JP2016092562A - Audio processing device and method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain audio having sound pressure distribution with less variation.SOLUTION: An acoustic sensor arranged at each of a plurality of control points collects audio output from a sound source device on the basis of a measurement acoustic signal and outputs a measurement signal. A transfer function calculator calculates a transfer function from the sound source device to the acoustic sensor on the basis of the measurement signal and the measurement acoustic signal. An information input unit sets a control point in a listening area and a non-listening area. A filter coefficient calculation unit calculates an input filter coefficient by solving a linear programming problem for determining, on the basis of the transfer function and the setting result of each control point, the input filter coefficient for minimizing the maximum value of sound pressure in the non-listening area under the condition that an error of sound pressure in the listening area is within an allowable error range. The technology can be applied to an acoustic system.SELECTED DRAWING: Figure 1

Description

  The present technology relates to a sound processing apparatus and method, and a program, and more particularly, to a sound processing apparatus and method, and a program capable of obtaining sound with a sound pressure distribution with less variation.

  For example, when a sound source such as a general speaker is used to output an acoustic signal into the space, almost all people around it listen to the sound output from the sound source, although it is affected by the directivity characteristics of the sound source. Is possible.

  Therefore, a space or direction in which sound can be heard (hereinafter also referred to as a listening area) and a space or direction in which sound cannot be heard (hereinafter also referred to as a non-listening area) are set, and the sound is limited to those who are in the listening area. In order to achieve such a purpose, some device for the playback device is required.

  If it is possible to reproduce the sound limited to a specific position, it is possible to prevent the sound from becoming a noise to a person other than the listener, and a plurality of listeners are individually separated in one space. An acoustic signal can be heard. It is conceivable to use a directional speaker system that emits sound waves in a desired direction as means for realizing such a requirement.

  Here, the directional speaker system is equipped with a horn speaker that creates directivity mainly by its geometric shape, a parametric speaker that realizes narrow directivity by self-demodulation of an amplitude-modulated strong ultrasonic beam, and multiple speakers. Thus, there is a speaker array that forms directivity as a whole.

  Among these directional speaker systems, a horn speaker must have a large geometric structure in order to obtain directivity at a low frequency, and its application is limited.

  Moreover, a parametric speaker using ultrasonic waves can easily obtain directivity in a narrow range, but is insufficient in sound quality due to a lack of sound pressure in a low frequency range. Furthermore, since the ratio of the self-demodulated sound wave to the radiated ultrasonic wave is low in the parametric speaker, it is necessary to allow a very strong ultrasonic wave exposure to ensure a sufficient sound pressure.

  On the other hand, in the method using the speaker array, the directivity can be obtained while being smaller than the horn speaker, and since the sound in the audible band is used, no ultrasonic exposure occurs. Furthermore, in the method using the speaker array, it is possible to change the directivity without changing the arrangement of the speaker array by changing the relative amplitude and phase characteristics between the input signals input to the speakers using a filter. is there.

  However, in the method using the speaker array, it is not obvious how the desired directional characteristics can be obtained by changing the amplitude and phase characteristics, and the way the speakers are arranged has a great influence on the directivity. For this reason, various filter coefficient calculation methods and speaker arrangement methods have been proposed.

  For example, the average acoustic energy density in the listening area and the non-listening area and the difference between the average acoustic energy densities are calculated, and the ratio of the difference in acoustic energy density to the total acoustic energy output from the sound source is maximized. There has been proposed a method for calculating a signal radiated from (see, for example, Patent Document 1). According to this method, only the listener in the listening area can hear the sound when the sound is reproduced by the speaker array.

Japanese Patent No. 5073724

  However, in the above-described technique, calculation is performed using the average acoustic energy density of the listening area and the non-listening area, so that spatial variation occurs in the sound pressure distribution in the listening area and the non-listening area.

  The present technology has been made in view of such a situation, and makes it possible to obtain sound with a sound pressure distribution with less variation.

  An audio processing device according to an aspect of the present technology sets an area setting unit that sets several control points in a non-listening area, and sets each control point in the listening area. A target sound pressure setting unit that sets a target sound pressure at a control point; an allowable error setting unit that sets an allowable error of the target sound pressure; and a predetermined number of sound sources less than the total number of all the control points On the basis of the transfer function between the control point and each of the control points, under the condition that the error of the sound pressure of the control point in the listening area with respect to the target sound pressure falls within the allowable error range. A filter coefficient calculation unit for calculating a filter coefficient of a filter corresponding to the sound source that minimizes the sound pressure of the control point in the non-listening area.

  The filter coefficient calculation unit may calculate the filter coefficient by further adding a constraint that the gain of the filter does not exceed a predetermined upper limit value.

  The area setting unit is configured to set some control points in other areas different from the listening area and the non-listening area, and the filter coefficient calculation unit is configured to generate sound of the control points in the other areas. The filter coefficient can be calculated by further adding a constraint that the pressure does not exceed a predetermined maximum sound pressure.

  The maximum sound pressure may be determined based on a minimum value of the target sound pressure at the control point in the listening area.

  The filter coefficient calculation unit can calculate the filter coefficient by solving a linear programming problem.

  An audio processing method or program according to one aspect of the present technology sets several control points in a listening area, sets several control points in a non-listening area, and sets each control point in the listening area. A target sound pressure, a tolerance of the target sound pressure, and a transfer function between the sound source and each of the control points for a predetermined number of sound sources less than the total number of all the control points. The sound pressure of the control point in the non-listening area is minimized under the condition that the error of the sound pressure of the control point in the listening area with respect to the target sound pressure is within the allowable error range. Calculating a filter coefficient of a filter corresponding to the sound source.

  In one aspect of the present technology, several control points are set in the listening area and several control points are set in the non-listening area, and the target sound pressure of each control point in the listening area is set. Is set, an allowable error of the target sound pressure is set, and for a predetermined number of sound sources less than the total number of all the control points, based on a transfer function between the sound source and each of the control points, Minimizing the sound pressure of the control point in the non-listening area under the condition that the error of the sound pressure of the control point in the listening area with respect to the target sound pressure is within the allowable error range, A filter coefficient of a filter corresponding to the sound source is calculated.

  According to one aspect of the present technology, it is possible to obtain sound having a sound pressure distribution with less variation.

  Note that the effects described here are not necessarily limited, and may be any of the effects described in the present disclosure.

It is a figure which shows the structural example of an acoustic system. It is a flowchart explaining a filter coefficient calculation process. It is a flowchart explaining audio | voice reproduction | regeneration processing. It is a figure which shows the structural example of an acoustic system. It is a flowchart explaining audio | voice reproduction | regeneration processing. It is a figure explaining a use case. It is a figure explaining a use case. It is a figure which shows the structural example of a computer.

  Hereinafter, embodiments to which the present technology is applied will be described with reference to the drawings.

<First Embodiment>
<Outline of this technology>
This technology changes the sound field characteristics of a space by changing the relative amplitude and phase characteristics among a plurality of speakers so that an acoustic signal can be heard only within a predetermined desired listening area. is there.

  First, an outline of the present technology will be described.

(About the basic principle of this technology)
For example, consider a certain frequency component ω included in an acoustic signal. On the sound reproduction space based on the acoustic signal, a listening area for allowing the listener to listen to the sound and a non-listening area for preventing the sound from being heard are provided, and the listening area is represented as V _target, and the non-listening area. _Is expressed as V _cancel . Further, it is assumed that N speakers exist as sound sources in the reproduction space.

