JP2009246827A - Device for determining positions of sound source and virtual sound source, method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device for determining positions of a sound source and virtual sound source that obtains the intended measurement results without using measurement signals, a method and program. <P>SOLUTION: The device for determining the position of the sound source determines the positions of the sound source and virtual sound source by using three or more microphones disposed on a predetermined sound field space. The three or more microphones comprising the number greater than that of spatial dimension include: an acquiring means 11 for acquiring plurality of pairs of acoustic signals comprised of a pair of the acoustic signals from two microphones out of three or more microphones; and a calculating means 14 for calculating a plurality of linear equations by estimating the sound source direction expressed by the linear equations passing a middle point between the microphones in each pair of microphones from a phase difference of each pair of acoustic signals in the acquired plurality pairs of the acoustic signals. Further, the device includes a determining means 15 for determining a minimum square resolution of a simultaneous equation comprising the plurality of the linear equations as the positions of the sound source and virtual sound source. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to sound source identification and estimation of a virtual sound source distribution based on reflected sound, and more particularly to a sound source and virtual sound source position identification device, method, and program.

  In general, when there is reflected sound or stereo reproduction, the sound may be heard from a direction different from the position where the sound source such as a speaker is originally located. For example, for a listener who listens to music recorded in stereo on a music playback medium (for example, a CD) in advance, music is played between two speakers (direct sound sources) arranged in front of the listener, that is, from the front of the listener. Can be heard. In addition, when playing music (direct sound source) in a large concert hall or the like, the sound of a musical instrument played a little behind the listener may be heard. Such a virtual sound source that can be heard from a direction different from the position of the direct sound source is referred to as a virtual sound source.

  In order to provide the listener with effective acoustic effects such as the above-described stereo recording and performance in a large concert hall, it is desirable to objectively specify the positions of these sound sources and virtual sound sources based on the sound.

  On the other hand, four microphones arranged close to each other are arranged in a predetermined sound field space, and an observation signal such as a measurement signal is generated on the performance stage. By measuring the impulse response using this four-point microphone, sound is emitted from the sound source to each microphone. A measurement technique for observing the arrival time and identifying the position of a sound source (a virtual sound source when there is a reflected sound) is known (see Non-Patent Documents 1 and 2, for example). For example, when it is desired to measure a virtual sound source distribution in a three-dimensional space, it is possible to specify a sound source position by determining a distance from a four-point microphone.

  There is also known a technique for estimating the direction of arrival of sound from the phase difference between acoustic signals observed with two microphones (see Non-Patent Document 3, for example).

  A so-called “stencil filter” is used as a technique for estimating the direction of arrival of the sound source from the phase difference between the acoustic signals observed with these two microphones, and the stencil filter processing estimates the direction of arrival of the sound. In order to improve the estimation accuracy at the time of performing, the accumulation process of the phase difference which arises periodically for every frequency is performed.

Yoshio Yamazaki, Satoshi Ishihara, Yukimi Sakurai, Osamu Ebina, Satoshi Ito, “Spatial information obtained from impulse responses of four adjacent points,” Proc. Of the Acoustical Society of Japan, May 1986, pp. 759-760 Yoshio Yamazaki, Atsushi Ito, “Acoustic measurement of concert halls by the proximity four-point method,” JAS Journal, October 1987 issue Chen Liu, Bruce C.I. Wheeler, William D .; 0'Brien, Jr. , Robert C. Bi1ger, Charisa R. Lansing, and Albert S. Feng, “Localization of multiple sources with two microphones,” J. Feng. Acoustic. Soc. Am. 108 (4), 1888-1905.2000.

  In order to measure the virtual sound source distribution in a concert hall or the like by measuring the impulse response with a four-point microphone in the prior art, an observation signal such as a measurement signal is generated on the performance stage, and the sound reaches each microphone from the sound source. It was necessary to observe time. However, generating an observation signal and analyzing it has the problem that it takes a relatively long time for preparation and analysis. In addition, even when generating an observation signal for each frequency that is suitable for the actual performance, analysis is performed in consideration of the effects of sound pressure and range in the actual performance. Can not.

