired by the first microphone and the sound pressure of the sound acquired by the second microphone is 2.9 [dB] in the related technology. Yes, in the first embodiment, it is 7.2 [dB].

When the sound source is in front of the sound source direction determination device, the sound pressure difference between the sound pressure of the sound acquired by the first microphone and the sound pressure of the sound acquired by the second microphone is -2.9 [dB] in the related technology. In the first embodiment, it is -4.2 [dB]. That is, when the sound source is above the sound source direction determination device, the sound pressure difference calculated in the first embodiment is 4.3 [dB] larger than that of the related technology, and when the sound source is in front of the sound source direction determination device, the first The sound pressure difference calculated in one embodiment is 1.3 [dB] smaller than that of the related technology.

Therefore, in the present embodiment, the possibility of obtaining an erroneous determination result in the determination in step 104 and step 106 in FIG. 11 is reduced. Therefore, according to the present embodiment, the accuracy of the sound source direction determination using the omnidirectional microphone is reduced. It is possible to improve.

The following additional notes will be further disclosed with respect to each of the above embodiments.

但し、
phase[j]=atan(phase_im[j]/phase_re[j])、
phase_re[j]=re1[j]*re2[j]+im1[j]*im2[j]、
phase_im[j]=im1[j]*re2[j]-re1[j]*im2[j]、
C_n[j]=λ[j]/λ_cであり、
ｊは周波数帯域数であり、
re1[j]は、j番目の周波数帯域の前記第１音圧のスペクトルの実部であり、
re2[j]は、j番目の周波数帯域の前記第２音圧のスペクトルの実部であり、
im1[j]は、j番目の周波数帯域の前記第１音圧のスペクトルの虚部であり、
im2[j]は、j番目の周波数帯域の前記第２音圧のスペクトルの虚部であり、
λ[j]は、j番目の周波数帯域の音の波長であり、
λ_cは、基準周波数の音の波長であり、
ｅｅは、前記対象周波数帯域の上限であり、
ｓｓは、前記対象周波数帯域の下限である。
（付記９）
前記音圧の相違は、前記第１音圧のパワーの対数から前記第２音圧のパワーの対数を減算したフレーム毎の音圧差の複数フレームの平均値である音圧差平均値であり、
前記位相の相違は、フレーム毎の対象周波数帯域の位相差の複数フレームの平均値である位相差平均値であり、
前記音圧差平均値が第５閾値よりも大きい場合、及び、前記位相差平均値が第６閾値よりも大きい場合の内少なくとも一方の場合、前記音源が前記第１平坦面に対向する位置に存在すると判定し、
前記第５閾値は、前記音源が前記第１平坦面に対向する位置に存在する場合の前記音圧差平均値と、前記音源が前記第２平坦面に対向する位置に存在する場合の前記音圧差平均値と、の平均値であり、
前記第６閾値は、前記音源が前記第１平坦面に対向する位置に存在する場合の前記位相差平均値と、前記音源が前記第２平坦面に対向する位置に存在する場合の前記位相差平均値と、の平均値である、
付記１～付記５の何れかの音源方向判定装置。
（付記１０）
前記音圧差平均値が前記第５閾値以下の場合、及び、前記位相差平均値が前記第６閾値以下の場合の内少なくとも一方の場合、前記音源が前記第２平坦面に対向する位置に存在すると判定する、
付記９の音源方向判定装置。
（付記１１）
前記音源が前記第１平坦面に対向する位置に存在する場合の前記音圧差平均値と、前記音源が前記第２平坦面に対向する位置に存在する場合の前記音圧差平均値と、の平均値は、前記音圧差の第１発話区間の平均値である第１平均値と、前記音圧差の第２発話区間の平均値である第２平均値と、の平均値であり、前記第１平均値と前記第２平均値との相違は、第２所定値を超え、
前記音源が前記第１平坦面に対向する位置に存在する場合の前記位相差平均値と、前記音源が前記第２平坦面に対向する位置に存在する場合の前記位相差平均値と、の平均値は、前記位相差の第３発話区間の平均値である第３平均値と、前記位相差の第４発話区間の平均値である第４平均値と、の平均値であり、前記第３平均値と前記第４平均値との相違は、第３所定値を超える、
付記９または付記１０の音源方向判定装置。
（付記１２）
前記音に対応する信号の信号対雑音比が第４所定値より小さい場合、前記第５閾値及び前記第６閾値を第５所定値分低減する、
付記９～付記１１の何れかの音源方向判定装置。
（付記１３）
