KR100730297B1 - Sound source localization method using Head Related Transfer Function database - Google Patents

Abstract

The present invention relates to a sound source location tracking method using a head transfer function. More specifically, a first step of acquiring a sound wave signal through two microphones 54, 55 while varying the position of a sound source on a three-dimensional space (for example, an anechoic room) with respect to the platform 50; A second step of constructing a database by obtaining a head transfer function of signals acquired in the first step in a certain range of the audio frequency band by signal processing; A third step of acquiring a sound wave signal generated by the sound source to be tracked through two microphones; A fourth step of calculating a phase difference between signals acquired in the third step in a certain range of the audio frequency band by signal processing; In the second step, the value obtained by multiplying the square of the difference between the phase differences of the head transfer function and the phase difference obtained in the fourth step stored in the database by the square of the coherence function is integrated in a certain range of the audio frequency band A fifth step of obtaining a phase scale; And a sixth step of estimating a position of a sound source on a database having a phase scale relatively smaller than that of the other phase measures obtained in the fifth step to a position of the subject sound source to be traced. To a sound source location tracking method using the same.

Sound source, position tracking, head transfer function, phase scale, size scale, microphone

Description

[0001] The present invention relates to a sound source localization method using a head transfer function database,

FIG. 1A is a conceptual diagram of a conventional method for estimating a sound source position, and FIG. 1B is a conceptual diagram of a cone-shaped confusion plane.

2 is a conceptual diagram of a head transfer function.

FIG. 3 is experimental data showing the HRTF magnitude of the ears according to the horizontal angle change when the altitude angle is 0 ° for a sound source at a distance of 1m.

FIG. 4 is experimental data showing the HRTF size of the left ear according to the altitude change when the horizontal angle is 30 ° with respect to a sound source at a distance of 1m.

5A is a phase scale calculated by using HRTF data of a sound source according to a horizontal angle change at an altitude angle of 0 DEG, and FIG. 5B is a phase scale calculated by considering a free sound field.

FIG. 6A is a phase scale calculated using HRTF data according to elevation angle change when the horizontal angle is 30 DEG, and FIG. 6B is a scale scale. FIG. 6C is a phase scale calculated using HRTF data according to the elevation angle change when the horizontal angle is 60 DEG, and FIG. 6D is a size scale.

7 is a flowchart illustrating a sound source position estimation method using a head transfer function database according to an embodiment of the present invention.

DESCRIPTION OF THE EMBODIMENTS

10: sound waves,

20: microphone,

30: cone confusion plane,

40: Sound source,

50: Platform,

54: first microphone,

55: second microphone,

100: Flowchart.

A second step of acquiring a signal for database construction, a second step of building a database, a third step of acquiring a sound wave signal generated by the sound source to be tracked through a microphone, a phase difference and / A fifth step of calculating a phase scale and / or a scale of scale, and a sixth step of estimating a location of a sound source to be traced, to a method of estimating a sound source position using a head transfer function database .

사이의 관계는 다음과 같다.FIG. 1A is a conceptual diagram of a conventional sound source position estimation method. As shown in FIG. 1A, the prior art of this field is based on the assumption of a free sound field and a method for estimating the position of a sound source on a three-dimensional space from the delay time of a sound wave 10 signal input to each microphone 20 . At this time, the position angle of the sound source and the delay time

The relationship between the two is as follows.

Where c is the sound velocity, d is the delay distance, 2a is the distance between the two microphones, and Pi is the angle at which the sound source is present. Therefore, by estimating the delay time, the angle of presence of the sound source can be known. At this time, a generalized cross-correlation method (GCC) is widely used for estimating the delay time between output signals of the respective microphones 20. This is a method of estimating the delay time by obtaining a cross-correlation coefficient of the signals of the sound waves 10 entering each microphone 20 and obtaining a time delay having a maximum value.

This conventional technique does not take into consideration the shape of the platform on which the microphone 20 is mounted. However, a target platform such as an actual robot has a specific shape, and reflection, diffraction, scattering, etc. occur due to the shape of the platform until the sound waves 10 radiated from the sound source reach each microphone 20. Therefore, phase and size distortions occur due to these causes. In case of assuming only a free sound field without considering the platform shape as in the prior art, accurate sound source position estimation may not be performed.

 Also, since the positions of the sound sources having the same delay time are numerous in the three-dimensional space, a cone of confusion 30 as shown in FIG. 1B is generated. That is, if the location of the sound source is estimated based only on the delay time difference, there is a problem in that it can not be divided into the forward-backward division and the up-down division. That is, there is a problem in that it is impossible to simultaneously estimate the horizontal angle and the altitude angle in which the sound sources exist using only two microphones 20.

