Generation apparatus, generation method, and program

Abstract

To provide a generation device, a generation method, and a program thereof, which achieve high suppression performance even near a user's ear away from an actual error microphone.SOLUTION: In a noise suppression system, two error microphones 93-1 and 93-2 are arranged in the vicinity of the head of a user, a virtual microphone 130 is located at a position closer to an observation point than the two error microphones, the virtual microphone is located on a straight line connecting the two error microphones, and the distance between a first error microphone 93-1 located at a position far from the virtual microphone and a second error microphone 93-2 located at a position close to the virtual microphone is equal to the distance between the second error microphone and the virtual microphone. The generation device calculates a transfer function from the first error microphone to the second error microphone from a sound collection signal of the first error microphone and a sound collection signal of the second error microphone, estimates a sound collection signal obtained by collecting sound by the virtual microphone from the calculated transfer function and the sound collection signal of the second error microphone, and obtains an estimated sound collection signal.SELECTED DRAWING: Figure 6

Description

[Technical Field]

[0001] This invention relates to an active noise control (ANC) technique for suppressing external noise at a specific location. [Background technology]

[0002] Non-patent document 1 is known as a conventional active noise suppression technology. Active noise suppression generally uses a reference microphone, an error microphone, and a cancellation speaker. Figure 1 shows an example of the configuration of a conventional noise suppression device. The reference microphone 91 picks up the noise emitted by the noise source. The cancellation speaker 92 reproduces the cancellation signal generated by the suppression signal generator 90 and emits a cancellation sound that cancels out the noise. Furthermore, the error microphone 93 picks up any remaining noise and feeds it back. The suppression signal generator 90 uses the sound picked up by the reference microphone 91 and the sound picked up by the error microphone 93 to actively control and generate a cancellation signal so that the remaining noise is minimized. Since the cancellation speaker 92 emits a cancellation sound so that the remaining noise is minimized at the installation location of the error microphone 93, the cancellation sound suppresses noise most efficiently at the installation location of the error microphone 93. For this reason, the error microphone 93 is installed close to the user's ear. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Kajikawa, "Recent Topics and Applications of Active Noise Control," Research Report on Music Information Science (MUS), vol. 2015-MUS-107, no. 3, pp. 1-6, May 2015. [Overview of the project] [Problems that the invention aims to solve]

[0004] However, in actual use, it may not be possible to place the error microphone 93 close to the user's ear. As the distance between the error microphone 93 and the user's ear increases, as mentioned above, the noise is most efficiently suppressed at the installation location of the error microphone 93. Closer to the user's ear, more noise remains, reducing the suppression performance and potentially preventing the user from fully benefiting from noise suppression. For example, simulations confirmed that when the distance from the noise source to the ear is 100 mm and the error microphone 93 is placed at the user's ear (0 mm), the suppression performance is -∞ dB, while when the error microphone 93 is placed at an intermediate point between the noise source and the ear, the suppression performance is -7.38 dB. Figure 2 illustrates the difference between the conventional suppression range (sweet spot) S1 and the desired sweet spot S2.

[0005] The present invention aims to provide a generation device, generation method, and program thereof that achieve high suppression performance even when the sound is picked up near the user's ear, far from the actual error microphone, by estimating the sound pickup signal obtained when sound is picked up near the user's ear from the sound pickup signal picked up at the actual placement of the error microphone, and using the estimated sound pickup signal instead of the sound pickup signal picked up at the actual placement of the error microphone in order to actively control the cancellation signal. [Means for solving the problem]

[0006] To solve the above problems, according to one aspect of the present invention, the generating device generates a cancellation signal used for active noise control. Two error microphones are placed near the user's head, and a virtual microphone is located closer to the observation point than the two error microphones. The virtual microphone is located on a straight line connecting the two error microphones, and the distance between the first error microphone, which is far from the virtual microphone, and the second error microphone, which is close to the virtual microphone, is equal to the distance between the second error microphone and the virtual microphone. The generating device includes a sound pressure estimation unit that calculates a transfer function from the first error microphone to the second error microphone from the sound pickup signals of the first error microphone and the sound pickup signals of the second error microphone, estimates the sound pickup signal picked up by the virtual microphone from the calculated transfer function and the sound pickup signal of the second error microphone, and obtains an estimated sound pickup signal. It also includes a suppression signal generation unit that generates a cancellation signal to suppress noise at the installation location of the virtual microphone using the sound pickup signal of the noise to be suppressed and the estimated sound pickup signal. The virtual microphone is a microphone that is virtually installed but is not actually installed. [Effects of the Invention]

[0007] According to the present invention, it is possible to achieve higher suppression performance than conventional methods when it is not possible to place the error microphone near the user's ear. [Brief explanation of the drawing]

[0008] [Figure 1] A diagram illustrating conventional active noise control. [Figure 2] A diagram illustrating the suppression range of conventional technologies. [Figure 3] Functional block diagram of the noise suppression system according to the first embodiment. [Figure 4] A diagram showing an example of the processing flow of the noise suppression system according to the first embodiment. [Figure 5] A diagram illustrating the method for estimating the estimated sound signal. [Figure 6] Diagram to explain the relative positions of the error microphones. [Figure 7]A plan view for explaining a simulation situation to measure the effect of the noise suppression system according to the second embodiment. [Figure 8] A diagram showing simulation results. [Figure 9] A diagram for explaining the positional relationship of error microphones. [Figure 10] A plan view for explaining a simulation situation to measure the effect of the noise suppression system according to a modification of the second embodiment. [Figure 11] A diagram showing simulation results. [Figure 12] A diagram for explaining the positional relationship between the actual error microphones and the virtual error microphones when calculating virtual sound collection signals. [Figure 13] A diagram showing a configuration example of a computer to which this method is applied.

