Acoustic reproduction system

Abstract

This acoustic reproduction system comprises a controller for causing a plurality of speakers to reproduce sound of two channels. The controller causes the plurality of speakers to reproduce the sound by applying, to a left recording signal and a right recording signal, an approximate inverse matrix obtained by approximating an inverse matrix of a plant matrix indicating a transmission system from the plurality of speakers to both ears. The approximate inverse matrix is expressed by a product of a gain correction term for correcting the signal strength of a direct component according to a frequency band, a coefficient correction term for correcting a difference in strength between the direct component and a cross component according to the frequency band, a negative phase term for reversing the phase of the cross component, and a delay correction term for providing a time difference according to the frequency band of the cross component. The controller has a plurality of circuits for implementing the gain correction term, the coefficient correction term, the negative phase term, and the delay correction term.

Description

Sound reproduction system

[0001] The present disclosure relates to sound reproduction systems.

[0002] Patent Literature 1 discloses an audio reproduction system that reproduces the direction, distance, and spread of sound in a three-dimensional manner. The system reproduces two-channel sound, left and right, from multiple speakers based on a left recording signal corresponding to the listener's left ear and a right recording signal corresponding to the listener's right ear. The system has multiple left speakers corresponding to the left channel and multiple right speakers corresponding to the right channel. The system configures a filter for each speaker to cancel crosstalk. Crosstalk is a component of sound that reaches an ear other than the ear being controlled.

[0003] A filter that can cancel crosstalk is ideally determined by measuring the entire transfer system, represented by a plant matrix from the loudspeaker to both ears, and calculating the inverse matrix of the measured plant matrix. Hereinafter, the inverse matrix of the plant matrix will be referred to as the inverse filter matrix, and the system that controls acoustics by inversely transforming the transfer system will be referred to as the inverse system. Implementing an inverse filter matrix presents many problems, including loss of dynamic range, shifts in the listening position, lack of robustness against reflections in the control space, and individual differences in head-related transfer functions.

[0004] The system described in Patent Document 1 solves the above-mentioned problem by using a unique speaker whose position changes linearly depending on the frequency of the sound output. However, it is difficult to linearly change the speaker position depending on the frequency as defined. For this reason, in practice, the speaker positions are discretized, and each speaker is assigned one of the frequency bands divided by a predetermined bandwidth as its frequency band. Such a discretized system determines the frequency band of the sound output for each left speaker and each right speaker, and distributes the signal of the appropriate frequency band to the appropriate speaker, thereby achieving independent control of both ears.

[0005] Special Publication No. 2004-511118

[0006] The system described in Patent Document 1 requires the implementation of a circuit that performs a convolution operation of a large number of filter coefficients in order to precisely configure and control the inverse filter matrix. Therefore, the system described in Patent Document 1 may increase implementation costs. Implementation costs include labor costs required for development, the computational load of the operation itself, and the unit cost of parts. The present disclosure provides technology that can reduce implementation costs related to independent control of both ears.

[0007] An audio reproduction system according to one aspect of the present disclosure reproduces sound in a listening space in which a listener is present. The audio reproduction system includes a plurality of speakers and a controller. The plurality of speakers are arranged in the listening space. The controller reproduces two-channel sound, i.e., a left channel and a right channel, from the plurality of speakers based on a left recording signal corresponding to the listener's left ear and a right recording signal corresponding to the listener's right ear. The plurality of speakers include a plurality of left speakers corresponding to the left channel and a plurality of right speakers corresponding to the right channel. Discretized frequency bandwidths are set in advance for each of the plurality of left speakers and the plurality of right speakers so that the frequency bandwidths of the sounds that can be output differ. The controller reproduces sound from the plurality of speakers by applying an approximate inverse matrix, which is an approximation of an inverse matrix of a plant matrix that indicates a transfer system from the plurality of speakers to both ears, to the left recording signal and the right recording signal. The sounds output from the multiple speakers include direct components, which are sounds output from the multiple left speakers to the listener's left ear and sounds output from the multiple right speakers to the listener's right ear, and cross components, which are sounds output from the multiple left speakers to the listener's right ear and sounds output from the multiple right speakers to the listener's left ear. The approximate inverse matrix is ​​expressed as the product of a gain correction term that corrects the signal strength of the direct component according to a frequency band, a coefficient correction term that corrects the difference in signal strength between the direct component and the cross component according to a frequency band, an anti-phase term that inverts the phase of the cross component, and a delay correction term that provides a time difference according to the frequency band of the cross component. The controller has multiple circuits that implement the gain correction term, the coefficient correction term, the anti-phase term, and the delay correction term.