Here, the transfer function from the n-th speaker among the N speakers (where 0 ≦ n ≦ N−1) to a predetermined position r within the listening area V _target or the non-listening area V _{cancel is} represented by H _n. (r). Also, let W _n be an input filter coefficient used for filter processing performed on the acoustic signal supplied to the nth speaker. In other words, it is assumed that the input acoustic signal is filtered using the input filter coefficient W _n and then reproduced by a speaker.

Here, the input filter coefficient W _n may be any value, but here, it is assumed that the coefficient is a coefficient for each frequency component, that is, a matrix having a coefficient multiplied by each frequency component as an element. To do. Further, hereinafter, it will be referred to as both an input filter filters consisting of the input filter coefficients W _n.

Further, a desired sound pressure characteristic in the listening area V _target , that is, a target sound pressure level is d (r), and an allowable error from the sound pressure characteristic d (r) is σ. Here, the sound pressure characteristic d (r) is a complex sound pressure.

  At this time, the directivity characteristic D (r) at a predetermined position r of the speaker array composed of N speakers, that is, the sound pressure that would actually be obtained is expressed by the following equation (1).

That is, the directivity characteristic D (r) can be represented by the sum of products of the transfer function H _n (r) of each of the N speakers and the input filter coefficient W _n .

Therefore, the maximum value minimization problem shown in the following equation (2) using the input filter coefficient W _n as a variable is considered.

Maximum minimization problem shown in equation (2), the sound pressure .SIGMA.H _n (r) W _n (where position r within the listening area V _target) in the listening area V _target and the desired sound pressure characteristic d (r) within the error tolerances σ between, under the condition that that is in the range of allowable error, non-listening area V _cancel sound pressure at ΣH _n (r) W _n (where position r is in a non-listening area V _cancel) To obtain an input filter coefficient W _n that minimizes. More specifically, the input filter coefficient W _n that minimizes the sound pressure at the position r where the sound pressure is maximum in the non-listening area V _cancel is obtained.

  The sound pressure characteristic d (r) and the allowable error σ may be determined for each position r, or the same one may be used at several positions r.

If this maximum value minimization problem is solved and the input filter coefficient W _n is obtained, the frequency characteristics and amplitude phases between the sounds output from the speakers constituting the speaker array by the filter processing using the input filter coefficient W _n Properties can be controlled.

Thus, compared to the sound pressure of the sound heard at the listening area V _target, to reduce the sound pressure of the sound heard by the non-listening area V _cancel, so as to be able to listen to audio only listening area V _target can do.

In addition, in the maximum value minimization problem shown in Expression (2), the sound quality variation in the listening area V _target that is relatively difficult to cause human hearing problems is allowed by the allowable error σ. Therefore, it is possible to suppress the sound pressure in the non-listening area V _cancel , which is relatively problematic from the viewpoint of hearing, to the maximum within the permissible range, and to suppress variations in spatial sound pressure in the listening and non-listening areas. At the same time, higher separation performance can be obtained.

Further, in this maximum minimization problem, by adjusting the tolerance σ sound pressure at the listening area V _target, easily a trade-off between sound quality and non-listening area V _cancel sound in pressure in the listening area V _target Can be adjusted.

By the way, the maximum value minimization problem shown in Equation (2) is difficult to mathematically handle because it includes a non-linear element such as a complex absolute value calculation. Therefore, if this maximum value minimization problem is transformed into a linear programming problem, the input filter coefficient W _n can be easily obtained.

In performing the transformation to the linear planning problem, first, the maximum value δ of the sound pressure of the voice in the non-listening area V _cancel is expressed by the following equation (3).

  Then, using the maximum value δ shown in Equation (3), Equation (2) described above is rewritten as shown in Equation (4) below.

Here, | ΣH _n (r) W _n | ≦ δ in Expression (4) represents a constraint that the sound pressure at each position r in the non-listening area V _cancel does not exceed the maximum value δ.

  Next, according to the real rotation theorem, the following equation (5) holds for any complex number Z. Using this equation (5), the condition relating to the absolute value of the complex number in equation (4) is converted into a real number condition. can do. Thereby, Formula (4) is deform | transformed into the following formula | equation (6).

  In Equations (5) and (6), θ represents a complex phase, that is, a rotation angle. J represents an imaginary unit, and Re {} represents a real part of a complex number.

Further, it is assumed that an imaginary part of a complex number is represented by Im {}, and a function _An (r) having a position r as a variable is represented by the following equation (7). In Equation (7), m is a natural number.

Further, a function a (r) corresponding to the position r is represented by the following equation (8), and a row vector A (2N + 1) having the function A _n (r) represented by equation (7) as an element: r, θ) is represented by equation (9).

Further, a column vector x having a size 2N + 1 having the real part and the imaginary part of each input filter coefficient W _n and the maximum value δ as elements is shown in Expression (10), and a row vector having a size 2N + 1. It is assumed that c is expressed by equation (11), and a function b (r, θ) having variables of position r and rotation angle θ is expressed by equation (12). In Expression (11), the values of the elements other than the last element of the row vector c are 0.

  At this time, using the row vector c, the column vector x, the row vector A (r, θ), and the function b (r, θ), the equation (6) is expressed as a linear semi-infinite programming problem represented by the following equation (13). Can be represented.

  As a result, the maximum value minimization problem shown in Expression (2) is transformed into a linear semi-infinite planning problem shown in Expression (13).

In the linear semi-infinite programming problem shown in Equation (13), the position r and the rotation angle θ are continuous values, and thus there are an infinite number of constraints. By discretizing the position r and the rotation angle θ, The linear semi-infinite programming problem expressed by Equation (13) can be treated as a normal linear programming problem. Thus, the optimal input filter coefficient W _n can be calculated by calculating the linear programming problem by a linear programming method such as the main dual interior point method.

Here, the discretization of the rotation angle θ is related only to the approximation accuracy of the actual rotation theorem. On the other hand, in the discretization of the position r, both the listening area V _target and the non-listening area V _cancel are discretized at the control point that is the position r for controlling the sound pressure, and the sound pressure of the sound field in the space is determined. The distribution is represented by the value of a sensor such as a microphone arranged at the control point. Therefore, discretization of the position r has a great influence on the properties of the obtained solution. Specifically, when the number of sensors such as microphones is equal to or less than the number of sound sources such as speakers, the control point in the non-listening area V _cancel can be used even under the condition that the tolerance σ = 0 in the listening area V _target . There may be a solution x with a maximum sound pressure of 0.

However, since the solution x thus obtained can be a solution where the sound pressure is zero at a certain position of the sensor as in the standing wave section, the non-listening area V _cancel other than the position where the sensor is located. The sound pressure at is not necessarily minimized. Therefore, it is required to arrange at least one sensor more than the number of sound sources.

(Stabilization of input filter)
Moreover, it is possible to adjust the property of the input filter obtained by adding another constraint condition to the linear programming problem obtained from Expression (13). For example, the upper limit can be set for the gain of the input filter by adding the constraint condition shown in the following equation (14) as the constraint condition of the linear programming problem.

In Expression (14), g _max is a scalar value and indicates the upper limit value of the input filter coefficient W _n . Therefore, the constraint condition expressed by the equation (14) is a condition that the size (absolute value) of each element of the input filter coefficient W _n does not exceed the upper limit value g _max .