Modes for Carrying Out the Invention

[0009] Hereinafter, embodiments of the present invention will be described. In the drawings used in the following description, components having the same function and steps performing the same process are denoted by the same reference numerals, and redundant explanations are omitted. In the following description, symbols " ^ " " - ", etc. should be described directly above the following character, but due to text notation limitations, they are described immediately before the character. In the formula, these symbols are described in their original positions. Also, the processing performed for each element unit of a vector or matrix is applied to all elements of the vector or matrix unless otherwise specified.

[0010] <Highlights of the First Embodiment> In this embodiment, the observed sound pressure near the ear is estimated from the sound pickup signal of an error microphone placed at a distance from the ear. For example, the sound pickup signal of a virtual error microphone placed near the ear is estimated from the sound pickup signal of the actual error microphone, and in ANC, the sound pickup signal of the virtual error microphone is used as the sound pickup signal of the conventional error microphone. With this configuration, the sweet spot is changed from the installation location of the error microphone to the location of the virtual error microphone, and a sound is produced that cancels out any remaining noise near the ear.

[0011] Various methods can be considered for estimating the sound signal picked up by a virtual error microphone. For example, the sound pressure can be estimated by considering the distance attenuation and phase delay from the actual error microphone's placement to the ear. Alternatively, for example, the sound pressure at the ear can be estimated using spherical harmonics from an actual error microphone placed on a sphere.

[0012] <First Embodiment> Figure 3 shows a functional block diagram of the noise suppression system according to the first embodiment, and Figure 4 shows its processing flow.

[0013] The noise suppression system includes a reference microphone 91, a cancellation speaker 92, an error microphone 93, a suppression signal generation unit 110, and a sound pressure estimation unit 120. The device consisting of the suppression signal generation unit 110 and the sound pressure estimation unit 120 is also called a suppression signal generation device.

[0014] The suppression signal generator takes the sound pickup signal x(r) from the reference microphone 91 and the sound pickup signal x(e) from the error microphone 93 as inputs, and generates a cancellation signal (hereinafter also referred to as "suppression signal") y such that the point where the amount of noise suppression is maximum is located closer to the user than the installation position of the error microphone 93, and outputs it to the cancellation speaker 92.

[0015] A suppression signal generator is a special device configured by loading a special program into a known or dedicated computer having, for example, a central processing unit (CPU) and main memory (RAM). The suppression signal generator executes each process under the control of, for example, the central processing unit. Data input to the suppression signal generator and data obtained from each process are stored in, for example, main memory, and the data stored in main memory is read to the central processing unit as needed and used for other processes. Each processing unit of the suppression signal generator may be composed of hardware such as integrated circuits, at least in part. Each storage unit of the suppression signal generator can be composed of, for example, main memory such as RAM (Random Access Memory), or middleware such as a relational database or key-value store. However, each storage unit does not necessarily have to be located inside the suppression signal generator; it may be composed of auxiliary storage devices made of semiconductor memory elements such as hard disks, optical disks, or flash memory, and may be located outside the suppression signal generator.

[0016] The following describes each part.

[0017] <Reference Microphone 91> Reference microphone 91 picks up the sound to be suppressed (S91) and outputs the picked-up signal x(r). The sound to be suppressed picked up by reference microphone 91 will be referred to as "noise" below.

[0018] <Cancel Speaker 92> The cancellation speaker 92 receives the cancellation signal y as input and reproduces the cancellation signal y (S92). When the reproduced sound from the cancellation speaker 92 and the noise to be suppressed are in perfect opposite phase, the reproduced sound and the noise to be suppressed overlap, that is, the sound waves superimpose, causing the waves to cancel each other out, and thus the noise is suppressed.

[0019] <Error Microphone 93> The error microphone 93 picks up sounds that were not suppressed by the playback sound played back from the cancellation speaker 92, including any remaining noise (S93), and outputs a pickup signal x(e). The error microphone 93 is positioned closer to the noise source than the observation point (e.g., the user's ear). For example, as shown in Figure 5, the error microphone 93 is positioned 0.05m closer to the noise source than the user's ear.

[0020] <Sound pressure estimation unit 120> The sound pressure estimation unit 120 takes the output signal (acquired signal) x(e) of the error microphone 93 as input and calculates and outputs an estimated acquired signal x(v), which is the signal that is estimated to be acquired when the microphone 130 is placed closer to the observation point than the error microphone 93. In other words, the sound pressure estimation unit 120 estimates the acquired signal obtained when the sound that was not suppressed by the playback sound reproduced from the cancellation speaker 92 is acquired at the placement of the microphone 130 (S120), and outputs the estimated acquired signal as the estimated acquired signal x(v). The estimation method for the estimated acquired signal x(v) is described below as an example. Here, the microphone 130 is not actually installed but is installed virtually, and will be referred to as the virtual microphone 130 below.