An object of the present invention is to provide an intelligent robot industry which can simultaneously estimate the horizontal angle and altitude angle of a sound source in a three-dimensional space by measuring a head transfer function in advance, And to provide a method for estimating an available sound source position.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings, so that those skilled in the art can easily carry out the present invention.

A sound source location estimation method using a head transfer function database according to the present invention includes a first step of acquiring a signal for building a database, a second step of processing a signal acquired in the first step to construct a database, A fifth step of calculating a phase scale and / or a scale scale by calculating a phase difference and / or a magnitude ratio by processing the signals acquired in the third stage, a fourth step of acquiring a sound wave signal through a microphone, And a sixth step of estimating a sound source location to be tracked.

2 is a conceptual diagram of a head transfer function. 2, the sound waves generated in the sound source 40 reach the center of the head of the platform 50, the first microphone microphone 54, and the second microphone 55, which are two measurement points. The definition of the head related transfer function is as follows.

,

,

Here, P _S is the pressure in the sound source, sound pressure P _ff is the center point of the head, P _S is the pressure at the measuring point. That is, when the platform 50 is present, the sound waves radiated from the sound source 40 reach the first and second microphones 54 and 55 of the platform 50 and there is no sound pressure and the shape of the platform 50 And the ratio of the free-filed pressure caused by the sound waves radiated from the sound source 40 reaching the center of the head of the platform 50. Therefore, the HRTF reflects the influence of reflection, diffraction, scattering, and the like depending on the shape of the platform 50. The HRTF depends on the position where the sound source 40 is present, that is, the horizontal angle, the altitude angle, and the distance.

Fig. 3 is experimental data showing the HRTF magnitude of the ears according to the horizontal angle change at an altitude angle of 0 DEG for a sound source at a distance of 1 m. 4 is experimental data showing the HRTF size of the left ear according to the elevation angle change at a horizontal angle of 30 degrees with respect to a sound source at a distance of 1 m. 3 and 4 show HRTF measurement results of the platform (dummy head) 50 according to the horizontal angle and altitude angle of the sound source in the anechoic room. The vertical axis is the frequency, and the horizontal axis is the size of the HRTF.

As shown in Figs. 3 and 4, the HRTF has a property of being dependent on the horizontal angle and the altitude angle. According to the present invention, after these features are measured in advance and stored in a database, the positions of the sound sources are estimated by comparing the phase difference and the magnitude ratio of the signals input to the respective microphones with the HRTF database. In order to quantitatively compare the measurement results with the retained HRTF data, the present invention uses a phase scale and a size scale.

,

,

은 오른쪽 마이크로폰과 왼쪽 마이크로폰 출력신호의 코히어런스 함수(coherence function),

는 오른쪽 머리전달함수와 왼쪽 머리전달함수의 위상차,

는 오른쪽 머리전달함수와 왼쪽 머리전달함수의 크기비,

는 추적 대상 음원에 대한 오른쪽 마이크로폰과 왼쪽 마이크로폰 출력신호의 위상차,

: 추적 대상 음원에 대한 오른쪽 마이크로폰과 왼쪽 마이크로폰 출력신호의 크기비이다. 적분 기호는 추정하고자 하는 음원의 주파수 대역만을 선택하여 더하는 것을 의미한다. 한예로, 사람의 음성을 추정하는 경우 음성주파수 대역인 300Hz~3.4kHz 대역의 차이만을 더하는 것이 된다. 이렇게 함으로써 원치 않는 주파수대역의 신호를 배제하고 음원의 위치를 추정할 수 있게 된다. 또한, 코히어런스 함수는 두 신호간의 일관성 혹은 데이터의 신뢰성을 나타내는 척도로써, 측정 신호의 위상차(크기비)와 HRTF 데이터의 위상차(크기비)간의 차이에 상관 함수를 곱해줌으로 인해 신뢰성 있는 정보만을 이용하게 되는 결과를 가져온다. here,

Is the coherence function of the right and left microphone output signals,

Is the phase difference between the right head transfer function and the left head transfer function,

The ratio of the size of the right head transfer function to the size of the left head transfer function,

The phase difference between the right microphone and the left microphone output signal for the tracked sound source,

: The ratio of the size of the right microphone and left microphone output signal to the source being tracked. The integral symbol means that only the frequency band of the sound source to be estimated is selected and added. For example, when estimating a human voice, only the difference of the frequency band of 300 Hz to 3.4 kHz is added. By doing this, it is possible to exclude the signal of the undesired frequency band and to estimate the position of the sound source. The coherence function is a measure of consistency between two signals or reliability of data. The coherence function multiplies the difference between the phase difference (magnitude ratio) of the measurement signal and the phase difference (magnitude ratio) of the HRTF data, It is possible to use only the result.