  The constraint condition represented by the equation (14) includes the absolute value of a complex number, but this constraint condition can be obtained by performing the same modification as the case where the maximum value minimization problem described above is transformed into a linear programming problem. Can be easily incorporated into linear planning problems.

Thus, by setting the upper limit value g _max for the gain of the input filter (input filter coefficient W _n ), a sound source device such as a speaker or an amplifier is used to reproduce sound based on the input acoustic signal. Thus, it is possible to avoid an input filter having a frequency band with a very high gain that may cause problems.

  That is, depending on the type of sound source such as a speaker, there is a problem that sound is broken when a signal having a high amplitude component in a specific frequency band is input. Therefore, by adding the constraint condition regarding the upper limit value of the filter gain shown in Expression (14), the sound quality in the listening area is ensured while suppressing the occurrence of such a problem, and the sound pressure in the non-listening area is minimized. be able to.

The input when the filter coefficients W _n to set the upper limit value g _max, since the size of each element of the input filter coefficient W _n is equal to or less than the upper limit value g _max, the input filter frequency axis excessive gain Variation can be suppressed. As a result, an input filter with a short impulse response and easy mounting can be obtained. In other words, by setting the upper limit value g _max , it is possible to obtain a stable filter that has a small number of taps when viewed in the time domain and does not cause aliasing or the like as an input filter.

Such an effect of suppressing the variation in the gain of the input filter is a trade-off with the sound pressure of the sound in the non-listening area obtained by solving the above-described linear planning problem. That is, if the value of the upper limit value g _max is made smaller, variations in gain can be suppressed, and conversely, if the value of the upper limit value g _max is made larger, the maximum value of the sound pressure in the non-listening area is made smaller, The sound pressure difference between the listening area and the non-listening area, that is, separation performance can be improved.

Therefore, by adjusting the value of the upper limit value g _max of the input filter coefficient W _n , it is easy to make a trade-off between the calculated stability of the input filter and the separation performance of the obtained listening area and non-listening area. Will be able to adjust.

(Regarding non-limited area settings and constraint conditions)
In addition, a space or direction that is not included in the above-described listening area V _target or non-listening area V _cancel is set as another area (hereinafter also referred to as a non-limited area), and another restriction condition is set for the non-limited area. You may do it.

Here, the non-limited area is, for example, a space or a direction that is not the listening area V _target but has a low necessity of actively reducing the sound pressure.

  As an example, let us consider that it is possible to listen to different acoustic signals (voices) in a plurality of listening areas.

  As a method for enabling different sounds to be heard in a plurality of listening areas as described above, the following realization methods can be considered. That is, first, a listening area is set on the reproduction space for each sound signal, and an input filter is calculated for each sound signal with a space other than the listening area as a non-listening area. Then, filtering processing using an input filter may be performed on each acoustic signal, and the resultant acoustic signals may be added and supplied to a speaker or the like as a sound source to reproduce sound.

  Here, when attention is paid to one acoustic signal, a space that is set as a listening area for another acoustic signal different from the acoustic signal of interest in the reproduction space is calculated when calculating an input filter of the acoustic signal of interest. It is a true non-listening area. However, a space that is not a listening area for any acoustic signal in the reproduction space can be a non-limited area.

  In this case, by allowing a certain amount of sound leakage in the non-limited area, the degree of freedom of the variable is reduced to the suppression of the sound pressure in the true non-listening area, and it can be expected to improve the separation performance of each acoustic signal. .

_Assuming that the non-limited area is V _otherwise , for example, as a constraint condition in the non-limited area V _otherwise , the following equation (15) can be considered.

Note that d _min in Expression (15) indicates the _target sound pressure of the sound in the listening area V _target , that is, the minimum value of the sound pressure characteristic d (r), as shown by the following Expression (16).

Therefore, the constraint condition shown in Expression (15) is that the sound pressure at each position r in the non-limited area V _otherwise , that is, at each control point, is the minimum of the target sound pressure in the listening area V _target considering the allowable error σ. The condition is that the value is below (not exceeding) the value.

In other words, the constraints shown in equation (15), the constraint that the sound pressure of the respective control points within the non-limiting area V _otherwise does not exceed a predetermined maximum sound pressure determined for non-limiting area V _otherwise It becomes a condition. In this example, d _min -σ, which is the right side of Equation (15), indicates the maximum sound pressure.

The maximum sound pressure is determined based on the minimum value of the target sound pressure in the listening area V _target , such as αd _min (where α <1), that is, the minimum sound pressure in the listening area V _target . As long as it does not exceed, it may be anything. Further, the maximum sound pressure determined for the non-limited area V _otherwise may be a different value for each control point. In such a case, for example, the maximum sound pressure can be set stepwise for each of the plurality of non-limited areas V _otherwise .

  The constraint condition shown in the equation (15) is also applied to the linear programming problem easily obtained from the equation (13) by performing the same modification as the case where the maximum value minimization problem is transformed into the linear programming problem. It can be incorporated as a constraint.

  The constraint condition shown in the equation (15) is that a non-limited area is simply set, and if there is no particular constraint condition in the non-limited area, strong sound pressure may be generated in an area other than the listening area. This is a condition for preventing such a situation.

  As described above, the unrestricted area is set in the reproduction space, and the constraint condition in the unrestricted area is further added to the above-mentioned linear planning problem, thereby preventing the occurrence of an unintended high sound pressure area and the optimum input filter. You will be able to get

<Example of acoustic system configuration>
Next, specific embodiments of the present technology described above will be described.

  FIG. 1 is a diagram illustrating a configuration example of an embodiment of an acoustic system to which the present technology is applied.

  The acoustic system 11 shown in FIG. 1 includes a filter calculation unit 21 and an audio reproduction unit 22, and the acoustic system 11 obtains a transfer function and an input filter in advance by prior measurement before reproducing the audio. It is a feedforward type system.

  That is, the filter calculation unit 21 calculates an input filter based on various types of input information and transfer functions, and supplies the input filter to the audio reproduction unit 22.

  The audio reproduction unit 22 is arranged in a reproduction space, for example, and performs filtering processing on the supplied input acoustic signal using the input filter supplied from the filter calculation unit 21, and audio is generated based on the output acoustic signal obtained as a result. Is output.

  In this example, the reproduction space is divided into three areas R1 to R3. Area R1 and area R3 are non-listening areas that are spaces or directions in which sound should not be reproduced, and area R2 is a listening area that is a space or direction in which sounds are to be reproduced.

  The filter calculation unit 21 includes switches 31-1 to 31-N, D / A (Digital / Analog) converters 32-1 to D / A converters 32-N, sound source devices 33-1 to 33-N. , Acoustic sensor 34-1 to acoustic sensor 34-K, transfer function calculator 35, information input unit 36, and filter coefficient calculation unit 37.

  Each of the switches 31-1 to 31-N has an output terminal connected to the D / A converter 32-1 to D / A converter 32-N, and according to the control of the filter calculation unit 21. Turn on or off. When the switches 31-1 to 31-N are in the on state, the supplied measurement acoustic signals are supplied to the D / A converter 32-1 to D / A converter 32-N.