[0021] The sound pressure estimation unit 120 estimates the sound signal of a virtual error microphone using spherical harmonic function expansion coefficients from the sound signals of multiple error microphones placed at equal intervals near the head. Figure 5 is a diagram illustrating the positional relationship of the actual error microphones.

[0022] In this estimation method, the radius r e Error microphones are placed at equal intervals on the surface of a sphere with radius r, and the sound pressure on the sphere is estimated. For example, the distance from the center to the error microphone is r. e Assuming r = 0.15m, error microphones can be placed at equal intervals by (i) placing 6 error microphones at the center of each face of a regular hexahedron (see Figure 5(i)) and (ii) placing 12 error microphones at the center of each face of a regular dodecahedron (see Figure 5(ii)). For example, the distance from the center to the observation point (position of virtual microphone 130) is estimated as r = 0.08m.

[0023] By using the spherical harmonic function expansion, it is possible to estimate the observed sound pressure on an arbitrary spherical surface from the observed sound pressure on a certain spherical surface.

[0024] Radius r e Observed values p(θ1, φ1), p(θ2, φ2), …, p(θ L , φ L ) of the sound pressure are obtained from the L error microphones above. For example, the sound collection signal x(e) = [p(θ1, φ1), p(θ2, φ2), …, p(θ L , φ L )] of the L error microphones 93 is used.

[0025] The sound pressure estimation unit 120 obtains the sound field coefficient P m n (·) for the radius r e on the sound field according to the following formula. nm (r e )

Equation

Equation

Equation

[0026] The derivation of equation (5) will be explained below.

[0027] When the noise source is considered a point source and reflection from a rigid sphere of radius a is taken into account, the sound pressure at point (r,θ,φ) is:

number

number

number

number

number

number

number

[0028] (N+1) 2 <L Here, the spherical harmonics Y m n The number of speakers required corresponds to each mode (·). If L=6, then N=1; if L=12, then N=2.

[0029] Furthermore, N is subject to the following constraints under which spatial aliasing does not occur.

[0030] kr <N The distance between the head and the virtual error microphone is limited. For example, with a frequency of 300Hz, when N=1, the estimable range is limited to a distance of r=0.18m from the head.

[0031] <Suppression signal generation unit 110> The suppression signal generation unit 110 takes the sound-collected signal x(r) and the estimated sound-collected signal x(v) as inputs, generates a cancellation signal y to suppress noise at the installation location of the virtual microphone 130 (S110), and outputs it.

[0032] Conventional techniques can be used for generating the cancellation signal. For example, the method described in Non-Patent Document 1 can be used. In this embodiment, feedforward ANC is realized using a sound-collected signal x(r), an estimated sound-collected signal x(v), and a cancellation signal y. The sound-collected signal that would be obtained when the interference sound between the noise from the noise source and the reproduced sound of the cancellation signal y is detected by the virtual microphone 130 is estimated, and the noise from the noise source is detected by the reference microphone 91 and input to a noise control filter realized by an adaptive digital filter to generate the cancellation signal y, which is then reproduced by the cancellation speaker 92. It is assumed that the reproduced sound of the cancellation signal y propagates through a secondary path, which is a series of transmission systems from the cancellation speaker 92 to the virtual microphone 130. The coefficients of the noise control filter are then updated by an adaptive algorithm so that the input to the virtual microphone 130 is minimized. Conventional update methods can be used for updating the coefficients of the noise control filter, so a description is omitted. In feedforward ANC, a secondary path model that estimates the secondary path is used in the adaptive algorithm to compensate for the influence of the secondary path.

[0033] <Effects> With the above configuration, higher suppression performance than conventional systems can be achieved when it is not possible to place an error microphone near the user's ear. Simulation results of the noise suppression system according to the first embodiment showed that when the noise was a 300Hz plane wave, the suppression performance was -19.04dB for the right ear and -19.25dB for the left ear, and when the noise was a 100Hz plane wave, the suppression amount was -24.91dB for the right ear and -24.90dB for the left ear.

[0034] <Second Embodiment> This explanation will focus on the differences from the first embodiment. The noise suppression system of the first embodiment achieves higher suppression performance than conventional systems when it is not possible to place an error microphone near the user's ear by estimating the sound pressure at a desired observation point (e.g., near the user's ear) and then applying ANC processing. However, in order to improve the accuracy of sound pressure estimation at the observation point, it is necessary to place multiple error microphones at equal intervals on a spherical surface around the user's head. Therefore, when implementing this as a noise suppression system, it is also necessary to place multiple error microphones in front of the user's head, which may be disruptive to the user.