前記音源が前記第１平坦面と対向する位置に存在すると判定された場合、前記音に対応する信号を第１言語に翻訳し、前記音源が前記第２平坦面に対向する位置に存在すると判定された場合、前記音に対応する信号を第２言語に翻訳する、
付記１～付記１２の何れかの音源方向判定装置。
（付記１４）
第１平坦面に開口した第１開口部を一端部に備え、前記第１開口部から音が伝搬する第１音道、及び、前記第１平坦面と交差する第２平坦面に開口した第２開口部を一端部に備え、前記第２開口部から音が伝搬する第２音道が内部に設けられたマイク設置部と、
前記第１音道の他端部に設置された無指向性の第１マイクロフォンと、
前記第２音道の他端部に設置された無指向性の第２マイクロフォンと、
コンピュータと、
を含む音源方向判定装置の前記コンピュータが、
前記第１マイクロフォンで取得された音の第１周波数成分の音圧である第１音圧と、前記第２マイクロフォンで取得された音の前記第１周波数成分の音圧である第２音圧との音圧の相違、及び、前記第１マイクロフォンで取得された音の第２周波数成分の位相である第１位相と、前記第２マイクロフォンで取得された音の前記第２周波数成分の位相である第２位相との位相の相違の少なくとも一方に基づいて、音源が存在する方向を判定する、
音源方向判定方法。
（付記１５）
前記音圧の相違は、前記第１音圧のパワーの対数から前記第２音圧のパワーの対数を減算した音圧差の平均値であり、
前記位相の相違は、対象周波数帯域の位相差の平均値であり、
前記音圧差の平均値が正の第１閾値よりも大きい場合、及び、前記位相差の平均値が正の第３閾値よりも大きい場合の内少なくとも一方の場合、前記音源が前記第１平坦面に対向する位置に存在すると判定する、
付記１４の音源方向判定方法。
（付記１６）
前記音圧の相違は、前記第１音圧のパワーの対数から前記第２音圧のパワーの対数を減算したフレーム毎の音圧差の複数フレームの平均値である音圧差平均値であり、
前記位相の相違は、フレーム毎の対象周波数帯域の位相差の複数フレームの平均値である位相差平均値であり、
前記音圧差平均値が第５閾値よりも大きい場合、及び、前記位相差平均値が第６閾値よりも大きい場合の内少なくとも一方の場合、前記音源が前記第１平坦面に対向する位置に存在すると判定し、
前記第５閾値は、前記音源が前記第１平坦面に対向する位置に存在する場合の前記音圧差平均値と、前記音源が前記第２平坦面に対向する位置に存在する場合の前記音圧差平均値と、の平均値であり、
前記第６閾値は、前記音源が前記第１平坦面に対向する位置に存在する場合の前記位相差平均値と、前記音源が前記第２平坦面に対向する位置に存在する場合の前記位相差平均値と、の平均値である、
付記１４の音源方向判定方法。
（付記１７）
第１平坦面に開口した第１開口部を一端部に備え、前記第１開口部から音が伝搬する第１音道、及び、前記第１平坦面と交差する第２平坦面に開口した第２開口部を一端部に備え、前記第２開口部から音が伝搬する第２音道が内部に設けられたマイク設置部と、
前記第１音道の他端部に設置された無指向性の第１マイクロフォンと、
前記第２音道の他端部に設置された無指向性の第２マイクロフォンと、
コンピュータと、
を含む音源方向判定装置のコンピュータに、
前記第１マイクロフォンで取得された音の第１周波数成分の音圧である第１音圧と、前記第２マイクロフォンで取得された音の前記第１周波数成分の音圧である第２音圧との音圧の相違、及び、前記第１マイクロフォンで取得された音の第２周波数成分の位相である第１位相と、前記第２マイクロフォンで取得された音の前記第２周波数成分の位相である第２位相との位相の相違の少なくとも一方に基づいて、音源が存在する方向を判定する、
音源方向判定処理を実行させるためのプログラム。
（付記１８）
前記音圧の相違は、前記第１音圧のパワーの対数から前記第２音圧のパワーの対数を減算した音圧差の平均値であり、
前記位相の相違は、対象周波数帯域の位相差の平均値であり、
前記音圧差の平均値が正の第１閾値よりも大きい場合、及び、前記位相差の平均値が正の第３閾値よりも大きい場合の内少なくとも一方の場合、前記音源が前記第１平坦面に対向する位置に存在すると判定する、
付記１７のプログラム。
（付記１９）
前記音圧の相違は、前記第１音圧のパワーの対数から前記第２音圧のパワーの対数を減算したフレーム毎の音圧差の複数フレームの平均値である音圧差平均値であり、
前記位相の相違は、フレーム毎の対象周波数帯域の位相差の複数フレームの平均値である位相差平均値であり、
前記音圧差平均値が第５閾値よりも大きい場合、及び、前記位相差平均値が第６閾値よりも大きい場合の内少なくとも一方の場合、前記音源が前記第１平坦面に対向する位置に存在すると判定し、
前記第５閾値は、前記音源が前記第１平坦面に対向する位置に存在する場合の前記音圧差平均値と、前記音源が前記第２平坦面に対向する位置に存在する場合の前記音圧差平均値と、の平均値であり、
前記第６閾値は、前記音源が前記第１平坦面に対向する位置に存在する場合の前記位相差平均値と、前記音源が前記第２平坦面に対向する位置に存在する場合の前記位相差平均値と、の平均値である、
付記１７のプログラム。 (Appendix 1)
A first opening opened in the first flat surface is provided at one end, and a first sound path through which sound propagates from the first opening and a second flat surface intersecting with the first flat surface are opened. A microphone installation unit having two openings at one end and an internal second sound path through which sound propagates from the second opening.
An omnidirectional first microphone installed at the other end of the first sound path, and
An omnidirectional second microphone installed at the other end of the second sound path, and