  In addition, the stencil filter processing used to estimate the direction of arrival of a sound from the phase difference of signals observed with two microphones can estimate the “direction of arrival” of a specific sound. Further ingenuity is necessary to identify the “position” of the sound source and grasp the virtual sound source distribution.

  Therefore, it is very efficient if the position of the sound source or virtual sound source can be identified from the direct sound such as the sound being played in a predetermined sound field space and the virtual sound source distribution can be estimated. It is desirable to have a method for identifying

  An object of the present invention is to provide a sound source and a virtual sound source position specifying apparatus, method, and program capable of obtaining a measurement result suitable for a purpose without using a measurement signal.

  A sound source position specifying device according to the present invention is a sound source position specifying device for specifying the positions of a sound source and a virtual sound source using three or more microphones provided in a predetermined sound field space, wherein the three or more microphones are used. Is configured with a number larger than the number of spatial dimensions of the measurement target, and means for acquiring a plurality of sets of acoustic signals, each of which is a set of acoustic signals from two microphones, out of the three or more microphones. Means for calculating a plurality of linear equations by estimating a sound source direction represented by a linear equation passing through a midpoint between microphones in each set of microphones from a phase difference of each set of acoustic signals in a plurality of sets of acoustic signals; And means for specifying a least squares solution of simultaneous equations composed of the plurality of linear equations as positions of the sound source and the virtual sound source.

  Further, in the sound source location device according to the present invention, the means for estimating the plurality of sound source directions is a stencil filter processing means for accumulating a phase difference periodically generated for each frequency with respect to a phase difference between the plurality of sets of sound signals. It is characterized by having. Thereby, a sound source position can be specified more correctly.

  In the sound source location device according to the present invention, at least three of the three or more microphones are configured to be positioned at the apex of a right triangle.

  The sound source location device according to the present invention is characterized in that at least four of the three or more microphones are positioned at the vertices of a cube representing a three-dimensional direction.

  Furthermore, the sound source position specifying method according to the present invention is a sound source position specifying method for specifying the positions of a sound source and a virtual sound source using three or more microphones provided in a predetermined sound field space. The microphones of each set are obtained from the step of acquiring a plurality of sets of acoustic signals, each of which is a set of acoustic signals from two microphones, and the phase difference of each set of acoustic signals in the acquired sets of acoustic signals. A sound source direction represented by a linear equation passing through a midpoint between microphones in the above is estimated, and a step of calculating a plurality of linear equations, and a least square solution of simultaneous equations composed of the plurality of linear equations, a sound source and a virtual sound source And a step of specifying the position as a position.

  Furthermore, the sound source position specifying program according to the present invention is provided on a computer configured as a sound source position specifying device that specifies the positions of a sound source and a virtual sound source using three or more microphones provided in a predetermined sound field space. Of each of the two or more microphones, each set is obtained from a step of acquiring a plurality of sets of acoustic signals, each of which is a set of acoustic signals from two microphones, and a phase difference between each set of acoustic signals in the acquired plurality of sets of acoustic signals. Estimating a sound source direction represented by a linear equation passing through a midpoint between the microphones in the microphone of the microphone, calculating a plurality of linear equations, and obtaining a least squares solution of simultaneous equations composed of the plurality of linear equations, It is characterized as a program for executing the step of specifying the position of the virtual sound source.

  According to the present invention, for example, when measuring a virtual sound source distribution by reflected sound of a concert hall, the position of the sound source and the virtual sound source can be determined without using a measurement signal or the like, even if the distance from the sound source to each microphone is not known. It is possible to detect and measure the virtual sound source distribution.

  In addition, when a certain sound field is reproduced, it can be used to evaluate how much the sound field is similar to the original sound field.