[0008] According to this sound reproduction system, an inverse filter matrix, i.e., an inverse matrix of a plant matrix representing a transfer system from multiple speakers to both ears, is approximated by an approximate inverse matrix. The approximate inverse matrix is ​​applied to left and right recording signals, thereby causing the multiple speakers to reproduce sound. The approximate inverse matrix is ​​expressed as a product of a gain correction term, a coefficient correction term, a phase-inverse term, and a delay correction term. In other words, the inverse filter matrix, which performs a convolution operation of a lengthy filter coefficient, is approximately divided into four elements that perform simpler operations, such as gain correction, coefficient correction, phase-inverse, and delay correction. The functions of the four elements are then shared and implemented by multiple circuits. Therefore, this sound reproduction system can reduce the implementation costs associated with independent control of both ears compared to implementing the inverse filter matrix directly.

[0009] In one embodiment, the controller includes an amplifier circuit that implements a gain correction term and a coefficient correction term, and a delay circuit that implements an anti-phase term and a delay correction term. The delay circuit may uniformly set, for each discretized frequency band, the sum of the time difference at which the phase difference between the direct component and the cross component becomes 180 degrees at the center frequency of the frequency bandwidth and the arrival time difference between the direct component and the cross component. This allows the sound reproduction system to realize an inverse filter matrix that performs a convolution operation of a long filter coefficient using simpler circuits with low implementation costs, such as an amplifier circuit and a delay circuit. Furthermore, the delay circuit sets a time difference (simple delay) between the sound from the left speaker and the sound from the right speaker so that crosstalk is canceled at a representative frequency (center frequency) within the frequency band assigned to the speaker. In other words, rather than calculating an inverse filter matrix for each frequency included in the frequency bandwidth assigned to a certain speaker, a time difference at which crosstalk is canceled at the center frequency of the frequency bandwidth of that speaker is calculated, and this time difference is applied to all frequencies included in the frequency band assigned to the speaker. As a result, this sound reproduction system calculates time differences and the like for each speaker (each frequency band), and therefore, compared to calculating an inverse filter matrix for each frequency, the accuracy of crosstalk cancellation for sounds at frequencies that deviate from the representative frequency is somewhat reduced, but the calculation cost can be significantly reduced. The calculation cost refers to the amount of calculation effort, the number of tasks, the complexity, etc.

[0010] In one embodiment, the controller may include an amplifier circuit implementing a gain correction term and a coefficient correction term, an inverter circuit implementing an antiphase term, and a delay circuit implementing a delay correction term, wherein the inverter circuit inverts the polarity of the signal, and the delay circuit uniformly sets the arrival time difference between the direct component and the cross component for each discrete frequency band. This allows the sound reproduction system to realize an inverse filter matrix that performs a convolution operation on a large number of filter coefficients using a simpler, less expensive circuit, namely, an amplifier circuit, an inverter circuit, and a delay circuit. Furthermore, as described above, the sound reproduction system calculates time differences and the like for each speaker (each frequency band). Therefore, compared to calculating an inverse filter matrix for each frequency, the flatness of the frequency characteristics of the sound generated at the ear of the controlled subject may be somewhat reduced for sounds at frequencies other than the representative frequency, but the calculation cost can be significantly reduced.

[0011] In one embodiment, the controller includes a feedback circuit that implements a gain correction term, an amplifier circuit that implements a coefficient correction term, and a delay circuit that implements an anti-phase term and a delay correction term. The feedback circuit calculates a gain correction term according to frequency, and the delay circuit may uniformly set, for each discretized frequency band, the sum of the time difference at which the phase difference between the direct component and the cross component becomes 180 degrees at the center frequency of the frequency bandwidth and the arrival time difference between the direct component and the cross component. This sound reproduction system can realize an inverse filter matrix that performs a convolution operation of a large number of filter coefficients using a simpler circuit with low implementation cost, namely, a feedback circuit, an amplifier circuit, and a delay circuit. Furthermore, because the feedback circuit sets a gain correction term having frequency characteristics, crosstalk can be canceled more accurately than when the gain correction term is a constant. Furthermore, as described above, this sound reproduction system calculates the time difference and the like for each speaker (each frequency band). Therefore, compared to calculating an inverse filter matrix for each frequency, the accuracy of crosstalk cancellation for sounds at frequencies shifted from the representative frequency is somewhat reduced, but the calculation cost can be significantly reduced.