  The D / A converter 32-1 to D / A converter 32-N performs D / A conversion on the measurement acoustic signals supplied via the switches 31-1 to 31-N, and generates a sound source device. 33-1 to the sound source device 33-N.

  The sound source device 33-1 to sound source device 33-N include, for example, speakers that emit sound (acoustic signals) in the space, and are supplied from the D / A converter 32-1 to D / A converter 32-N. Based on the measured measurement acoustic signal, the transfer function measurement sound is reproduced.

  Hereinafter, when it is not necessary to particularly distinguish the switches 31-1 to 31 -N, they are also simply referred to as switches 31, and it is necessary to particularly distinguish the D / A converters 32-1 to 32-N. When there is no signal, it is also simply referred to as a D / A converter 32. Further, hereinafter, the sound source device 33-1 to the sound source device 33-N are also simply referred to as the sound source device 33 when it is not necessary to particularly distinguish them. Here, the number N of the sound source devices 33 is 2 or more.

  The acoustic sensors 34-1 to 34-K are composed of, for example, a microphone and the like, pick up the sound output from the sound source device 33, and supply the measurement signal obtained as a result to the transfer function calculator 35. Hereinafter, the acoustic sensors 34-1 to 34-K are also simply referred to as acoustic sensors 34 when it is not necessary to distinguish them.

  In the space where the filter calculation unit 21 is arranged, for example, L control points that are points (positions) for controlling the sound pressure of sound are set in areas corresponding to areas R1 and R3 that are non-listening areas. Is done. Similarly, M control points are set in an area corresponding to the area R2 that is the listening area in the space where the filter calculation unit 21 is arranged. Here, L + M = K. In the following, the area on the space where the filter calculation unit 21 corresponding to the listening area and the non-listening area on the reproduction space is also simply referred to as a listening area and a non-listening area.

  Further, an acoustic sensor 34 is arranged at each control point. At this time, the total number of all control points, that is, the total number K of all acoustic sensors 34 is set to be equal to or greater than the total number N of all sound source devices 33. This is to ensure that the sound pressure in the non-listening area other than the control point is minimized as described above. In more detail, the information input unit determines which control point is set in which area, in other words, whether the position of each acoustic sensor 34 is set as a listening area or a non-listening area. 36.

The transfer function calculator 35 is provided between each sound source device 33 and the acoustic sensor 34 based on the measurement signal supplied from the acoustic sensors 34-1 to 34-K and the supplied measurement acoustic signal. A transfer function H _n (r) indicating transfer characteristics is calculated and supplied to the filter coefficient calculation unit 37. In the transfer function calculator 35, a transfer function is obtained for each sound sensor 33 for each acoustic sensor 34. That is, a total of N × K transfer functions are obtained.

  The information input unit 36 performs various settings necessary for calculating the input filter in accordance with an operation input by the administrator, and supplies the setting result to the filter coefficient calculation unit 37. The information input unit 36 includes an area setting unit 41, a target sound pressure setting unit 42, and an allowable error setting unit 43.

  The area setting unit 41 sets K control points on the space in accordance with, for example, an operation input by the administrator. Specifically, several control points are set in various areas such as a listening area, a non-listening area, and a non-limited area, and the position r of the set control points and the type of the area including the control points are set. Information indicating the control point setting result and the like.

  The target sound pressure setting unit 42 controls the control points set in the listening area among the K control points, that is, the M control points described above, for example, in response to an operation input by the administrator. Set the target sound pressure at the point. Here, the target sound pressure at the control point in the listening area is the sound pressure characteristic d (r) described above.

  The allowable error setting unit 43 sets an allowable error σ of the target sound pressure at each control point for each control point in the listening area, for example, according to an operation input by the administrator.

  The information input unit 36 uses the area setting unit 41 to the allowable error setting unit 43 to set the result of the setting, that is, the position of each control point, the type of area in which each control point is located, the target sound pressure, and the allowable error. 37.

The filter coefficient calculation unit 37 calculates an input filter based on the transfer function H _n (r) supplied from the transfer function calculator 35 and various setting results supplied from the information input unit 36, and the sound reproduction unit 22 is supplied.

  The audio reproduction unit 22 arranged in the reproduction space includes an amplitude phase control filter processing unit 61-1 to an amplitude phase control filter processing unit 61-N, a D / A converter 62-1 to a D / A converter 62-. N, and sound source devices 63-1 to 63-N.

  An input acoustic signal of the sound to be reproduced is supplied to the amplitude phase control filter processing unit 61-1 to the amplitude phase control filter processing unit 61-N.

  The amplitude phase control filter processing unit 61-1 to the amplitude phase control filter processing unit 61-N perform filter processing using the input filter supplied from the filter coefficient calculation unit 37 on the supplied input acoustic signal, The output acoustic signal obtained as a result is supplied to the D / A converter 62-1 to D / A converter 62-N. The relative amplitude and phase between the N input acoustic signals are controlled by the filter processing in the amplitude phase control filter processing unit 61-1 to the amplitude phase control filter processing unit 61-N.

  The D / A converter 62-1 through D / A converter 62-N perform D / A conversion on the output acoustic signals supplied from the amplitude phase control filter processing unit 61-1 through amplitude phase control filter processing unit 61-N. A conversion is performed and supplied to the sound source device 63-1 to 63-N.

  The sound source devices 63-1 to 63-N include, for example, speakers that emit sound into the space, and output acoustic signals supplied from the D / A converter 62-1 to D / A converter 62-N. Play audio based on Here, each of the sound source devices 63-1 to 63-N corresponds to each of the sound source devices 33-1 to 33-N.

  Hereinafter, when it is not necessary to distinguish between the amplitude phase control filter processing unit 61-1 to the amplitude phase control filter processing unit 61-N, the amplitude phase control filter processing unit 61-N is also simply referred to as an amplitude phase control filter processing unit 61, and The 1 to D / A converters 62 -N are also simply referred to as D / A converters 62 when it is not necessary to distinguish them. Hereinafter, the sound source device 63-1 to the sound source device 63-N are also simply referred to as the sound source device 63 when it is not necessary to distinguish between them.

  In the acoustic system 11 shown in FIG. 1, since it is not necessary to arrange the acoustic sensor 34 for obtaining the transfer function in the reproduction space, the user listens to the voice at a desired position in the reproduction space or moves to the desired position. Or more free listening.

<Description of filter coefficient calculation processing>
Next, the operation of the acoustic system 11 shown in FIG. 1 will be described.

  For example, the filter calculation unit 21 of the acoustic system 11 calculates the input filter by performing filter coefficient calculation processing in advance when the audio reproduction unit 22 reproduces sound in the reproduction space, and supplies the input filter to the audio reproduction unit 22.

  Hereinafter, the filter coefficient calculation processing by the acoustic system 11 will be described with reference to the flowchart of FIG.

  In step S11, the area setting unit 41 performs area setting for K control points arranged in the space. That is, the area setting unit 41 sets various types of areas such as a listening area and a non-listening area on the reproduction space as necessary, and on the space where the filter calculation unit 21 corresponding to each area is arranged. Set control points in the area.

  In this example, some control points are set in the areas corresponding to the non-listening areas R1 and R3, and some control points are set in the areas corresponding to the listening area R2. .

  In step S12, the target sound pressure setting unit 42 sets, for each control point in the listening area, the sound pressure characteristic d (r) at those control points as the target sound pressure.