If the HRTF data corresponding to the position of the sound source to be tracked is used, the difference between the phase difference (size ratio) of the measurement signals and the phase difference (size ratio) using the HRTF data becomes small, The location (horizontal angle, altitude) of the sound source can be estimated by finding the HRTF dataset which minimizes the phase and size scales by obtaining the scale and size scales, respectively.

5A is a phase scale calculated by using HRTF data of a sound source according to a horizontal angle change at an altitude angle of 0 DEG, and FIG. 5B is a phase scale calculated by considering a free sound field. The results of the assumption of free field for comparison are also shown. In the first figure of Fig. 5A, the abscissa axis is the phase scale, the ordinate axis is the HRTF angle, and the other axis is the actual angle. In the second figure, the horizontal axis is the HRTF angle and the vertical axis is the actual angle. Figure b is the same as above.

Referring to FIGS. 5A and 5B, it can be seen that the value of the phase scale is smaller when the head transfer function near the actual sound source position is used. That is, when the phase scale is calculated by changing the HRTF angle with respect to the angle of the actual sound source, the phase scale value becomes smallest when the angle of the head transfer function is equal to the angle of the actual sound source. This shows that the position of the sound source can be estimated by obtaining the head transfer function angle that minimizes the phase scale. However, the value of the phase scale is small even at an angle symmetric to the actual sound source position, which means that front-back confusion occurs when the phase difference is used only when the front-back division becomes ambiguous. Similar to humans, the front-rear of the dummy head is almost symmetrical, resulting in a free-field. However, as shown in the above results, estimating the horizontal angle by considering the head transfer function improves the sound source localization performance than simply estimating the horizontal angle at the free field. Especially when the sound source is in this position, the sound emitted from the sound source hits the head of the dummy head and returns to the opposite microphone for the longest time.

FIG. 6A is a phase scale calculated using HRTF data according to elevation angle change when the horizontal angle is 30 DEG, and FIG. 6B is a scale scale. FIG. 6C is a phase scale calculated using HRTF data according to the elevation angle change when the horizontal angle is 60 DEG, and FIG. 6D is a size scale. 6A, 6B, 6C, and 6D are the HRTF angles and the ordinate is the actual angle.

As shown in Figs. 6A and 6C, when the phase scale is used, the front-to-rear distinction becomes ambiguous in the horizontal angle estimation, Results can be seen. However, the reason why it is not clear as in the case of the horizontal angle change is that the dummy head is not up-and-down symmetrical with respect to the microphone, as a person is not symmetrical up and down with respect to the poppy. Even though the dummy head has almost symmetry before and after the dummy head, it is not symmetrical due to the torso or the like, and the influence is included in the size information of the head transfer function have.

As shown in Figs. 6B and 6D, using the magnitude scale for altitude angle estimation shows better estimation performance. In the conventional free sound field, the horizontal angle and the altitude angle can not be estimated at the same time using the delay time of each microphone.

7 is a flowchart illustrating a sound source position estimation method using a head transfer function database according to an embodiment of the present invention. As shown in FIG. 7, the sound source position estimation method according to the present invention includes a first step (S1) of acquiring a signal for building a database, a signal processing step (S1) A second step S2, a third step S3 of acquiring a sound wave signal generated by the sound source to be tracked through a microphone, a fourth step of processing the signals learned in the third step to calculate a phase difference and / A fifth step S5 of calculating a phase scale and / or a size scale, and a sixth step S6 of estimating a sound source location to be tracked.

In the first step S1, the position (for example, the horizontal angle and the altitude angle) of the sound source 40 on a three-dimensional space (for example, an anechoic room) with respect to a platform (robot, dummy head, And acquires the sound wave signal through the two first and second microphones 54 and 55 while varying the frequency.

In the second step S2, a head transfer function of signals acquired in the first step S1 in a certain range of the audio frequency band is obtained by signal processing, and a database is constructed.

In the third step S3, a sound wave signal generated by the sound source to be tracked is acquired through two microphones.

In the fourth step S4, a phase difference, a magnitude ratio, or a phase difference and a magnitude ratio between the signals acquired in the third step in a certain range of the audible frequency band are calculated by the signal processing. In the fourth step S4, a coherence function, an auto-spectral density function, and a cross-spectral density function can be obtained. Using these functions, Or the size ratio is preferably obtained.