The first sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the second microphone. The difference in sound pressure between the two, and the first phase, which is the phase of the second frequency component of the sound acquired by the first microphone, and the phase of the second frequency component of the sound acquired by the second microphone. A determination unit that determines the direction in which the sound source exists based on at least one of the phase differences from the second phase.
including,
Sound source direction determination device.
(Appendix 2)
The first frequency component is a high frequency component.
The sound source direction determination device of Appendix 1.
(Appendix 3)
The first flat surface and the second flat surface are orthogonal to each other.
The area of the first flat surface is equal to or less than the first predetermined value, and the area of the second flat surface is larger than the first predetermined value.
The first sound path has a first diffracting portion that diffracts sound in the first opening, and has a second diffracting portion that is a bending portion that diffracts sound in the middle.
The second sound path has a third diffracting portion that diffracts sound in the second opening.
The sound source direction determination device of Appendix 1 or Appendix 2.
(Appendix 4)
The first flat surface and the second flat surface are orthogonal to each other.
The area of the first flat surface is equal to or less than the first predetermined value, and the area of the second flat surface is larger than the first predetermined value.
The first sound path has a first diffracting portion that diffracts sound in the first opening, and has a second diffracting portion that is a bending portion that diffracts sound in the middle.
The second sound path has a third diffracting portion that diffracts sound in the second opening, and has a fourth diffracting portion that is a bending portion that diffracts sound in the middle.
The sound source direction determination device of Appendix 1 or Appendix 2.
(Appendix 5)
The first flat surface and the second flat surface are orthogonal to each other.
The areas of the first flat surface and the second flat surface are larger than the first predetermined value.
The first sound path has a first diffracting portion that diffracts sound in the first opening.
The second sound path has a second diffracting portion that diffracts sound in the second opening.
The sound source direction determination device of Appendix 1 or Appendix 2.
(Appendix 6)
The difference in sound pressure is the average value of the sound pressure difference obtained by subtracting the logarithm of the power of the second sound pressure from the logarithm of the power of the first sound pressure.
The phase difference is the average value of the phase difference in the target frequency band.