  In step S <b> 13, the allowable error setting unit 43 sets the allowable error σ of the target sound pressure at each control point for each control point in the listening area. The allowable error σ at each control point may be the same value or a different value.

  When the above processing is performed, the information input unit 36 calculates the filter coefficient of the position r of each control point, the type of area where each control point is, the sound pressure characteristic d (r) as the target sound pressure, and the allowable error σ. To the unit 37.

  In step S <b> 14, the filter calculation unit 21 selects one sound source device 33 that has not yet been processed from among the N sound source devices 33. For example, the sound source devices 33 to be processed are sequentially selected one by one from the sound source device 33-1 to the sound source device 33-N.

  When the filter calculation unit 21 selects one sound source device 33 to be processed, the filter calculation unit 21 controls the opening and closing of the switch 31 so that the measurement sound signal is supplied only to the sound source device 33.

  That is, the filter calculation unit 21 turns on the switch 31 connected to the sound source device 33 to be processed, and turns off the switch 31 connected to the sound source device 33 that is not another processing target.

  In step S15, the sound source device 33 outputs a transfer function measurement sound.

  That is, the D / A converter 32 performs D / A conversion on the measurement acoustic signal supplied via the switch 31 and supplies it to the sound source device 33. As a result, the measurement acoustic signal is converted from a digital signal to an analog signal. The sound source device 33 outputs a transfer function measurement sound based on the measurement acoustic signal supplied from the D / A converter 32.

  Here, since the measurement acoustic signal is supplied only to the sound source device 33 to be processed by the opening / closing control of the switch 31 by the filter calculation unit 21, the sound for transfer function measurement is output from the sound source device 33 to be processed, No sound is output from the other sound source device 33.

  In step S <b> 16, each of the K acoustic sensors 34 collects the sound output from the sound source device 33 to be processed, and supplies the measurement signal obtained as a result to the transfer function calculator 35.

In step S <b> 17, the transfer function calculator 35 determines each of the K acoustic sensors 34 from the sound source device 33 to be processed based on the measurement signal supplied from the acoustic sensor 34 and the supplied measurement acoustic signal. A transfer function H _n (r) indicating the transfer characteristic until is calculated and supplied to the filter coefficient calculation unit 37. Thereby, a transfer function between the K acoustic sensors 34 and the sound source device 33 to be processed is obtained.

In step S <b> 18, the filter calculation unit 21 determines whether or not all sound source devices 33 have been processed. For example, when all the sound source devices 33 are to be processed and the transfer function H _n (r) for each sound source device 33 is obtained, it is determined that the processing has been performed for all the sound source devices 33.

If it is determined in step S18 that processing has not been performed for all the sound source devices 33, the processing returns to step S14, and the above-described processing is repeated. That is, the sound source device 33 to be processed next is selected, and the transfer function H _n (r) is calculated.

  On the other hand, if it is determined in step S18 that the processing has been performed for all sound source devices 33, the process proceeds to step S19.

In step S19, the filter coefficient calculation unit 37 includes the transfer function H _n (r) supplied from the transfer function calculator 35, the position r of each control point supplied from the information input unit 36, and the area where each control point exists. The input filter coefficient of the input filter is calculated based on the type of sound pressure, the sound pressure characteristic d (r) as the target sound pressure, and the allowable error σ, and supplied to the amplitude phase control filter processing unit 61.

Specifically, the filter coefficient calculation unit 37 solves the linear programming problem obtained by discretizing the position r and the rotation angle θ with respect to the linear semi-infinite programming problem shown in the above-described equation (13), thereby obtaining N N An input filter coefficient W _n for each sound source device 33 is calculated.

In other words, the filter coefficient calculator 37 minimizes the maximum value of the sound pressure in the non-listening area under the condition that the error of the sound pressure in the listening area with respect to the target sound pressure is within the allowable error σ. by solving the linear programming problem of finding W _n, to calculate the input filter coefficients W _n.

At this time, the filter coefficient calculation unit 37 may solve the linear planning problem by adding a constraint condition for the upper limit value g _max shown in the equation (14) to the linear planning problem, for example. In addition, when the non-limited area is set, the filter coefficient calculation unit 37 adds the constraint condition for the non-limited area shown in Expression (15) to the linear programming problem to solve the linear programming problem. May be.

When the input filter coefficient W _n is obtained, the filter coefficient calculating unit 37, the input filter coefficients W _n obtained for each sound source device 33, their source device corresponding amplitude phase that is connected to the tone generator 63 to 33 This is supplied to the control filter processing unit 61.

That is, the input filter coefficient W _n obtained for the _n -th sound source device 33-n is the amplitude / phase control filter processing unit 61- connected to the n-th sound device 63-n corresponding to the sound source device 33-n. n.

When the input filter coefficient W _n is supplied to the amplitude / phase control filter processing unit 61, the filter coefficient calculation process ends.

  As described above, the acoustic system 11 sets information regarding each control point such as the type of area and the target sound pressure, and obtains a transfer function by measurement. Then, the acoustic system 11 calculates the maximum value of the sound pressure in the non-listening area under the condition that the sound pressure error in the listening area is within the allowable error σ based on the setting and transfer function regarding each control point. An input filter coefficient is calculated by solving a linear programming problem of obtaining an input filter coefficient to be minimized.

  As a result, it is possible to obtain an input filter coefficient that can suppress the sound pressure in the non-listening area to the maximum while keeping the sound pressure in the listening area within the range determined by the allowable error. It is possible to obtain a sound with a small sound pressure distribution.

<Description of audio playback processing>
Next, a sound reproduction process in which the sound system 11 reproduces sound using the input filter obtained by the filter coefficient calculation process will be described.

  In step S41, the amplitude phase control filter processing unit 61 performs a filtering process using the input filter supplied from the filter coefficient calculation unit 37 on the supplied input acoustic signal, and the output acoustic signal obtained as a result thereof Is supplied to the D / A converter 62.

  Specifically, the amplitude phase control filter processing unit 61 performs time-frequency conversion on the supplied input acoustic signal, thereby converting the input acoustic signal, which is a time domain signal (time signal), into a frequency domain signal. Is converted to a frequency spectrum.

Then, the amplitude / phase control filter processing unit 61 includes, for each frequency component of the obtained frequency spectrum, elements (coefficients) of each frequency component constituting the input filter coefficient W _n of the input filter supplied from the filter coefficient calculation unit 37. Is used to adjust the relative amplitude and phase of the input acoustic signal. That is, the input acoustic signal is subjected to filter processing using an input filter in the frequency domain.

  Further, the amplitude phase control filter processing unit 61 performs frequency-time conversion, that is, inverse conversion of time-frequency conversion, on the filtered frequency spectrum, thereby outputting an output acoustic signal that is a filtered time-domain signal. Get.

  In step S <b> 42, the D / A converter 62 performs D / A conversion on the output acoustic signal supplied from the amplitude phase control filter processing unit 61 and supplies it to the sound source device 63. As a result, the output acoustic signal is converted from a digital signal to an analog signal.

  In step S43, the sound source device 63 reproduces sound based on the output acoustic signal supplied from the D / A converter 62, and the sound reproduction process ends.

  When the sound is reproduced by each sound source device 63, the users in the areas R1 and R3 in the reproduction space, that is, the non-listening area cannot listen to the reproduced sound, and the area R2 in the reproduction space, that is, A user in the listening area can listen to the reproduced sound.