[0035] Furthermore, implementing the sound pressure estimation method in the first embodiment as a time-domain filter may result in instability. Therefore, in this embodiment, a microphone array of known shape, including two error microphones 93-1 and 93-2, is arranged to satisfy predetermined constraints, and a noise control filter b1 is calculated by an adaptive algorithm using the sound signals from the two error microphones 93-1 and 93-2. The sound pressure at a desired observation point is estimated using the calculated noise control filter b1 and the sound signals from the error microphones. Furthermore, the noise control filter used in the suppression signal generation unit 110 is configured using the estimated sound pressure, thereby improving the suppression performance.

[0036] The noise suppression system according to the second embodiment will be described below.

[0037] <Noise suppression system according to the second embodiment> Figure 3 shows a functional block diagram of the noise suppression system according to the second embodiment, and Figure 4 shows its processing flow.

[0038] The noise suppression system includes a reference microphone 91, a cancellation speaker 92, an error microphone 93, a suppression signal generation unit 110, and a sound pressure estimation unit 220. However, in this embodiment, the error microphone 93 consists of a microphone array of a known shape including two error microphones. The device consisting of the suppression signal generation unit 110 and the sound pressure estimation unit 220 is also called a suppression signal generation device.

[0039] The following describes a sound pressure estimation unit 220 that differs from that of the first embodiment.

[0040] <Sound pressure estimation unit 220> The sound pressure estimation unit 220 takes the output signal (acquired sound signal) x(e) of the error microphone 93 as input and calculates and outputs an estimated acquired sound signal x(v), which is the signal that is estimated to be acquired when the virtual microphone 130 is placed closer to the observation point than the error microphone 93. In other words, the sound pressure estimation unit 220 estimates the acquired sound signal obtained when sound that was not suppressed by the playback sound reproduced from the cancellation speaker 92 is acquired at the placement of the virtual microphone 130 (S220), and outputs the estimated acquired sound signal as the estimated acquired sound signal x(v). The estimation method for the estimated acquired sound signal x(v) is described below as an example.

[0041] The sound pressure estimation unit 220 calculates a transfer function from one error microphone 93-1, which is located far from the noise source, to the other error microphone 93-2, based on the sound signals from two error microphones 93-1 and 93-2 included in a microphone array placed near the head. It then estimates the sound signal of a virtual microphone from the calculated transfer function and the sound signal from the other error microphone 93-2. If the noise source does not change, the transfer function may be calculated in advance prior to the noise suppression process. If the noise source changes, the transfer function may be calculated and updated at predetermined time intervals, or a sequential transfer function may be calculated and updated.

[0042] Figure 6 is a diagram illustrating the positional relationship of the error microphones. The microphone array, including error microphones 93-1 and 93-2, is arranged to satisfy the following constraints.

[0043] (i) Arrange the microphone array such that the desired observation point lies on a straight line connecting the two error microphones 93-1 and 93-2.

[0044] (ii) The microphone array is arranged such that the distance between error microphone 93-1 and error microphone 93-2 is equal to the distance between error microphone 93-2 and the desired observation point (location of virtual microphone 130).

[0045] If conditions (i) and (ii) above are met, moving one of the error microphones 93-1 of the microphone array in parallel to the position where the other error microphone 93-2 was located will cause the position of the other error microphone 93-2 to coincide with the position of the desired observation point (the position of the virtual microphone 130).

[0046] In the microphone array, one of the two error microphones (error microphone 93-2 in Figure 6) is positioned closer to the noise source, while the other (error microphone 93-1 in Figure 6) is positioned further away. Therefore, the noise source is not equidistant from the two error microphones.

[0047] The two error microphones are positioned far enough apart to prevent spatial aliasing depending on the wavelength of the noise being suppressed.

[0048] Furthermore, the microphone array is positioned so that the noise source is sufficiently far from the microphone array and the desired observation point. "Sufficiently far" means that the noise arriving at the microphone array and the desired observation point can be considered as a plane wave.

[0049] Furthermore, a microphone array is placed between the noise source and the observation point. This arrangement ensures that the noise reaches the microphone array before it reaches the observation point.

[0050] By arranging the microphone array in this way, the transfer function from error microphone 93-1 to error microphone 93-2 can be used to replace the transfer function from error microphone 93-2 to the desired observation point (virtual microphone 130).

[0051] Let x1(t) be the sound signal picked up by error microphone 93-1, and x2(t) be the sound signal picked up by error microphone 93-2. Let x(e) = [x1(t), x2(t)] be the sound signal picked up by error microphone 93, which consists of a microphone array including error microphones 93-1 and 93-2. Here, t is an index representing time. If a1 is the transfer function from error microphone 93-1 to error microphone 93-2,

number

[0052] Here, using error microphone 93-1 as the reference microphone and error microphone 93-2 as the error microphone, the noise control filter b1 is calculated using a fitting algorithm.

[0053] In other words, the noise control filter b1(t) is updated so that the error between the acquired sound signal x2(t) and the estimated value y2(t) of the sound signal acquired by the error microphone 93-2, which is estimated by the following equation, is minimized.

number

[0054] Since the above constraints are satisfied, the transfer function from error microphone 93-1 to error microphone 93-2 can be used to replace the transfer function from error microphone 93-2 to the desired observation point (virtual microphone 130). Furthermore, in equation (22), the estimated sound-receiving signal x(v) can be calculated by replacing the sound-receiving signal x1(t) from error microphone 93-1 with the sound-receiving signal x2(t) from error microphone 93-2. In other words, the estimated sound-receiving signal x(v) = x(t) at time t can be calculated using the following equation.

number

[0055] As shown in equation (23), in this embodiment, the sound pressure estimation method can be implemented as a time-domain filter.