[0012] In one embodiment, the controller includes a feedback circuit implementing a gain correction term, an amplifier circuit implementing a coefficient correction term, an inverting circuit implementing an antiphase term, and a delay circuit implementing a delay correction term. The feedback circuit calculates the gain correction term according to frequency, the inverting circuit inverts the polarity of the signal, and the delay circuit uniformly sets the arrival time difference between the direct component and the cross component for each discrete frequency band. This sound reproduction system can realize an inverse filter matrix that performs a convolution operation of a large number of filter coefficients using a simpler, less expensive circuit, including a feedback circuit, an amplifier circuit, an inverting circuit, and a delay circuit. Furthermore, as described above, the feedback circuit can more accurately cancel crosstalk. Furthermore, as described above, this sound reproduction system calculates time differences and the like for each speaker (each frequency band). Therefore, compared to calculating an inverse filter matrix for each frequency, the flatness of the frequency characteristics of the sound generated at the ear of the controlled subject may be somewhat reduced for sounds at frequencies other than the representative frequency, but the computational cost can be significantly reduced.

[0013] In one embodiment, the arrival time difference may be the time difference at which the correlation between the impulse response of the direct component and the impulse response of the cross component is maximized, in which case the sound reproduction system can set an appropriate arrival time difference based on the impulse responses.

[0014] In one embodiment, the gain correction term may be a numerical value in the range of 0.1 to 10. If the gain correction term is a constant with no frequency characteristics, the sound reproduction system may implement the gain correction term in an amplifier circuit.

[0015] In one embodiment, the gain correction term is γ / (1+g ｉ ２ ) where γ is a number ranging from 0.1 to 10, and g ｉmay be a transmission gain ratio, which is a ratio of signal strengths resulting from a path difference between the direct component and the cross component of the speaker in the i-th frequency band. When the gain correction term is a constant based on the path difference between the direct component and the cross component and does not have frequency characteristics, the sound reproduction system can implement the gain correction term in an amplifier circuit.

[0016] In one embodiment, the transmission gain ratio may be a ratio between the square root of the time-averaged integral of the squared value of the impulse response of the direct component at each time and the square root of the time-averaged integral of the squared value of the impulse response of the cross component at each time, in which case the sound reproduction system can set an appropriate transmission gain ratio based on the impulse response.

[0017] In one embodiment, the transmission gain ratio may be a ratio between the maximum absolute value of the impulse response of the direct component at each time and the maximum absolute value of the impulse response of the cross component at each time, in which case the sound reproduction system can set an appropriate transmission gain ratio based on the impulse responses.

[0018] In one embodiment, the gain correction term and the coefficient correction term are determined by a transmission gain ratio g ｉ In this case, the transmission gain ratio g ｉ The gain correction term and coefficient correction term can be implemented in the feedback circuit and amplifier circuit using

[0019] According to the present disclosure, a technique is provided that can reduce implementation costs related to independent control of both ears.

[0020] 10 is a schematic diagram showing an example of a sound reproduction system according to an embodiment; FIG. 11 is a block diagram illustrating binaural synthesis using two-channel speakers; FIG. 12 is a diagram illustrating the geometric positional relationship between a sound source and a listener; FIG. 13 is a diagram illustrating the definition of azimuth angle intervals; FIG. 14 is a diagram illustrating the concept of a pair of monopole sound sources whose azimuth angle changes continuously as a function of frequency; FIG. 15 is an example of a feedback circuit; FIG. 16 is a diagram illustrating four examples of implementation methods; FIG. 17 is a diagram illustrating evaluation results of the implementation method shown in FIG. 7; FIG. 18 is a graph illustrating a simulation of frequency response at a control point; FIG. 19 is a simulation result of sound pressure distribution by a sound reproduction system realized by an approximate inverse matrix; FIG. 19 is a simulation result of the impulse response of sound reaching the left ear from a pair of left and right speakers; FIG. 19 is a calculation result of the cross-correlation function between the impulse response of the direct component and the impulse response of the cross component.