  As described above, the acoustic system 11 performs the filtering process on the input acoustic signal using the input filter calculated in advance, and reproduces the sound based on the output acoustic signal obtained as a result. As described above, by performing the filtering process on the input acoustic signal using the input filter, it is possible to obtain sound with a sound pressure distribution with less spatial variation and to obtain higher separation performance.

<Second Embodiment>
<Example of acoustic system configuration>
By the way, as described above, a listening area can be set for each of a plurality of input sound signals. In such a case, the acoustic system 11 is configured as shown in FIG. 4, for example. In FIG. 4, parts corresponding to those in FIG. 1 are denoted by the same reference numerals, and the description thereof is omitted.

  The acoustic system 11 shown in FIG. 4 includes a filter calculation unit 21 and an audio reproduction unit 22. In addition, two input acoustic signals for reproducing different sounds are supplied to the audio reproducing unit 22, and the audio reproducing unit 22 reproduces these two input acoustic signals simultaneously.

  Here, in order to distinguish these input sound signals, one input sound signal is also referred to as an input sound signal SG1 of the input IP1, and the other input sound signal is also referred to as an input sound signal SG2 of the input IP2.

  Further, in this example, the reproduction space is divided into five areas R11 to R15. The area R12 is the listening area for the input IP1, the area R14 is the listening area for the input IP2, and the other areas R11, R13, and R15 are non-listening areas.

  Therefore, in this example, when calculating the input filter coefficient for the input acoustic signal SG1 of the input IP1, the area R12 is set as a listening area, and the other areas R11 and the areas R13 to R15 are set as non-listening areas.

  Further, when calculating the input filter coefficient for the input acoustic signal SG2 of the input IP2, the area R14 is set as a listening area, and the other areas R11 to R13 and the area R15 are set as non-listening areas.

  Here, an example in which each area is a listening area or a non-listening area will be described. For example, the input IP1 and the input IP2 are the non-listening areas of the areas R11, R13, and R15. Some may be non-limited areas.

  For example, when calculating the input filter coefficient of the input IP1, the area R12 and the area R13 may be set as the listening area, and when calculating the input filter coefficient of the input IP2, the area R13 and the area R14 may be set as the listening area. By doing so, only the voice of the input IP1 can be heard in the area R12, only the voice of the input IP2 can be heard in the area R14, and the voices of both the input IP1 and the input IP2 can be heard in the area R13.

  In the example of FIG. 4, the filter calculation unit 21 includes the switches 31-1 to 31-N, the D / A converters 32-1 to 32-N, the sound source device 33-1 to the sound source. The apparatus 33-N, the acoustic sensors 34-1 to 34-K, the transfer function calculator 35, the information input unit 36, the information input unit 101, the filter coefficient calculation unit 37, and the filter coefficient calculation unit 102 are included. .

  The configuration of the filter calculation unit 21 shown in FIG. 4 is different from the configuration of the filter calculation unit 21 shown in FIG. 1 in that an information input unit 101 and a filter coefficient calculation unit 102 are newly provided. The configuration is the same as that of one filter calculation unit 21.

In the filter calculation unit 21 of FIG. 4, the information input unit 36 performs various settings for the input acoustic signal SG1 of the input IP1, and supplies the setting results to the filter coefficient calculation unit 37. Further, the filter coefficient calculation unit 37 inputs the input filter for the input IP1 based on the transfer function H _n (r) supplied from the transfer function calculator 35 and various setting results supplied from the information input unit 36. Is calculated and supplied to the audio playback unit 22.

  On the other hand, the information input unit 101 performs various settings for the input acoustic signal SG2 of the input IP2 according to, for example, an operation input by the administrator, and supplies the setting results to the filter coefficient calculation unit 102. Further, the information input unit 101 includes an area setting unit 111, a target sound pressure setting unit 112, and an allowable error setting unit 113 corresponding to each of the area setting unit 41 to the allowable error setting unit 43.

Further, the filter coefficient calculation unit 102 inputs the input filter for the input IP2 based on the transfer function H _n (r) supplied from the transfer function calculator 35 and various setting results supplied from the information input unit 101. Is calculated and supplied to the audio playback unit 22.

  Also, the audio reproduction unit 22 shown in FIG. 4 includes an amplitude phase control filter processing unit 61-1 through an amplitude phase control filter processing unit 61-N, an amplitude phase control filter processing unit 131-1 through an amplitude phase control filter processing unit 131-. N, an adding unit 132-1 to an adding unit 132-N, a D / A converter 62-1 to a D / A converter 62-N, and a sound source device 63-1 to a sound source device 63-N.

  The configuration of the audio reproduction unit 22 shown in FIG. 4 is newly provided with an amplitude phase control filter processing unit 131-1 through amplitude phase control filter processing unit 131-N and an addition unit 132-1 through addition unit 132-N. 1 is different from the audio reproduction unit 22 in FIG. 1 in other respects and has the same configuration as the audio reproduction unit 22 in FIG.

  In the audio reproduction unit 22 of FIG. 4, the amplitude phase control filter processing unit 61-1 to the amplitude phase control filter processing unit 61-N apply the input acoustic signal SG1 of the input IP1 from the filter coefficient calculation unit 37. Filter processing using the supplied input filter is performed, and the output acoustic signal obtained as a result is supplied to the adder 132-1 to adder 132-N.

  On the other hand, the amplitude phase control filter processing unit 131-1 through the amplitude phase control filter processing unit 131-N receive the input supplied from the filter coefficient calculation unit 102 with respect to the input acoustic signal SG2 of the supplied input IP2. Filter processing using a filter is performed, and an output acoustic signal obtained as a result is supplied to the adder 132-1 to adder 132-N.

  The adding units 132-1 to 132-N respectively output acoustic signals supplied from the amplitude phase control filter processing unit 61-1 to the amplitude phase control filter processing unit 61-N and the amplitude phase control filter processing unit 131-. 1 to the output acoustic signal supplied from the amplitude / phase control filter processor 131-N are added and supplied to the D / A converter 62-1 to D / A converter 62-N.

  Hereinafter, the amplitude phase control filter processing unit 131-1 to the amplitude phase control filter processing unit 131 -N are also simply referred to as the amplitude phase control filter processing unit 131 when it is not necessary to distinguish between them. Hereinafter, the addition units 132-1 to 132-N are also simply referred to as the addition unit 132 when it is not necessary to distinguish between them.

<Description of audio playback processing>
Next, the operation of the acoustic system 11 shown in FIG. 4 will be described.

  First, the acoustic system 11 calculates and obtains an input filter for the input IP1 and an input filter for the input IP2 in advance before reproducing the audio of the input IP1 and the input IP2. At this time, the acoustic system 11 performs the filter coefficient calculation process described with reference to FIG. 2 for the input IP1 and the input IP2.

  Here, in the filter coefficient calculation process for obtaining the input filter for the input IP1, as in the case of the first embodiment, the process from step S11 to step S13 is performed by the information input unit 36, and the process of step S19 is performed. This is performed by the filter coefficient calculation unit 37. In this case, the area setting unit 41 of the information input unit 36 sets control points with the area R12 as a listening area and the areas R11 and R13 to R15 as non-listening areas.