In the fifth step S5, the square of the difference between the phase differences of the head transfer function for each sound source position stored in the database and the phase difference obtained in the fourth step S4 in the second step S2 is multiplied by the square of the coherence function Is integrated in a range of audio frequency bands to obtain a phase scale. In the fifth step, the squared value of the difference between the magnitude ratios of the head transfer function per sound source location stored in the database and the magnitude ratio obtained in the fourth step S4 in the second step S2 is used as the square of the coherence function Can be integrated in a range of audible frequency bands to obtain a magnitude scale. Alternatively, both the phase scale and the magnitude scale may be obtained.

In the sixth step S6, the position of the sound source on the database having the phase scale, the phase scale, the phase scale, and the size scale, which are relatively smaller than the size scale and the scale scale, obtained in the fifth step S5, .

The operation of the sound source position estimation method using the head transfer function database according to the present invention having the above-described structure will be described.

2 and 3, a conventional signal amplifier, a microphone, and a frequency analyzer are used for acquiring and processing a sound signal, but alternatively, alternate equipment capable of signal acquisition and processing and frequency analysis software may be used have.

According to the present invention, a phase scale and a size scale obtained by using a head transfer function measured at a location of a sound source to be tracked include a sound field (for example, an infinite plane) Assuming that the sound field (for example, the infinite plane) on which it lies is exactly the same and there is no error in signal acquisition and processing, theoretically zero. The small phase scale and size scale mean that the source being tracked is near the source location of the head transfer function measurement. Even considering the differences in the sound field and small errors in the signal processing, the functions of these measures are sufficiently maintained.

Although the invention has been described primarily with reference to the above embodiments, many other possible modifications and variations can be made without departing from the spirit and scope of the invention. For example, in the second step, the absolute value of the difference, which is not the square of the difference between the phase differences of the head transfer function for each sound source position stored in the database and the phase difference obtained in the fourth step, It will be readily apparent to those skilled in the art from the above-described embodiments of the invention.

When the sound source location tracking method using the head transfer function according to the present invention is used, the horizontal angle and the altitude angle of a sound source in a three-dimensional space can be simultaneously estimated. Accordingly, a sound source position estimation method applicable to the intelligent robot industry, security CCTV, and remote conference camera is provided. In addition, by using the phase scale and the size scale, it is possible to solve the above-mentioned problems of the conventional art and to estimate the sound source position more accurately than in the prior art.

The scope of the present invention is defined by the following claims, and is not limited to the description of the specification, and all variations and modifications falling within the scope of the claims are included in the scope of the present invention.

Claims (6)

A method for tracking a sound source location, A first step of acquiring a sound wave signal through two microphones with different positions of sound sources on a three-dimensional space with respect to a platform 50 to be applied; A second step of constructing a database by obtaining a head transfer function of the sound signals in a certain range of the audio frequency band by signal processing; A third step of acquiring a sound wave signal generated by the sound source to be tracked through two microphones; A fourth step of calculating a phase difference between signals obtained in the third step in a certain range of the audio frequency band by signal processing; A value obtained by multiplying the square of the difference between the phase differences of the head transfer function and the phase difference obtained in the fourth step stored in the database by the square of the coherence function is integrated in a range of audio frequency bands, A fifth step of obtaining; And And estimating a position of a sound source on a database having a phase measure of a minimum value among the phase measures as a location of a sound source to be tracked. The method according to claim 1, Wherein the fourth step further obtains a coherence function, an auto-spectral density function, and a cross-spectral density function. A method for tracking a sound source location, A first step of acquiring a sound wave signal through two microphones while changing the position of the sound source 40 in a three-dimensional space with respect to a platform 50 to be applied; A second step of constructing a database by obtaining a head transfer function of the sound signals in a certain range of the audio frequency band by signal processing; A third step of acquiring a sound wave signal generated by the sound source to be tracked through two microphones; A fourth step of calculating a size ratio between signals acquired in the third step in a certain range of the audio frequency band by signal processing; A value obtained by multiplying the squared value of the difference between the magnitude of the head transfer function for each sound source position stored in the database and the magnitude ratio obtained in the fourth step by the square of the coherence function is integrated in a certain range of the audio frequency band, (5) And And estimating a position of a sound source on a database having a scale of a minimum value among the size scales as a location of a sound source to be tracked. The method of claim 3, Wherein the fourth step further obtains a coherence function, an auto-spectral density function, and a cross-spectral density function. The method of claim 3, The fourth step further comprises calculating a phase difference between the signals acquired in the third step in a certain range of the audio frequency band by the signal processing, The fifth step is to multiply the squared value of the difference between the phase differences of the head transfer function for each sound source location stored in the database and the phase difference obtained in the fourth step by the square of the coherence function, And obtaining a phase scale by integrating the frequency of the sound source in the audio frequency band of the sound source. 6. The method of claim 5, Wherein the fourth step further obtains a coherence function, an auto-spectral density function, and a cross-spectral density function.

Images