When the average value of the sound pressure difference is larger than the positive first threshold value and at least one of the cases where the average value of the phase difference is larger than the positive third threshold value, the sound source is the first flat surface. Judged to exist at a position facing
The sound source direction determination device according to any one of Supplementary note 1 to Supplementary note 5.
(Appendix 7)
When the average value of the sound pressure difference is smaller than the negative second threshold value and at least one of the cases where the average value of the phase difference is smaller than the negative fourth threshold value, the sound source is the second flat surface. Judged to exist at a position facing
The sound source direction determination device of Appendix 6.
(Appendix 8)
The average value a_phase of the phase difference of the target frequency band is represented by the following equation (10), and is the sound source direction determination device of Appendix 6 or Appendix 7.

However,
phase [j] = atan (phase_ im [j] / phase_ re [j]),
phase_re [j] = re1 [j] * re2 [j] + im1 [j] * im 2 [j],
phase_im [j] = im1 [j] * re2 [j]-re1 [j] * im2 [j],
C_n [j] = λ [j] / λ_c,
j is the number of frequency bands
re1 [j] is the real part of the spectrum of the first sound pressure in the jth frequency band.
re2 [j] is the real part of the spectrum of the second sound pressure in the jth frequency band.
im1 [j] is an imaginary part of the spectrum of the first sound pressure in the jth frequency band.
im2 [j] is an imaginary part of the spectrum of the second sound pressure in the jth frequency band.
λ [j] is the wavelength of the sound in the jth frequency band.
λ_c is the wavelength of the sound of the reference frequency,
ee is the upper limit of the target frequency band, and is
ss is the lower limit of the target frequency band.
(Appendix 9)
The difference in sound pressure is a sound pressure difference average value which is an average value of a plurality of frames of the sound pressure difference for each frame obtained by subtracting the logarithm of the power of the second sound pressure from the logarithm of the power of the first sound pressure.
The phase difference is a phase difference average value which is an average value of a plurality of frames of the phase difference of the target frequency band for each frame.
When the sound pressure difference average value is larger than the fifth threshold value and at least one of the cases where the phase difference average value is larger than the sixth threshold value, the sound source exists at a position facing the first flat surface. Then,
The fifth threshold is the average value of the sound pressure difference when the sound source is present at a position facing the first flat surface, and the sound pressure difference when the sound source is present at a position facing the second flat surface. The average value and the average value of
The sixth threshold is the phase difference average value when the sound source is present at a position facing the first flat surface, and the phase difference when the sound source is present at a position facing the second flat surface. The average value and the average value of,
The sound source direction determination device according to any one of Supplementary note 1 to Supplementary note 5.
(Appendix 10)
When the sound pressure difference average value is equal to or less than the fifth threshold value and at least one of the cases where the phase difference average value is equal to or less than the sixth threshold value, the sound source exists at a position facing the second flat surface. Judge,
The sound source direction determination device of Appendix 9.
(Appendix 11)
The average of the sound pressure difference average value when the sound source is present at a position facing the first flat surface and the sound pressure difference average value when the sound source is present at a position facing the second flat surface. The value is an average value of a first mean value which is an average value of the first speech section of the sound pressure difference and a second mean value which is an average value of the second speech section of the sound pressure difference, and is the first mean value. The difference between the mean value and the second mean value exceeds the second predetermined value,
The average of the phase difference average value when the sound source is present at a position facing the first flat surface and the phase difference average value when the sound source is present at a position facing the second flat surface. The value is an average value of a third mean value which is an average value of the third speech section of the phase difference and a fourth mean value which is an average value of the fourth speech section of the phase difference, and is the third mean value. The difference between the mean value and the fourth mean value exceeds the third predetermined value.
The sound source direction determination device of Appendix 9 or Appendix 10.
(Appendix 12)
When the signal-to-noise ratio of the signal corresponding to the sound is smaller than the fourth predetermined value, the fifth threshold value and the sixth threshold value are reduced by the fifth predetermined value.
A sound source direction determination device according to any one of Supplementary note 9 to Supplementary note 11.
(Appendix 13)
When it is determined that the sound source exists at a position facing the first flat surface, the signal corresponding to the sound is translated into the first language, and it is determined that the sound source exists at a position facing the second flat surface. If so, the signal corresponding to the sound is translated into a second language.