  In contrast, in the filter coefficient calculation process for obtaining the input filter for the input IP2, the area setting unit 111, the target sound pressure setting unit 112, and the allowable error setting unit 113 of the information input unit 101 perform steps S11, S12, And the process of step S13 is performed. At this time, in step S11, the area setting unit 111 sets a control point with the area R14 as a listening area and the areas R11 to R13 and the area R15 as non-listening areas.

In addition, when the filter coefficient calculation process for the input IP1 is performed before the filter coefficient calculation process for the input IP2, the transfer function H _n (r) has already been obtained. In the filter coefficient calculation process of step S14 to step S18, the process may not be performed.

Furthermore, the filter coefficient calculation processing for the input IP2, the process of step S19 by the filter coefficient calculation unit 102 is performed, the input filter coefficients W _n for the input IP2 is calculated and supplied to the amplitude phase control filter processing unit 131 .

  When the input filter for the input IP1 and the input filter for the input IP2 are obtained in this way, the acoustic system 11 then performs a sound reproduction process for reproducing the sound of the input IP1 and the sound of the input IP2 at an arbitrary timing. Do.

  Hereinafter, the sound reproduction process performed by the acoustic system 11 shown in FIG. 4 will be described with reference to the flowchart of FIG.

  In step S71, the sound reproducing unit 22 performs a filtering process on the input acoustic signal for each input.

  That is, the amplitude phase control filter processing unit 61 performs the filtering process using the input filter supplied from the filter coefficient calculation unit 37 on the input acoustic signal SG1 of the supplied input IP1, and the output obtained as a result thereof The acoustic signal is supplied to the adding unit 132. The filter process performed by the amplitude / phase control filter processing unit 61 is the same as the filter process performed in step S41 of FIG.

  Similarly, the amplitude / phase control filter processing unit 131 performs a filter process using the input filter supplied from the filter coefficient calculation unit 102 on the input acoustic signal SG2 of the supplied input IP2, and obtained as a result. The output acoustic signal is supplied to the adding unit 132. Note that the filter processing performed by the amplitude phase control filter processing unit 131 is the same as the filter processing performed by the amplitude phase control filter processing unit 61, and thus the description thereof is omitted.

  In step S72, the adding unit 132 adds the output acoustic signal of the input IP1 supplied from the amplitude / phase control filter processing unit 61 and the output acoustic signal of the input IP2 supplied from the amplitude / phase control filter processing unit 131. To the D / A converter 62.

  When the output acoustic signal of the input IP1 and the output acoustic signal of the input IP2 are added, the processes of step S73 and step S74 are performed thereafter, and the sound reproduction process is terminated. These processes are performed in steps S42 and S42 of FIG. Since it is the same as the process of step S43, the description is abbreviate | omitted.

  When the sound based on the output acoustic signal is reproduced in this way, the sound of the input IP1 can be heard in the area R12 in the reproduction space, and the sound of the input IP2 can be heard in the area R14. . In addition, in area R11, area R13, and area R15, it is impossible to listen to audio.

  As described above, the acoustic system 11 performs the filtering process on the input acoustic signal using the input filter calculated in advance for each input, and adds the output acoustic signals of the respective inputs obtained as a result to generate the sound. Reproduce. As described above, by performing the filtering process on the input acoustic signal using the input filter for each input, it is possible to obtain sound with a sound pressure distribution with less spatial variation and to obtain higher separation performance. it can.

<Use Case 1>
Various cases are conceivable as use cases of the acoustic system 11 shown in FIG. Specifically, for example, as shown in FIG. 6, the acoustic system 11 can be applied when the audio of the content is reproduced by the speaker array 182 while the video of the content is reproduced by the image presentation device 181.

  In FIG. 6, the acoustic system 11 is not shown, but each speaker constituting the speaker array 182 corresponds to the sound source device 63, and an output acoustic signal from the D / A converter 62 is supplied to the speaker array 182.

  For example, there may be cases where a primary audio of language 1 and a secondary audio of language 2 are included, such as audio multiplex broadcasting in television broadcasting and movies. Normally, when those two sounds are played back at the same time, the main sound and sub sound will be mixed and heard, so it is difficult for people who understand language 1 and people who understand language 2 to listen to the content in the same room. Met.

  On the other hand, if the acoustic system 11 shown in FIG. 4 is used, it is possible to provide a listening area for only the main sound and a listening area for only the sub sound in the reproduction space. Therefore, the place where the person who understands language 1 is the main sound listening area and the sub sound non-listening area, and the place where the person who understands language 2 is the sub sound listening area and the main sound non-listening area. Thus, the user can enjoy the content without being disturbed by the voice of the other party's language.

  In this example, the reproduction space is divided into five areas R21 to R25. It is assumed that a user who understands language 1 is in area R22 and a user who understands language 2 is in area R24. In such a case, in the acoustic system 11, each of the input filters for the main sound is set as the area R22 as the main sound listening area and the sub sound non-listening area, and the area R24 as the sub sound listening area and the main sound non-listening area. And a sub audio input filter may be calculated. In this case, the area R21, the area R23, and the area R25 may be non-listening areas or non-limited areas.

<Use Case 2>
Further, as a use case of the acoustic system 11 shown in FIG. 4, for example, a case shown in FIG.

  For example, when a plurality of sounds corresponding to each video and image are reproduced simultaneously in the same room in a museum or an exhibition, the sound reproduced at a predetermined place obstructs the sound reproduced at another place. Can occur.

  In such a case, if the acoustic system 11 shown in FIG. 4 is used, the plurality of sounds can be reproduced without the mutual sounds being disturbed by other sounds.

  In FIG. 7, the video of the content C1 is reproduced by the image presentation device 211, the video of the content C2 is reproduced by the image presentation device 212, and the audio of the content C1 and the audio of the content C2 are reproduced by the speaker array 213. .

  Although the acoustic system 11 is not illustrated in FIG. 7, each speaker constituting the speaker array 213 corresponds to the sound source device 63, and an output acoustic signal from the D / A converter 62 is supplied to the speaker array 213. .

  In this example, the reproduction space is divided into five areas R31 to R35. Then, the content C1 and the content C1 and the content C1 are listened to as the area R32 as the audio listening area for the content C2 and the content C2 as the audio non-listening area, and the area R34 as the content C2 as the audio listening area. What is necessary is just to calculate the input filter of C2. In this case, the area R31, the area R33, and the area R35 may be non-listening areas or non-limited areas.

  By doing in this way, the user in the front of the image presentation device 211 where the video of the content C1 is reproduced, that is, the user in the area R32 can listen only to the sound of the content C1. Similarly, the user in the front of the image presentation device 212 where the video of the content C2 is reproduced, that is, the user in the area R34 can listen only to the audio of the content C2.

  By the way, the above-described series of processing can be executed by hardware or can be executed by software. When a series of processing is executed by software, a program constituting the software is installed in the computer. Here, the computer includes, for example, a general-purpose personal computer capable of executing various functions by installing a computer incorporated in dedicated hardware and various programs.

  FIG. 8 is a block diagram illustrating an example of a hardware configuration of a computer that executes the above-described series of processes using a program.

  In a computer, a CPU (Central Processing Unit) 501, a ROM (Read Only Memory) 502, and a RAM (Random Access Memory) 503 are connected to each other by a bus 504.

  An input / output interface 505 is further connected to the bus 504. An input unit 506, an output unit 507, a recording unit 508, a communication unit 509, and a drive 510 are connected to the input / output interface 505.