  Hereinafter, a sound source position specifying apparatus according to an embodiment of the present invention will be described. In addition, a sound source position specifying method will be described.

  FIG. 1 is a diagram showing a sound source position specifying apparatus according to an embodiment of the present invention. The sound source position specifying device 1 of the present embodiment acquires sound signals from n (n is a natural number of 3 or more) microphones provided in a predetermined sound field space, and specifies the positions of the sound source and the virtual sound source. It is a device to do. Here, the n microphones include the first microphone 10-1, the second microphone 10-2, the third microphone 10-3, ..., the n-1th microphone 10-n-1, and the nth microphone. These microphones 10-n will be described separately as individual microphones.

  The sound source position identification device 1 of the present embodiment includes an acoustic signal acquisition unit 11, a phase difference calculation unit 12, a stencil filter processing unit 13, a sound source direction estimation unit 14, and a sound source coordinate detection unit 15.

  The acoustic signal acquisition means 11 has a function of acquiring a plurality of sets of acoustic signals, one set of acoustic signals from two microphones among three or more microphones. For example, a set of the first microphone 10-1 and the second microphone 10-2, a set of the second microphone 10-2 and the third microphone 10-3, the first microphone 10-1 and the third microphone An acoustic signal is acquired as a group 10-3.

  As will be apparent from the description below, the number of n microphones increases as the number of microphones increases, and it is preferable that the number of microphones is greater than the number of spatial dimensions to be measured. In order to specify the positions of the sound source and the virtual sound source, it is only necessary to have three, and to specify the positions of the sound source and the virtual sound source in the three-dimensional space, it is only necessary to have four. In order to specify the positions of the sound source and the virtual sound source in the two-dimensional space, it is preferable to arrange and configure three microphones so as to be positioned at the apex of the right triangle. Furthermore, in order to specify the positions of the sound source and the virtual sound source in the three-dimensional space, it is preferable to arrange and configure three microphones so as to be positioned at the vertexes of a cube representing the three-dimensional direction. By configuring in this way, it is possible to detect a sound source or a virtual sound source in an arbitrary direction in each dimension space in a well-balanced manner, and the calculation method for each direction can be similarly configured. Reduction and calculation accuracy can be stably increased.

  The phase difference calculation means 12 calculates a phase difference for each frequency of the acquired plural sets of acoustic signals.

  The stencil filter processing means 13 has a function of accumulating phase differences that occur periodically for each frequency with respect to the phase differences of the plurality of sets of acoustic signals.

  The sound source direction estimating means 14 estimates a plurality of sets of sound source directions represented by a linear equation passing through a midpoint between microphones in each set of microphones from information on the phase difference for each frequency of the acoustic signal subjected to stencil filter processing. It has a function. As will be apparent from the description below, the linear equation estimated from the phase difference of each set of acoustic signals is also a linear equation that passes through the sound source position estimated from the phase difference of each set of acoustic signals.

  The sound source coordinate detecting means 15 has a function of specifying the least square solution of the simultaneous equations composed of the plurality of linear equations as the positions of the sound source and the virtual sound source. The positions of the identified sound source and virtual sound source can be displayed on a display device (not shown) as two-dimensional or three-dimensional spatial coordinates.