The sound source direction determination device according to any one of Supplementary note 1 to Supplementary note 12.
(Appendix 14)
A first opening opened in the first flat surface is provided at one end, and a first sound path through which sound propagates from the first opening and a second flat surface intersecting with the first flat surface are opened. A microphone installation unit having two openings at one end and an internal second sound path through which sound propagates from the second opening.
An omnidirectional first microphone installed at the other end of the first sound path, and
An omnidirectional second microphone installed at the other end of the second sound path, and
With a computer
The computer of the sound source direction determination device including
The first sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the second microphone. The difference in sound pressure between the two, and the first phase, which is the phase of the second frequency component of the sound acquired by the first microphone, and the phase of the second frequency component of the sound acquired by the second microphone. Determining the direction in which the sound source is located, based on at least one of the phase differences from the second phase.
Sound source direction determination method.
(Appendix 15)
The difference in sound pressure is the average value of the sound pressure difference obtained by subtracting the logarithm of the power of the second sound pressure from the logarithm of the power of the first sound pressure.
The phase difference is the average value of the phase difference in the target frequency band.
When the average value of the sound pressure difference is larger than the positive first threshold value and at least one of the cases where the average value of the phase difference is larger than the positive third threshold value, the sound source is the first flat surface. Judged to exist at a position facing
Appendix 14 Sound source direction determination method.
(Appendix 16)
The difference in sound pressure is a sound pressure difference average value which is an average value of a plurality of frames of the sound pressure difference for each frame obtained by subtracting the logarithm of the power of the second sound pressure from the logarithm of the power of the first sound pressure.
The phase difference is a phase difference average value which is an average value of a plurality of frames of the phase difference of the target frequency band for each frame.
When the sound pressure difference average value is larger than the fifth threshold value and at least one of the cases where the phase difference average value is larger than the sixth threshold value, the sound source exists at a position facing the first flat surface. Then,
The fifth threshold is the average value of the sound pressure difference when the sound source is present at a position facing the first flat surface, and the sound pressure difference when the sound source is present at a position facing the second flat surface. The average value and the average value of
The sixth threshold is the phase difference average value when the sound source is present at a position facing the first flat surface, and the phase difference when the sound source is present at a position facing the second flat surface. The average value and the average value of,
Appendix 14 Sound source direction determination method.
(Appendix 17)
A first opening opened in the first flat surface is provided at one end, and a first sound path through which sound propagates from the first opening and a second flat surface intersecting with the first flat surface are opened. A microphone installation unit having two openings at one end and an internal second sound path through which sound propagates from the second opening.
An omnidirectional first microphone installed at the other end of the first sound path, and
An omnidirectional second microphone installed at the other end of the second sound path, and
With a computer
To the computer of the sound source direction determination device including
The first sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the second microphone. The difference in sound pressure between the two, and the first phase, which is the phase of the second frequency component of the sound acquired by the first microphone, and the phase of the second frequency component of the sound acquired by the second microphone. Determining the direction in which the sound source is located, based on at least one of the phase differences from the second phase.
A program for executing sound source direction determination processing.
(Appendix 18)
The difference in sound pressure is the average value of the sound pressure difference obtained by subtracting the logarithm of the power of the second sound pressure from the logarithm of the power of the first sound pressure.
The phase difference is the average value of the phase difference in the target frequency band.
When the average value of the sound pressure difference is larger than the positive first threshold value and at least one of the cases where the average value of the phase difference is larger than the positive third threshold value, the sound source is the first flat surface. Judged to exist at a position facing
Appendix 17 program.
(Appendix 19)
The difference in sound pressure is a sound pressure difference average value which is an average value of a plurality of frames of the sound pressure difference for each frame obtained by subtracting the logarithm of the power of the second sound pressure from the logarithm of the power of the first sound pressure.
The phase difference is a phase difference average value which is an average value of a plurality of frames of the phase difference of the target frequency band for each frame.