  The input unit 506 includes a keyboard, a mouse, a microphone, an image sensor, and the like. The output unit 507 includes a display, a speaker, and the like. The recording unit 508 includes a hard disk, a nonvolatile memory, and the like. The communication unit 509 includes a network interface or the like. The drive 510 drives a removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory.

  In the computer configured as described above, the CPU 501 loads the program recorded in the recording unit 508 to the RAM 503 via the input / output interface 505 and the bus 504 and executes the program, for example. Is performed.

  The program executed by the computer (CPU 501) can be provided by being recorded on a removable medium 511 as a package medium or the like, for example. The program can be provided via a wired or wireless transmission medium such as a local area network, the Internet, or digital satellite broadcasting.

  In the computer, the program can be installed in the recording unit 508 via the input / output interface 505 by attaching the removable medium 511 to the drive 510. Further, the program can be received by the communication unit 509 via a wired or wireless transmission medium and installed in the recording unit 508. In addition, the program can be installed in the ROM 502 or the recording unit 508 in advance.

  The program executed by the computer may be a program that is processed in time series in the order described in this specification, or in parallel or at a necessary timing such as when a call is made. It may be a program for processing.

  The embodiments of the present technology are not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present technology.

  For example, the present technology can take a configuration of cloud computing in which one function is shared by a plurality of devices via a network and is jointly processed.

  In addition, each step described in the above flowchart can be executed by being shared by a plurality of apparatuses in addition to being executed by one apparatus.

  Further, when a plurality of processes are included in one step, the plurality of processes included in the one step can be executed by being shared by a plurality of apparatuses in addition to being executed by one apparatus.

  Furthermore, this technique can also be set as the following structures.

[1]
An area setting unit for setting some control points in the listening area and setting some control points in the non-listening area;
A target sound pressure setting unit for setting a target sound pressure of each of the control points in the listening area;
An allowable error setting unit for setting an allowable error of the target sound pressure;
For a predetermined number of sound sources less than the total number of all control points, the target sound of the sound pressure of the control points in the listening area is based on the transfer function between the sound source and each of the control points. A filter coefficient calculation unit for calculating a filter coefficient of a filter corresponding to the sound source that minimizes the sound pressure of the control point in the non-listening area under a condition that an error with respect to pressure falls within the allowable error range. And a voice processing device.
[2]
The sound processing apparatus according to [1], wherein the filter coefficient calculation unit further calculates a filter coefficient by further adding a constraint condition that a gain of the filter does not exceed a predetermined upper limit value.
[3]
The area setting unit sets some control points in other areas different from the listening area and the non-listening area,
The filter coefficient calculation unit calculates the filter coefficient by further adding a constraint that the sound pressure at the control point in the other area does not exceed a predetermined maximum sound pressure. [1] or [2] The speech processing apparatus according to the description.
[4]
The sound processing device according to [3], wherein the maximum sound pressure is determined based on a minimum value of the target sound pressure at the control point in the listening area.
[5]
The speech processing apparatus according to any one of [1] to [4], wherein the filter coefficient calculation unit calculates the filter coefficient by solving a linear programming problem.
[6]
Set some control points in the listening area and set some control points in the non-listening area,
Set the target sound pressure of each control point in the listening area,
Set the target sound pressure tolerance,
For a predetermined number of sound sources less than the total number of all control points, the target sound of the sound pressure of the control points in the listening area is based on the transfer function between the sound source and each of the control points. A sound including a step of calculating a filter coefficient of a filter corresponding to the sound source that minimizes a sound pressure at the control point in the non-listening area under a condition that an error with respect to pressure falls within the allowable error range. Processing method.
[7]
Set some control points in the listening area and set some control points in the non-listening area,
Set the target sound pressure of each control point in the listening area,
Set the target sound pressure tolerance,
For a predetermined number of sound sources less than the total number of all control points, the target sound of the sound pressure of the control points in the listening area is based on the transfer function between the sound source and each of the control points. A process including a step of calculating a filter coefficient of a filter corresponding to the sound source that minimizes the sound pressure of the control point in the non-listening area under a condition that an error with respect to pressure falls within the allowable error range. A program that causes a computer to execute.

  DESCRIPTION OF SYMBOLS 11 Sound system, 21 Filter calculation part, 22 Sound reproduction part, 33-1 thru | or 33-N, 33 Sound source device, 34-1 thru | or 34-N, 34 Acoustic sensor, 35 Transfer function calculator, 36 Information input part, 37 Filter coefficient calculation unit, 41 area setting unit, 42 target sound pressure setting unit, 43 tolerance setting unit, 61-1 to 61-N, 61 amplitude phase control filter processing unit, 63-1 to 63-N, 63 sound source device , 101 Information input unit, 102 Filter coefficient calculation unit, 111 Area setting unit, 112 Target sound pressure setting unit, 113 Allowable error setting unit, 131-1 to 131 -N, 131 Amplitude phase control filter processing unit, 132-1 to 132-N, 132 adder

Claims (7)

An area setting unit for setting some control points in the listening area and setting some control points in the non-listening area;
A target sound pressure setting unit for setting a target sound pressure of each of the control points in the listening area;
An allowable error setting unit for setting an allowable error of the target sound pressure;
For a predetermined number of sound sources less than the total number of all control points, the target sound of the sound pressure of the control points in the listening area is based on the transfer function between the sound source and each of the control points. A filter coefficient calculation unit for calculating a filter coefficient of a filter corresponding to the sound source that minimizes the sound pressure of the control point in the non-listening area under a condition that an error with respect to pressure falls within the allowable error range. And a voice processing device. The audio processing apparatus according to claim 1, wherein the filter coefficient calculation unit calculates the filter coefficient by further adding a constraint that the gain of the filter does not exceed a predetermined upper limit value. The area setting unit sets some control points in other areas different from the listening area and the non-listening area,
The audio processing according to claim 1, wherein the filter coefficient calculation unit further calculates the filter coefficient by further adding a constraint that a sound pressure at the control point in the other area does not exceed a predetermined maximum sound pressure. apparatus. The sound processing apparatus according to claim 3, wherein the maximum sound pressure is determined based on a minimum value of the target sound pressure at the control point in the listening area. The speech processing apparatus according to claim 1, wherein the filter coefficient calculation unit calculates the filter coefficient by solving a linear planning problem. Set some control points in the listening area and set some control points in the non-listening area,
Set the target sound pressure of each control point in the listening area,
Set the target sound pressure tolerance,
For a predetermined number of sound sources less than the total number of all control points, the target sound of the sound pressure of the control points in the listening area is based on the transfer function between the sound source and each of the control points. A sound including a step of calculating a filter coefficient of a filter corresponding to the sound source that minimizes a sound pressure at the control point in the non-listening area under a condition that an error with respect to pressure falls within the allowable error range. Processing method. Set some control points in the listening area and set some control points in the non-listening area,
Set the target sound pressure of each control point in the listening area,
Set the target sound pressure tolerance,
For a predetermined number of sound sources less than the total number of all control points, the target sound of the sound pressure of the control points in the listening area is based on the transfer function between the sound source and each of the control points. A process including a step of calculating a filter coefficient of a filter corresponding to the sound source that minimizes the sound pressure of the control point in the non-listening area under a condition that an error with respect to pressure falls within the allowable error range. A program that causes a computer to execute.

Images