[0021] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In the following description, identical or equivalent elements will be designated by the same reference numerals, and redundant description will not be repeated. The dimensional ratios of the drawings do not necessarily match those in the description. The terms "upper," "lower," "left," and "right" are based on the illustrated state and are for convenience only.

[0022] [Configuration of Sound Reproducing System] Fig. 1 is a schematic diagram showing an example of a sound reproducing system according to an embodiment. The sound reproducing system 1 shown in Fig. 1 reproduces sound in a listening space 10 in which a listener 11 is present. As shown in Fig. 1, the sound reproducing system 1 includes a multi-way speaker unit 2 (an example of a plurality of speakers) arranged in the listening space 10, and a controller 3.

[0023] The controller 3 is connected to be able to control the multiway speaker unit 2. The controller 3 is configured as a computer system including, for example, a processor such as a CPU (Central Processing Unit), memories such as RAM (Random Access Memory) and ROM (Read Only Memory), input / output devices such as a touch panel, mouse, keyboard, and display, and a communication device such as a network card. The controller 3 realizes its functions, which will be described later, by having the processor operate each piece of hardware based on a program stored in the memory or the like.

[0024] The controller 3 may be configured to be able to refer to a database 4. The database 4 stores sounds to be provided to the listener 11. As an example, the database 4 stores a left recording signal 41 corresponding to the left ear 11L of the listener 11 and a right recording signal 42 corresponding to the right ear 11R of the listener 11.

[0025] The controller 3 reproduces two-channel sound from the multi-way speaker unit 2 based on the left recording signal 41 and the right recording signal 42. The two channels are a left channel Lch and a right channel Rch, and details will be described later in the description of FIG. 2. The multi-way speaker unit 2 includes a plurality of left speakers LW corresponding to the left channel Lch, and a plurality of right speakers RW corresponding to the right channel Rch. Note that the number of speakers corresponding to each channel is arbitrary.

[0026] 1 , the multiple left speakers LW include, in order from a position close to the center of the multi-way speaker unit 2 to a position away from it, a first left speaker 21L, a second left speaker 22L, a third left speaker 23L, a fourth left speaker 24L, a fifth left speaker 25L, and a sixth left speaker 26L. The multiple right speakers RW include, in order from a position close to the center of the multi-way speaker unit 2 to a position away from it, a first right speaker 21R, a second right speaker 22R, a third right speaker 23R, a fourth right speaker 24R, a fifth right speaker 25R, and a sixth right speaker 26R.

[0027] The multiple left speakers LW and the multiple right speakers RW are each configured as a pair of left and right speakers, and the frequency bandwidths of the sounds that can be output from each pair are preset so that they differ. For example, the frequency bandwidths of the sounds that can be output are preset so that the bandwidths become lower the further away from the center position of the multiway speaker unit 2 (the position facing the median plane of the listener). As a result, when a high-frequency sound is played from the multiway speaker unit 2, the sound is output from the left and right speakers that are closer to the center position of the multiway speaker unit 2. When a low-frequency sound is played from the multiway speaker unit 2, the sound is output from the left and right speakers that are farther from the center position of the multiway speaker unit 2.

[0028] [Details of the Controller] The controller 3 applies an approximate inverse matrix, which is an approximation of the inverse matrix of the plant matrix that indicates the transfer system from the multiway speaker unit 2 to the left ear 11L and right ear 11R of the listener 11, to the left recorded signal 41 and the right recorded signal 42, to reproduce sound from the multiway speaker unit 2. To explain the approximate inverse matrix realized by the controller 3, an outline of the inverse filter matrices for two channels will first be given.

[0029] [Two-Channel Inverse Filter Matrix] Fig. 2 is a block diagram for explaining binaural synthesis using two-channel speakers. As shown in Fig. 2, the input signal for the right ear is denoted by d Ｒ (jω), the input signal for the left ear is d Ｌ (jω), where ω is the angular frequency.