When the sound pressure difference average value is larger than the fifth threshold value and at least one of the cases where the phase difference average value is larger than the sixth threshold value, the sound source exists at a position facing the first flat surface. Then,
The fifth threshold is the average value of the sound pressure difference when the sound source is present at a position facing the first flat surface, and the sound pressure difference when the sound source is present at a position facing the second flat surface. The average value and the average value of
The sixth threshold is the phase difference average value when the sound source is present at a position facing the first flat surface, and the phase difference when the sound source is present at a position facing the second flat surface. The average value and the average value of,
Appendix 17 program.

10 Sound source direction determination device 11 1st microphone 11R 1st sound path 11O 1st opening 11K Bending part 12 2nd microphone 12R 2nd sound path 12O 2nd opening 13 Judgment unit 14 Speech translation device 51 CPU
52 Primary storage 53 Secondary storage

Claims (14)

A first opening opened in the first flat surface is provided at one end, and a first sound path through which sound propagates from the first opening and a second flat surface intersecting with the first flat surface are opened. A microphone installation unit having two openings at one end and an internal second sound path through which sound propagates from the second opening.
An omnidirectional first microphone installed at the other end of the first sound path, and
An omnidirectional second microphone installed at the other end of the second sound path, and
The first sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the second microphone. The difference in sound pressure between the two, and the first phase, which is the phase of the second frequency component of the sound acquired by the first microphone, and the phase of the second frequency component of the sound acquired by the second microphone. A determination unit that determines the direction in which the sound source exists based on at least one of the phase differences from the second phase.
including,
Sound source direction determination device. The first frequency component is a high frequency component.
The sound source direction determination device according to claim 1. The first flat surface and the second flat surface are orthogonal to each other.
The area of the first flat surface is equal to or less than the first predetermined value, and the area of the second flat surface is larger than the first predetermined value.
The first sound path has a first diffracting portion that diffracts sound in the first opening, and has a second diffracting portion that is a bending portion that diffracts sound in the middle.
The second sound path has a third diffracting portion that diffracts sound in the second opening.
The sound source direction determination device according to claim 1 or 2. The first flat surface and the second flat surface are orthogonal to each other.
The area of the first flat surface is equal to or less than the first predetermined value, and the area of the second flat surface is larger than the first predetermined value.
The first sound path has a first diffracting portion that diffracts sound in the first opening, and has a second diffracting portion that is a bending portion that diffracts sound in the middle.
The second sound path has a third diffracting portion that diffracts sound in the second opening, and has a fourth diffracting portion that is a bending portion that diffracts sound in the middle.
The sound source direction determination device according to claim 1 or 2. The first flat surface and the second flat surface are orthogonal to each other.
The areas of the first flat surface and the second flat surface are larger than the first predetermined value.
The first sound path has a first diffracting portion that diffracts sound in the first opening.
The second sound path has a second diffracting portion that diffracts sound in the second opening.
The sound source direction determination device according to claim 1 or 2. The difference in sound pressure is the average value of the sound pressure difference obtained by subtracting the logarithm of the power of the second sound pressure from the logarithm of the power of the first sound pressure.
The phase difference is the average value of the phase difference in the target frequency band.
When the average value of the sound pressure difference is larger than the positive first threshold value and at least one of the cases where the average value of the phase difference is larger than the positive third threshold value, the sound source is the first flat surface. Judged to exist at a position facing
The sound source direction determination device according to any one of claims 1 to 5. When the average value of the sound pressure difference is smaller than the negative second threshold value and at least one of the cases where the average value of the phase difference is smaller than the negative fourth threshold value, the sound source is the second flat surface. Judged to exist at a position facing
The sound source direction determination device according to claim 6. The sound source direction determination device according to claim 6 or 7, wherein the average value a_phase of the phase difference in the target frequency band is represented by the following equation (1).

However,
phase [j] = atan (phase_ im [j] / phase_ re [j]),
phase_re [j] = re1 [j] * re2 [j] + im1 [j] * im 2 [j],
phase_im [j] = im1 [j] * re2 [j]-re1 [j] * im2 [j],
C_n [j] = λ [j] / λ_c,
j is the number of frequency bands
re1 [j] is the real part of the spectrum of the first sound pressure in the jth frequency band.
re2 [j] is the real part of the spectrum of the second sound pressure in the jth frequency band.
im1 [j] is an imaginary part of the spectrum of the first sound pressure in the jth frequency band.
im2 [j] is an imaginary part of the spectrum of the second sound pressure in the jth frequency band.