  In addition, the sound source position specifying device 1 of the present embodiment can be suitably configured using a computer. For example, the computer is controlled by a control unit (for example, CPU) provided in the computer, acoustic signal acquisition means 11, phase difference calculation means 12, stencil filter processing means 13, sound source direction estimation means 14, and sound source coordinate detection means. 15 can be configured. The acoustic signal acquisition means 11 converts a plurality of acquired acoustic signals into digital signals and stores them in a memory (not shown) provided inside or outside the computer, and the phase difference calculation means 12 A set of acoustic signals is read out, a phase difference for each frequency is calculated, a stencil filter processing unit 13 performs a stencil filter process, and a sound source direction estimating unit 14 estimates a plurality of sound source directions, and each set of microphones. Is expressed by a linear equation passing through the midpoint between microphones. As described above, the sound source position specifying device 1 according to the present embodiment is expressed by a linear equation that passes through the midpoint between the microphones in each set of microphones, based on the phase difference of each set of sound signals in the acquired plurality of sets of sound signals. The sound source direction is estimated, and a plurality of linear equations are calculated. The sound source coordinate detection means 15 calculates a least square solution (solution by the least square method) of the simultaneous equations composed of the plurality of linear equations, specifies the positions of the sound source and the virtual sound source, and specifies the positions of the specified sound source and the virtual sound source. Are displayed on a display device (not shown) as two-dimensional or three-dimensional spatial coordinates. The programs necessary for these controls can be stored in a memory (not shown) provided inside or outside the computer, or can be stored in a computer-readable portable recording medium.

  Next, the operation of the sound source position specifying device 1 of the present embodiment will be described in more detail.

  FIG. 2 shows an operation flow of the sound source position identification device 1 of the present embodiment. First, an acoustic signal acquired by a plurality of microphones is acquired by the acoustic signal acquisition means 11, and two sets of phase differences in the plurality of microphones are calculated for each frequency by the phase difference calculation means 12 (step S1). . FIG. 3 shows an example of the calculated phase difference-frequency pattern.

  Next, the stencil filter processing means 13 performs stencil filter processing for accumulating phase difference patterns that appear periodically for each frequency (step S2). FIG. 4 shows an example of stencil filter processing accumulated at a phase difference of + 60 ° with respect to the phase difference-frequency pattern in FIG. FIG. 5 shows the result of applying the stencil filter process to the total phase difference with respect to the phase difference-frequency pattern in FIG.

  Next, the sound source direction estimation means 14 accumulates, in the frequency direction, the portion corresponding to the frequency band for which the sound source position is to be estimated, out of the phase-frequency patterns subjected to the stencil filter processing for all the phase differences in FIG. The direction of arrival of the sound is estimated (step S3). For example, as shown in FIG. 6 as an example of the sound source direction estimation result, the angle indicating the sound source direction can be represented as a graph indicating the sound arrival direction with the horizontal axis representing the sound source direction and the vertical axis representing the accumulated value.

  Subsequently, the sound source direction estimation means 14 calculates an equation of a function representing a linear equation passing through the midpoint between the sound source and the microphone from the angle indicating the sound source direction (step S4). Here, FIG. 7 shows an example of microphone arrangement when performing sound source position estimation on a two-dimensional plane. In FIG. 7, with the first microphone 10-1 as the central axis O, the second microphone 10-2 (x axis) and the third microphone 10-3 (y axis) are arranged at a distance d, Each microphone is configured to be located at the apex of a right triangle. Also, examples of equations of functions of linear equations passing through the midpoint between the estimated sound source 20 and each microphone in this case are shown in FIGS.

FIG. 8 is a diagram illustrating an example of an equation of a function of a linear equation passing through the estimated sound source 20 and the midpoint between the first microphone 10-1 and the third microphone 10-3. _xo is an angle formed by the perpendicular line on the two-dimensional plane of the line connecting the first microphone 10-1 and the second microphone 10-2 and the line of the linear equation. The specified angle.

FIG. 9 is a diagram illustrating an example of an equation of a function of a linear equation passing through the estimated sound source 20 and the midpoint between the first microphone 10-1 and the third microphone 10-3. _yo is an angle formed by the perpendicular line on the two-dimensional plane of the line connecting the first microphone 10-1 and the third microphone 10-3 and the line of the linear equation. The specified angle.

FIG. 10 is a diagram illustrating an example of an equation of a function of a linear equation passing through the estimated sound source 20 and the midpoint between the second microphone 10-2 and the third microphone 10-3. _xy is an angle formed by a perpendicular line on the two-dimensional plane of a line connecting the second microphone 10-2 and the third microphone 10-3 and the line of the linear equation, and the sound source direction estimating unit 14 The specified angle.