[0056] <Effects> With the above configuration, similar to the first embodiment, higher suppression performance than conventional systems can be achieved when it is not possible to place the error microphone near the user's ear. Furthermore, since it is a time-domain process, it can be easily incorporated into ANC algorithms. Figure 7 is a plan view illustrating the simulation situation for measuring the effect of the noise suppression system according to the second embodiment. The noise source, cancellation speaker 92, observation point, and error microphones 93-1 and 93-2 are all placed at a height of 1 m. Figure 8 shows the simulation results. Figure 8 shows the suppression results for band noise (100 Hz to 800 Hz) and demonstrates the suppression performance at the desired observation point. From Figure 8, it can be seen that the noise is suppressed. Furthermore, there is no need to place an error microphone in front of the user's head, so it does not interfere with the user.

[0057] Furthermore, the number of error microphones included in the microphone array only needs to be two or more, and this embodiment can be applied by using two of those error microphones.

[0058] <Example 1> This explanation will focus on the differences from the second embodiment. In this embodiment, a microphone array of known shape, including four error microphones 93-1, 93-2, 93-3, and 93-4, is arranged to satisfy predetermined constraints. Using the sound signals from three error microphones 93-1, 93-2, and 93-3, noise control filters b1 and b3 are calculated by an adaptive algorithm. The sound pressure at a desired observation point is estimated using the calculated noise control filters b1 and b3 and the sound signals from error microphones 93-2 and 93-4. Furthermore, the noise control filter used in the suppression signal generation unit 110 is configured using the estimated sound pressure, thereby improving the suppression performance.

[0059] The following describes the sound pressure estimation unit 220, which differs from that of the second embodiment. <Sound pressure estimation unit 220> The sound pressure estimation unit 220 takes the output signal (acquired sound signal) x(e) of the error microphone 93 as input and calculates and outputs an estimated acquired sound signal x(v), which is the signal that is estimated to be acquired when the virtual microphone 130 is placed closer to the observation point than the error microphone 93. In other words, the sound pressure estimation unit 220 estimates the acquired sound signal obtained when sound that was not suppressed by the playback sound reproduced from the cancellation speaker 92 is acquired at the placement of the virtual microphone 130 (S220), and outputs the estimated acquired sound signal as the estimated acquired sound signal x(v). The estimation method for the estimated acquired sound signal x(v) is described below as an example.

[0060] The four error microphones included in the microphone array positioned near the head shall be positioned at the vertices of a square.

[0061] The sound pressure estimation unit 220 calculates the transfer function from error microphone 93-3 to error microphone 93-2 and from error microphone 93-1 to error microphone 93-2 based on the sound signals of three error microphones included in the microphone array placed near the head (the sound signal of error microphone 93-2 closest to the desired observation point, the sound signal of error microphone 93-1 located on the straight line connecting the desired observation point and error microphone 93-2, and the sound signal of error microphone 93-3 located diagonally opposite error microphone 93-2 in a square). The unit then estimates the sound signal of a virtual microphone from the calculated transfer functions and the sound signals of error microphones 93-2 and 93-4.

[0062] Figure 9 is a diagram illustrating the positional relationship of the error microphones. The microphone array, including error microphones 93-1, 93-2, 93-3, and 93-4, is arranged to satisfy the following constraints.

[0063] (i) Configure the microphone array so that the four error microphones 93-1, 93-2, 93-3, and 93-4 are located at the vertices of a square.

[0064] (ii) The microphone array is arranged such that error microphone 93-1 is located on a straight line connecting the desired observation point and the error microphone 93-2 closest to the desired observation point.

[0065] (iii) Arrange the microphone array such that the distance between error microphone 93-1 and error microphone 93-2 is equal to the distance between error microphone 93-2 and the desired observation point.

[0066] If conditions (i), (ii), and (iii) above are met, then when error microphones 93-3 and 93-1 are moved in parallel along with the microphone array to the position where error microphones 93-4 and 93-2 were located, the position of error microphone 93-2 coincides with the position of the desired observation point (the position of virtual microphone 130).

[0067] Of the four error microphones included in the microphone array, error microphone 93-3 is positioned closest to the noise source, while error microphone 93-2, located diagonally opposite error microphone 93-3, is positioned furthest from the noise source.

[0068] The four error microphones should be spaced apart to prevent spatial aliasing depending on the wavelength of the noise being suppressed.

[0069] Furthermore, the microphone array is positioned so that the noise source is sufficiently far from the microphone array and the desired observation point. "Sufficiently far" means that the noise arriving at the microphone array and the desired observation point can be considered as a plane wave.

[0070] Furthermore, a microphone array is placed between the noise source and the observation point. This arrangement ensures that the noise reaches the microphone array before the observation point. In this modified example, the microphone array is positioned such that the error microphone 93-2 is located away from the desired observation point in the vertical direction of the head surface.