[0030] The two-channel inverse filter matrix H processes two input signals and outputs two channels, here a sound source V corresponding to the right channel Rch. Ｒ , and the sound source V corresponding to the left channel Lch Ｌ The inverse filter matrix H is, for example, a matrix with 2 rows and 2 columns. The transfer matrix from the sound source to the listener's ears is taken as the plant matrix C. For example, the plant matrix C is a matrix with 2 rows and 2 columns. The signal received by the right ear is expressed as w Ｒ (jω), the signal received by the left ear is wＬ Let (jω).

[0031] For example, the input signal for the left ear d Ｌ (jω) is the matrix element H of the inverse filter matrix H １１ (jω), H １２ (jω) and the left and right sound sources V Ｌ , V Ｒ The sound is played from the left sound source V Ｌ The sound output from the right sound source V to the left ear is called the direct component. Ｒ The sound output from the right sound source V to the left ear is called the cross component. Ｒ The sound output from the left source V to the right ear is called the direct component. Ｌ The sound output from the left ear to the right ear is called the cross-component.

[0032] In order to mathematically describe the system shown in Fig. 2, the positional relationship between the sound source and the listener 11 is defined as shown in Fig. 3 and Fig. 4. Fig. 3 is a diagram explaining the geometrical positional relationship between the sound source and the listener, and Fig. 4 is a diagram explaining the definition of the azimuth angle interval. As shown in Fig. 3, the geometrical positional relationship between the sound source and the listener is expressed in a coordinate system in which the center point of both ears of the listener 11 in a planar view is the origin O of the x-axis. Ｌ Or sound source V Ｒ The angle between the line connecting the placement position of the sound source V and the origin O and the median plane of the listener (y direction) is defined as the azimuth angle interval θ. As shown in FIG. 4, the azimuth angle interval θ is the difference in azimuth angle, and Ｌ Or sound source V Ｒ The equivalent distance between the two ears is Δr. The equivalent distance is the distance corrected by taking into account the effect of diffraction in the head. Ｌ Or sound source V Ｒ Let the distance to be l.

[0033] In the positional relationship shown in FIGS. 3 and 4, the direct component p ｄ and cross component p ｃ The direct component p ｄ and cross component p ｃare expressed by the following equations (1) and (2). Here, P ０ is the loudness of the sound source, g ｄ is the direct component p ｄ The transmission gain of l ｄ is the direct component p ｄ Transmission distance, g ｃ is the cross component p ｃ The transmission gain of l ｃ is the cross component p ｃ is the transmission distance.

[0034] Transfer matrix C from the sound source to the listener's ears ０ is described by the following matrix (3). Here, g is g ｃ / g ｄ and the direct component p ｄ and cross component p ｃ The gain ratio is the transfer matrix C ０ In the direct component p ｄ The transfer function C normalized by is the following matrix (4). Therefore, the inverse filter matrix H, which is the inverse matrix of the transfer function C, is the following matrix (5).

[0035] [Approximate Inverse Matrix] The approximate inverse matrix used by the controller 3 is an element-divided inverse filter matrix H, in which some or all of the elements have been simplified for ease of implementation or improved controllability. The inverse filter matrix H shown in matrix (5) can be written as element-divided as in the following matrix (6). Here, the symbol "·" between the matrices is the Amalgamation product (element-wise product). α is the direct component p ｄ G is a gain correction term that corrects the signal strength of the direct component p ｄ and cross component p ｃ R is a coefficient correction term that corrects the difference in signal strength between the cross component p ｃ D is the antiphase term that reverses the phase of the cross component p ｃIn this way, the inverse filter matrix H is expressed as a product of the gain correction term α, the coefficient correction term G, the antiphase term R, and the delay correction term D.

[0036] [Gain of Inverse Filter Matrix] Next, we will outline the issues that arise when performing binaural control using speakers and how to deal with them. When performing binaural control using speakers, the magnitude of the gain of the inverse filter matrix H (generally the gain correction term α) becomes an issue. If the gain of the inverse filter matrix H is very large, the speaker output must be made very large, which leads to extreme deterioration of sound quality. The direct component p ｄ and cross component p ｃ In practice, the transmission gain ratio g is often close to 1, and in this case, the gain correction term α becomes very large when the wave number k and the direction of the sound source satisfy a specific relationship.