  Even when a larger number of microphones are used, it is preferable to perform the processing of steps S1 to S4 for a set of all microphones. Similarly, in the three-dimensional space, the midpoint between the microphones constituting the specific group, the perpendicular on the two-dimensional surface corresponding to the line connecting the microphones constituting the specific group, and the line of the linear equation are as follows. Each angle is defined and calculated.

  Next, a simultaneous solution of these equations is calculated by the sound source coordinate detection means 15 (step S5), and the coordinates of the sound source position 20 are specified (step S6). When there is no solution of simultaneous equations, it can be obtained as a least square error solution.

FIG. 11 shows a processing flow as a more specific example of the operation of the sound source position estimating apparatus 1 on a two-dimensional plane. Among the n microphones shown in FIGS. 8 to 10, the signals Sx (n), Sy (n), and So (n) of the microphones x, y, and o are acquired (steps S20a to S20c), and a set of each signal is obtained. Are used to estimate the sound source directions φ _xo , φ _yo , and φ _xy (steps S 21 a to S 21 c), calculate simultaneous equations composed of linear equations representing the sound source directions (step S 22), and solve the least square error solution. (Step S23), the coordinates of the sound source position 20 are specified (Step S24).

  FIG. 12 shows the first microphone 10-1 as the central axis O, the second microphone 10-2 (x axis), and the third microphone 10-3 (using the sound source position estimation device 1 on the two-dimensional plane. This is a result (indicated by a black dot) of estimating the position of the sound source 20-1 that emits a noise signal including all frequencies by using three microphones configured and arranged as y-axis). It can be seen that the position of the sound source 20-1 can be estimated almost correctly.

  FIG. 13 also shows that the first microphone 10-1 is the central axis O, the second microphone 10-2 (x-axis), and the third microphone 10- using the sound source position estimation device 1 on the two-dimensional plane. As a result (indicated by black dots) of the positions of the sound sources 20-1 and 20-2 that emit noise signals having two different frequency bands by three microphones configured and arranged as 3 (y-axis) is there. Although the estimation error is included in a part of the frequency band surrounded by the broken line 30, it can be seen that the sound source position is correctly shown in almost all frequency bands. In the case where the estimation error is a problem in a part of the frequency band surrounded by the broken line, the estimation accuracy is increased by increasing the number of microphones or moving the arrangement of the microphones.

  FIG. 14 shows the results when the stencil filter processing is not used in estimating the sound source direction under the conditions of the configuration and arrangement of the sound source and the microphone shown in FIG. It can be seen that the estimation error increases compared to FIG. It can be seen that it is preferable to use stencil filter processing in order to improve accuracy in sound source direction estimation of each microphone set.

  The present invention can be modified or replaced in many ways other than the embodiments described above, and the present invention is not limited to the embodiments described above, and is limited only by the scope of the claims.

  In the present invention, the sound source or virtual sound source position can be identified from the direct sound such as the sound being played in a predetermined sound field space, and the virtual sound source distribution can be estimated. This is useful for estimating the virtual sound source distribution of the sound source distribution.