[0071] By arranging the microphone array in this way, the transfer function from error microphone 93-3 and error microphone 93-1 to error microphone 93-2 can be used to replace the transfer function from error microphone 93-4 and error microphone 93-2 to the desired observation point (virtual microphone 130).

[0072] In the second embodiment, when there are two noise sources, the solution is indeterminate and the transfer function cannot be estimated. However, in this modified example, the transfer function can be estimated even when there are two noise sources, and the sound signal picked up by the virtual microphone can be estimated.

[0073] Let x1(t), x2(t), x3(t), and x4(t) be the sound signals picked up by error microphones 93-1, 93-2, 93-3, and 93-4, respectively. Let x(e) = [x1(t), x2(t), x3(t), x4(t)] be the sound signal picked up by error microphone 93, which consists of a microphone array including error microphones 93-1, 93-2, 93-3, and 93-4. Let a3 be the transfer function from error microphone 93-3 to error microphone 93-2, and a1 be the transfer function from error microphone 93-1 to error microphone 93-2.

number

[0074] Here, using error microphones 93-3 and 93-1 as reference microphones and error microphone 93-2 as the error microphone, noise control filters b3 and b1 are calculated using a fitting algorithm.

[0075] In other words, the noise control filters b3(t) and b1(t) are updated so as to minimize the error between the acquired sound signal x2(t) and the estimated value y2(t) of the sound signal acquired by the error microphone 93-2, which is estimated by the following equation.

number

[0076] Since the above constraints are satisfied, the transfer function from error microphone 93-3 to error microphone 93-2 and the transfer function from error microphone 93-1 to error microphone 93-2 can be used to replace the transfer function from error microphone 93-4 to the desired observation point (virtual microphone 130) and the transfer function from error microphone 93-2 to the desired observation point (virtual microphone 130), respectively. Furthermore, in equation (25), by replacing the sound pickup signal x3(t) from error microphone 93-3 with the sound pickup signal x4(t) from error microphone 93-4, and the sound pickup signal x1(t) from error microphone 93-1 with the sound pickup signal x2(t) from error microphone 93-2, the estimated sound pickup signal x(v) can be calculated. In other words, the estimated sound pickup signal x(v) = x(t) at time t can be calculated by the following equation.

number

[0077] As shown in equation (26), in this embodiment, the sound pressure estimation method can be implemented as a time-domain filter.

[0078] <Effects> With this configuration, the same effects as in the second embodiment can be obtained. Furthermore, even when there are two noise sources, the transfer function can be estimated, and the sound picked up by the virtual microphone can be estimated with high accuracy. Figure 10 is a plan view illustrating the simulation situation for measuring the effect of the noise suppression system according to a modified version of the second embodiment. The noise source, cancellation speaker 92, observation point, and error microphones 93-1, 93-2, 93-3, and 93-4 are all positioned at a height of 1 m. Figure 11 shows the simulation results. Figure 11 shows the suppression results for band noise (100 Hz to 800 Hz) and shows the suppression performance at the desired observation point. From Figure 11, it can be seen that the noise is suppressed.

[0079] Furthermore, the microphone array only needs to contain four or more error microphones, and this modified example can be applied by using four of those error microphones.

[0080] <Modification 2> This explanation will focus on the differences from the second embodiment. In the second embodiment, since the microphone array is placed between the noise source and the observation point, the noise reaches the error microphone first, but if the noise reaches the observation point first, in the second embodiment, it is not possible to estimate the sound pickup signal of the virtual microphone 130.

[0081] In this embodiment, a noise suppression system is realized that can estimate the sound pickup signal of the virtual microphone 130 even when the noise reaches the observation point first (when the observation point is between the noise source and the microphone array) by processing the transfer function from the noise source to the microphone array. In this embodiment, it is assumed that the noise source signal s (dry source) is known.

[0082] The following describes the sound pressure estimation unit 220, which differs from that of the second embodiment. <Sound pressure estimation unit 220> The sound pressure estimation unit 220 takes the output signal (acquired sound signal) x(e) of the error microphone 93 as input and calculates and outputs an estimated acquired sound signal x(v), which is the signal that is estimated to be acquired when the virtual microphone 130 is placed closer to the observation point than the error microphone 93. In other words, the sound pressure estimation unit 220 estimates the acquired sound signal obtained when sound that was not suppressed by the playback sound reproduced from the cancellation speaker 92 is acquired at the placement of the virtual microphone 130 (S220), and outputs the estimated acquired sound signal as the estimated acquired sound signal x(v). The estimation method for the estimated acquired sound signal x(v) is described below as an example.