λ [j] is the wavelength of the sound in the jth frequency band.
λ_c is the wavelength of the sound of the reference frequency,
ee is the upper limit of the target frequency band, and is
ss is the lower limit of the target frequency band. The difference in sound pressure is a sound pressure difference average value which is an average value of a plurality of frames of the sound pressure difference for each frame obtained by subtracting the logarithm of the power of the second sound pressure from the logarithm of the power of the first sound pressure.
The phase difference is a phase difference average value which is an average value of a plurality of frames of the phase difference of the target frequency band for each frame.
When the sound pressure difference average value is larger than the fifth threshold value and at least one of the cases where the phase difference average value is larger than the sixth threshold value, the sound source exists at a position facing the first flat surface. Then,
The fifth threshold is the average value of the sound pressure difference when the sound source is present at a position facing the first flat surface, and the sound pressure difference when the sound source is present at a position facing the second flat surface. The average value and the average value of
The sixth threshold is the phase difference average value when the sound source is present at a position facing the first flat surface, and the phase difference when the sound source is present at a position facing the second flat surface. The average value and the average value of,
The sound source direction determination device according to any one of claims 1 to 5. When the sound pressure difference average value is equal to or less than the fifth threshold value and at least one of the cases where the phase difference average value is equal to or less than the sixth threshold value, the sound source exists at a position facing the second flat surface. Judge,
The sound source direction determination device according to claim 9. The average of the sound pressure difference average value when the sound source is present at a position facing the first flat surface and the sound pressure difference average value when the sound source is present at a position facing the second flat surface. The value is an average value of a first mean value which is an average value of the first speech section of the sound pressure difference and a second mean value which is an average value of the second speech section of the sound pressure difference, and is the first mean value. The difference between the mean value and the second mean value exceeds the second predetermined value,
The average of the phase difference average value when the sound source is present at a position facing the first flat surface and the phase difference average value when the sound source is present at a position facing the second flat surface. The value is an average value of a third mean value which is an average value of the third speech section of the phase difference and a fourth mean value which is an average value of the fourth speech section of the phase difference, and is the third mean value. The difference between the mean value and the fourth mean value exceeds the third predetermined value.
The sound source direction determination device according to claim 9 or 10. When it is determined that the sound source exists at a position facing the first flat surface, the signal corresponding to the sound is translated into the first language, and it is determined that the sound source exists at a position facing the second flat surface. If so, the signal corresponding to the sound is translated into a second language.
The sound source direction determination device according to any one of claims 1 to 11. A first opening opened in the first flat surface is provided at one end, and a first sound path through which sound propagates from the first opening and a second flat surface intersecting with the first flat surface are opened. A microphone installation unit having two openings at one end and an internal second sound path through which sound propagates from the second opening.
An omnidirectional first microphone installed at the other end of the first sound path, and
An omnidirectional second microphone installed at the other end of the second sound path, and
With a computer
The computer of the sound source direction determination device including
The first sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the second microphone. The difference in sound pressure between the two, and the first phase, which is the phase of the second frequency component of the sound acquired by the first microphone, and the phase of the second frequency component of the sound acquired by the second microphone. Determining the direction in which the sound source is located, based on at least one of the phase differences from the second phase.
Sound source direction determination method. A first opening opened in the first flat surface is provided at one end, and a first sound path through which sound propagates from the first opening and a second flat surface intersecting with the first flat surface are opened. A microphone installation unit having two openings at one end and an internal second sound path through which sound propagates from the second opening.
An omnidirectional first microphone installed at the other end of the first sound path, and
An omnidirectional second microphone installed at the other end of the second sound path, and
With a computer
To the computer of the sound source direction determination device including
The first sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the first microphone, and the second sound pressure, which is the sound pressure of the first frequency component of the sound acquired by the second microphone. The difference in sound pressure between the two, and the first phase, which is the phase of the second frequency component of the sound acquired by the first microphone, and the phase of the second frequency component of the sound acquired by the second microphone. Determining the direction in which the sound source is located, based on at least one of the phase differences from the second phase.
A program for executing sound source direction determination processing.

Images