It is a figure which shows the sound source position identification device of one Example by this invention. It is a figure which shows the operation | movement flow of the sound source position identification apparatus of one Example by this invention. It is a figure which shows an example of the calculated phase difference-frequency pattern in the sound source location device of one Example by this invention. It is a figure which shows an example of the stencil filter process accumulate | stored in phase difference +60 degrees with respect to the phase difference-frequency pattern in the sound source position identification device of one Example by this invention. It is a figure which shows an example which performed the stencil filter process with respect to all the phase differences with respect to the phase difference-frequency pattern in the sound source position identification device of one Example by this invention. It is a figure which shows an example of the sound source direction estimation result in the sound source position specific device of one Example by this invention. It is a figure which shows the example of microphone arrangement | positioning in the case of performing the sound source position estimation on the two-dimensional surface in the sound source position identification device of one Example by this invention. It is a figure which shows the example of the equation of the function of the linear equation which passes along the midpoint between the sound source and each microphone in the sound source location device of one Example by this invention. It is a figure which shows the example of the equation of the function of the linear equation which passes along the midpoint between the sound source and each microphone in the sound source location device of one Example by this invention. It is a figure which shows the example of the equation of the function of the linear equation which passes along the midpoint between the sound source and each microphone in the sound source location device of one Example by this invention. It is a figure which shows the processing flow on the two-dimensional surface in the sound source location specification apparatus of one Example by this invention. It is a figure which shows the result of having estimated the position of the sound source which emits the noise signal containing all the frequencies on the two-dimensional surface in the sound source position identification device of one Example by this invention. It is a figure which shows the result of having estimated the position of the sound source which emits the noise signal which has two different frequency bands on the two-dimensional surface in the sound source position identification device of one Example by this invention. It is a figure which shows the result at the time of not using a stencil filter process in the case of estimation of the sound source direction in the sound source position identification device of one Example by this invention.

Explanation of symbols

1 sound source position specifying device 10-1 first microphone 10-1
10-2 2nd microphone 10-3 3rd microphone 10-n-1 n-1 microphone 10-n nth microphone 11 Acoustic signal acquisition means 11
12 Phase difference calculating means 12
13 stencil filter processing means 14 sound source direction estimating means 15 sound source coordinate detecting means 20-1, 20-2 sound source

Claims (6)

A sound source position specifying device for specifying the position of a sound source and a virtual sound source using three or more microphones provided in a predetermined sound field space,
The three or more microphones are configured with a number larger than the number of spatial dimensions to be measured,
Means for acquiring a plurality of sets of acoustic signals, wherein the acoustic signals from two microphones are one set out of the three or more microphones;
Means for calculating a plurality of linear equations by estimating a sound source direction represented by a linear equation passing through a midpoint between microphones in each set of microphones from the phase difference of each set of acoustic signals in the acquired plurality of sets of acoustic signals. When,
Means for specifying a least-squares solution of simultaneous equations composed of the plurality of linear equations as positions of the sound source and the virtual sound source;
A sound source position specifying device comprising:   The means for estimating the plurality of sets of sound source directions includes stencil filter processing means for accumulating phase differences generated periodically for each frequency with respect to phase differences of the plurality of sets of sound signals. The sound source location device described in 1.   3. The sound source location device according to claim 1, wherein at least three microphones of the three or more microphones are configured to be located at vertices of a right triangle.   3. The sound source localization device according to claim 1, wherein at least four of the three or more microphones are configured to be located at the vertices of a cube representing a three-dimensional direction. . A sound source position specifying method for specifying positions of a sound source and a virtual sound source using three or more microphones provided in a predetermined sound field space,
Obtaining a plurality of sets of acoustic signals in which the acoustic signals from two microphones are one set out of the three or more microphones;
A step of calculating a plurality of linear equations by estimating a sound source direction represented by a linear equation passing through a midpoint between microphones in each set of microphones from a phase difference of each set of acoustic signals in the acquired plurality of sets of acoustic signals. When,
Identifying a least squares solution of simultaneous equations composed of the plurality of linear equations as the positions of the sound source and the virtual sound source;
A sound source position specifying method comprising: A computer configured as a sound source position specifying device that specifies the position of a sound source and a virtual sound source using three or more microphones provided in a predetermined sound field space,
Obtaining a plurality of sets of acoustic signals in which the acoustic signals from two microphones are one set out of the three or more microphones;
A step of calculating a plurality of linear equations by estimating a sound source direction represented by a linear equation passing through a midpoint between microphones in each set of microphones from a phase difference of each set of acoustic signals in the acquired plurality of sets of acoustic signals. When,
Identifying a least squares solution of simultaneous equations composed of the plurality of linear equations as the positions of the sound source and the virtual sound source;
Sound source location program to execute.

Images