[0083] The sound pressure estimation unit 220 calculates the transfer function h1 from the noise source to error microphone 93-1 and the transfer function h2 from the noise source to error microphone 93-2, based on the sound signals x1(t) and x2(t) from two error microphones 93-1 and 93-2 included in a microphone array placed near the head, and the noise source signal s(t). For example, the transfer functions h1 and h2 are calculated from the following equations.

number

number

[0084] The sound pressure estimation unit 220 uses the transfer functions h'1 and h'2 obtained by removing the initial delays of the calculated transfer functions h1 and h2 to calculate the pseudo-sound pickup signals x'1(t) and x'2(t) obtained when the error microphones 93-1 and 93-2 are brought close to the noise source, according to the following equation.

number

number

[0085] Figure 12 is a diagram illustrating the positional relationship between the actual error microphones and the simulated error microphones used to calculate the simulated sound pickup signal. The amount of initial delay removal, in other words, the distance at which error microphones 93-1 and 93-2 are brought closer to the noise source, satisfies the following conditions.

[0086] (i) The desired observation point is located on a straight line connecting the pseudo-error microphones 93-1f and 93-2f.

[0087] (ii) The distance between the two pseudo-error microphones 93-1f and 93-2f is equal to the distance between the pseudo-error microphone 93-2f and the desired observation point.

[0088] Furthermore, since the simulated error microphone is not actually placed, it will not interfere with the user even if it is positioned in front of the user's head as shown in Figure 12.

[0089] By setting these conditions, the sound pickup signal of the virtual microphone can be estimated from the pseudo-sound pickup signals x'1(t) and x'2(t) in the same manner as in the second embodiment.

[0090] If conditions (i) and (ii) above are met, moving one of the error microphones 93-1f of the microphone array in parallel with the other error microphone 93-2f to the position where the other error microphone 93-2f of the two pseudo-error microphones was located will result in the position of the other error microphone 93-2f coinciding with the position of the desired observation point (the position of the virtual microphone 130).

[0091] Of the two simulated error microphones, one (error microphone 93-2f in Figure 12) is positioned closer to the noise source, while the other (error microphone 93-1f in Figure 12) is positioned further away. In other words, the microphone array is arranged so that the noise source is not equidistant from the two simulated error microphones.

[0092] The two simulated error microphones are assumed to be spaced apart to prevent spatial aliasing depending on the wavelength of the noise being suppressed.

[0093] Furthermore, the microphone array is positioned so that the noise source is sufficiently far from the two simulated error microphones and the desired observation point. "Sufficiently far" means that the noise arriving at the two simulated error microphones and the desired observation point is far enough away that it can be considered a plane wave.

[0094] Furthermore, two dummy error microphones are placed between the noise source and the observation point. This arrangement ensures that the noise reaches the two dummy error microphones before reaching the observation point.

[0095] By arranging these two pseudo-error microphones, the transfer function from error microphone 93-1f to error microphone 93-2f can be used to replace the transfer function from error microphone 93-2f to the desired observation point (virtual microphone 130).

[0096] Let x'1(t) be the signal picked up by the pseudo-error microphone 93-1f, and x'2(t) be the signal picked up by the pseudo-error microphone 93-2f. Here, t is an index representing time. If a1 is the transfer function from error microphone 93-1f to error microphone 93-2f,

number

[0097] Here, using error microphone 93-1f as the reference microphone and error microphone 93-2f as the error microphone, the noise control filter b1 is calculated using a fitting algorithm.

[0098] In other words, the noise control filter b1(t) is updated so as to minimize the error between the acquired sound signal x'2(t) and the estimated value y'2(t) of the acquired sound signal from the pseudo-error microphone 93-2f, which is estimated by the following equation.

number

[0099] Since the above constraints are satisfied, the transfer function from the pseudo-error microphone 93-1f to the pseudo-error microphone 93-2f can be used to replace the transfer function from the pseudo-error microphone 93-2f to the desired observation point (virtual microphone 130). Furthermore, in equation (32), the estimated sound-receiving signal x(v) can be calculated by replacing the sound-receiving signal x'1(t) from the pseudo-error microphone 93-1f with the sound-receiving signal x'2(t) from the pseudo-error microphone 93-2f. In other words, the estimated sound-receiving signal x(v)=x(t) at time t can be calculated using the following equation.

number

[0100] <Effects> This configuration allows for the same effects as the second embodiment. Furthermore, if the microphone array is not placed between the noise source and the observation point, and the noise reaches the observation point first, the sound signal picked up by the virtual microphone 130 can be estimated. The configuration may be switched between the second embodiment and this modified example depending on the positional relationship between the noise source, the microphone array, and the desired observation point. Furthermore, this modified example and modified example 1 may be combined.

[0101] <Other variations> The present invention is not limited to the embodiments and modifications described above. For example, the various processes described above may not only be performed sequentially as described, but may also be performed in parallel or individually as needed, depending on the processing capacity of the device performing the processes. Other modifications can be made as appropriate without departing from the spirit of the present invention.

[0102] <Program and recording medium> The various processes described above can be carried out by loading a program that executes each step of the above method into the computer's memory unit 2020 shown in Figure 13, and then causing the control unit 2010, input unit 2030, output unit 2040, etc. to operate.

[0103] The program describing this process can be recorded on a computer-readable recording medium. Any computer-readable recording medium can be used, such as a magnetic recording device, optical disc, magneto-optical recording medium, or semiconductor memory.

[0104] Furthermore, this program may be distributed, for example, by selling, transferring, or lending portable recording media such as DVDs or CD-ROMs on which the program is recorded. Alternatively, the program may be stored in the storage device of a server computer and distributed by transferring the program from the server computer to other computers via a network.