[0037] Let us consider the conditions under which sound reproduction becomes difficult to control. Ｌ Or sound source V Ｒ Assume that the distance l to the ear is sufficiently far compared to the equivalent distance Δr between the two ears. In this case, the transmission distance l ｄ， l ｃ can be approximated as in the following equations (7) and (8). direct component p ｄ and cross component p ｃ The difference in transmission distance between ｄ -l ｃ From the formulas (7) and (8), the difference Δl in the transmission distance is approximated as the following formula (9).

[0038] When the gain correction term α is rewritten using the transmission gain ratio g~1 and Δl~=Δr·sin θ, the following equation (10) is obtained. Here, when n = kΔr sin θ / (π / 2), the gain correction term α is minimized when n is an odd number, resulting in good control efficiency of sound reproduction. However, when n is an even number, the gain correction term α diverges, making it theoretically impossible to control sound reproduction. In practice, when n is an even number, the control efficiency of sound reproduction becomes very poor.

[0039] To efficiently control sound reproduction, a unique speaker can be used in which the position of the speaker changes linearly depending on the frequency of the output sound. A pair of conceptual monopole sound sources that satisfy the principle of optimal sound source distribution is shown in Figure 5. As shown in Figure 5, the monopole sound sources are those in which the frequency and the opening angle change continuously while maintaining a constant relationship so as to satisfy n = 1 in n = kΔr sin θ / (π / 2). In this case, the plant matrix C and the inverse filter matrix H are very simple matrices as follows:

[0040] Even when the principle of distributed sound source placement described above is used, there are still remaining issues, such as the need to implement a convolution operation and the need to design an inverse filter for each frequency in the transfer system of each discretized band.

[0041] For this reason, the sound reproduction system 1 employs a conceptual monopole sound source that satisfies the principle of optimal sound source distribution, while dividing the audible frequency band into a finite number of bands and using a representative frequency (center frequency) of each divided frequency band to represent the frequency band, thereby reducing the computational cost. Furthermore, by configuring an inverse system that does not require convolution operations, an embodiment is realized in which the implementation cost of the inverse filter is reduced. Below, an embodiment expanded to multiple sound sources is considered. Assume that the number of sound sources is 2m, and the audible frequency range is divided into m parts and assigned to left and right sound sources. In this case, the inverse filter matrix H, which is the inverse matrix of the transfer system for the i-th band, is ｉ can be written as the following matrix (11). where Δl ｉ is the direct component p of the sound source in the i-th band ｄ and cross component p ｃ is the path difference between g ｉ is the direct component p of the sound source in the i-th band ｄ and cross component p ｃ is the gain ratio.

[0042] The controller 3 calculates the inverse filter matrix H ｉ is not implemented as a filter as it is, but as an inverse filter matrix H ｉ The gain correction term α ｉ , coefficient correction term Gｉ , antiphase term R ｉ , and the delay correction term D ｉ Before explaining the implementation of each element, the direct component p ｄ and cross component p ｃ The time difference Δt ｉ is the direct component p ｄ and cross component p ｃ Path difference Δl ｉ and the speed of sound c, it is defined as follows:

[0043] [Implementation of gain correction term] Gain correction term α ｉ There are two ways to implement the gain correction term α ｉ is a term with frequency characteristics, and a feedback circuit is used to calculate the gain correction term α ｉ The second pattern is a method of implementing the gain correction term α ｉ is a term that does not have frequency characteristics, and the gain correction term α ｉ This is a method of implementing a constant value.

[0044] First, the first pattern will be described. ｉ As shown in the matrix (11), when the denominator becomes 0, the gain correction term α ｉ , the stabilization parameter β ｉ is used to approximate it as shown in the following equation (12). Approximated gain correction term α ｉ When the input is x and the output is y, the input x and the output y can be described by the following equation (13). Equation (13) can be implemented by the feedback circuit shown in FIG. 6. Note that the stabilization parameter β ｉ is the transmission gain ratio g ｉ After determining the above, it is set to satisfy the following formula (14). The transmission gain ratio g ｉ The method for determining this will be described later.

[0045] Next, the second pattern will be described. In the second pattern, the gain correction term α ｉ is approximated by the following equation (15). The formula (15) can be implemented by an amplifier circuit. Note that the numerator of the formula (15) is γ, which may be a value in the range of 0.1 to 10.