[0105] A computer executing such a program may, for example, first store the program recorded on a portable storage medium or a program transferred from a server computer in its own storage device. Then, when processing is to be executed, the computer reads the program stored on its own storage medium and executes the processing according to the read program. Alternatively, the computer may directly read the program from the portable storage medium and execute the processing according to that program, or it may sequentially execute the processing according to the received program each time a program is transferred to it from a server computer. Furthermore, the above processing may be executed by a so-called ASP (Application Service Provider) type service, where the server computer does not transfer programs to this computer, but the processing function is realized only by execution instructions and result acquisition. In this form, the program includes information used for processing by an electronic computer that is equivalent to a program (data that is not a direct instruction to the computer but has the property of defining the processing of the computer, etc.).

[0106] Furthermore, in this configuration, the device is configured by executing a predetermined program on a computer, but at least a part of these processes may be implemented in hardware.

Claims

1. A generator for generating cancellation signals used in active noise control, Two error microphones are positioned near the user's head, and a virtual microphone is positioned closer to the observation point than the two error microphones. The virtual microphone is positioned on a straight line connecting the two error microphones, and the distance between the first error microphone, which is far from the virtual microphone, and the second error microphone, which is close to the virtual microphone, is equal to the distance between the second error microphone and the virtual microphone. A sound pressure estimation unit calculates a transfer function from the first error microphone to the second error microphone from the sound-collected signal of the first error microphone and the sound-collected signal of the second error microphone, estimates the sound-collected signal to be picked up by the virtual microphone from the calculated transfer function and the sound-collected signal of the second error microphone, and obtains an estimated sound-collected signal. The system includes a suppression signal generation unit that generates a cancellation signal to suppress noise at the installation location of the virtual microphone using the sound pickup signal obtained from the noise to be suppressed and the estimated sound pickup signal, The aforementioned virtual microphone is a microphone that is virtually installed but is not actually installed. generator.

2. A generating apparatus according to claim 1, Four error microphones, including the two error microphones, are positioned near the user's head, and the four error microphones are located at the vertices of a square, with the second error microphone being the closest to the virtual microphone, and the third error microphone being positioned diagonally to the second error microphone. The sound pressure estimation unit calculates a transfer function from the third error microphone to the second error microphone and a transfer function from the first error microphone to the second error microphone from the sound-collected signals of the first error microphone, the sound-collected signals of the second error microphone, and the sound-collected signals of the third error microphone, and estimates the sound-collected signal to be picked up by the virtual microphone from the two calculated transfer functions, the sound-collected signals of the second error microphone, and the sound-collected signals of the fourth error microphone, thereby obtaining an estimated sound-collected signal. generator.

3. A generating apparatus according to claim 1, Assume that the noise source signal is known, and that the virtual microphone is located between the first and second error microphones and the noise source. The sound pressure estimation unit calculates a transfer function from the noise source to the first error microphone from the sound pickup signal of the first error microphone and the sound source signal, calculates a transfer function from the noise source to the second error microphone from the sound pickup signal of the second error microphone and the sound source signal, and uses the transfer function obtained by removing the initial delay of the two calculated transfer functions to calculate a pseudo sound pickup signal obtained when the first error microphone and the second error microphone are brought close to the noise source. From the pseudo-sound pickup signal of the first error microphone and the pseudo-sound pickup signal of the second error microphone, a transfer function h from the pseudo-first error microphone to the pseudo-second error microphone is calculated, and from the calculated transfer function h and the pseudo-sound pickup signal of the second error microphone, the sound pickup signal picked up by the virtual microphone is estimated to obtain the estimated sound pickup signal. The virtual microphone is positioned closer to the observation point than the two pseudo-error microphones, the virtual microphone is located on a straight line connecting the two pseudo-error microphones, and the distance between the pseudo-first error microphone, which is far from the virtual microphone, and the pseudo-second error microphone, which is close to the virtual microphone, is equal to the distance between the pseudo-second error microphone and the virtual microphone. generator.

4. A method for generating a cancellation signal used for active noise control, Two error microphones are positioned near the user's head, and a virtual microphone is positioned closer to the observation point than the two error microphones. The virtual microphone is positioned on a straight line connecting the two error microphones, and the distance between the first error microphone, which is far from the virtual microphone, and the second error microphone, which is close to the virtual microphone, is equal to the distance between the second error microphone and the virtual microphone. A sound pressure estimation step is performed to calculate a transfer function from the first error microphone to the second error microphone from the sound-collected signal of the first error microphone and the sound-collected signal of the second error microphone, estimate the sound-collected signal to be picked up by the virtual microphone from the calculated transfer function and the sound-collected signal of the second error microphone, and obtain an estimated sound-collected signal. The process includes a suppression signal generation step, which generates a cancellation signal for suppressing noise at the installation location of the virtual microphone, using the sound pickup signal obtained from the noise to be suppressed and the estimated sound pickup signal. The aforementioned virtual microphone is a microphone that is virtually installed but is not actually installed. Generation method.

5. A program for causing a computer to function as a generation device according to any of claims 1 to 3.

Images