[0046] As explained above, the gain correction term α ｉ can be implemented by a feedback circuit or an amplifier circuit. ｉ The transmission gain ratio g ｉ The gain correction term α may be a constant regardless of ｉ is a constant, for example, a value in the range of 0.1 to 10 is used as the gain correction term α ｉ It can be set as:

[0047] [Implementation of coefficient correction term] Coefficient correction term G ｉ is the direct component p ｄ Bypass the cross component p ｃ In other words, the coefficient correction term G ｉ is the transmission gain ratio g ｉ This can be implemented using an amplifier circuit.

[0048] [Implementation of the antiphase term] The antiphase term R ｉ is the coefficient correction term G ｉ Similarly, the direct component p ｄ Bypass the cross component p ｃ It is implemented as a circuit that processes only the antiphase term R ｉ There are two ways to implement the matrix (16). The first way is to implement the matrix (16) by using the cross component p ｃ This is a pattern that performs inversion processing. The matrix (16) can be implemented with an inverting circuit. The second pattern is the antiphase term R ｉ is a pattern that approximates the following matrix (17). where ω ｉ = 2πf ｉ and f ｉ is the center frequency of the i-th frequency band. The matrix (17) is ｉBy replacing the antiphase processing with a simple delay circuit, the antiphase term R ｉ and a delay correction term D ｉ It is also possible to integrate these and implement them as a single simple delay circuit.

[0049] [Implementation of Delay Correction Term] Delay Correction Term D ｉ is the coefficient correction term G ｉ Similarly, the direct component p ｄ Bypass the cross component p ｃ In other words, it is implemented as a circuit that adds a simple delay process to only the delay correction term D ｉ can be implemented with a simple delay circuit.

[0050] [Summary of implementation method] The gain correction term α ｉ , coefficient correction term G ｉ , antiphase term R ｉ , and the delay correction term D ｉ The mounting methods are combinations of possible patterns for each element. Fig. 7 is a diagram illustrating four examples of mounting methods. As shown in Fig. 7, the mounting methods for the controller 3 include a first mounting method M1, a second mounting method M2, a third mounting method M3, and a fourth mounting method M4.

[0051] The first implementation method M1 uses the inverse filter matrix H ｉ An approximate inverse matrix is ​​used, which is approximated as in matrix (18). In the following, as a notation method, elements assigned to different circuits are explicitly divided by operators. For example, "ab" indicates that ab is realized by one circuit, and "a·b" indicates that ab is realized by two circuits. As shown in matrix (18) and in FIG. 7, the gain correction term α ｉ and coefficient correction term G ｉ is implemented in an amplifier circuit, and the antiphase term R ｉ and delay correction term D ｉ is implemented with a delay circuit. Each coefficient matrix is ​​implemented in the order of an amplifier circuit and a delay circuit.

[0052] In order to achieve independent control of both ears in two channels, the signal related to the left channel and the signal related to the right channel must have the opposite sign after correcting the arrival time difference between the direct component and the cross component in the processing of the recorded signal on one side. In the sound reproduction system 1, the above condition is satisfied when there is a delay time Δt between the signals. ｉ In the first implementation method M1, the delay time Δt ｉ The polarity inversion is implemented by the delay circuit. ｄ and cross component p ｃ The time difference (delay amount π / ω) at which the phase difference with ｉ ) can be realized. In other words, the delay circuit can be realized by ｄ and cross component p ｃ The time difference between the arrival times of the direct component p ｄ and cross component p ｃ The arrival time difference between ｉ ) is the sum of the final time difference. The delay circuit then sets the final time difference at the center frequency of the frequency bandwidth assigned to the speaker, and applies this time difference to all frequencies included in the frequency band assigned to the speaker. In this way, the sound reproduction system 1 can significantly reduce calculation costs by uniformly setting a time difference for each discretized frequency band.

[0053] The second implementation method M2 uses the inverse filter matrix H ｉ An approximate inverse matrix is ​​used, which is approximated as in matrix (19). As shown in matrix (19) and in FIG. 7, the gain correction term α ｉ and coefficient correction term G ｉ is implemented in an amplifier circuit, and the antiphase term R ｉ is implemented as an inverting circuit, and the delay correction term D ｉ is implemented by a delay circuit. Each coefficient matrix is ​​implemented in the order of an amplifier circuit, an inverting circuit, and a delay circuit. Compared with the first implementation method M1, the antiphase term R ｉ In the second implementation method M2, the delay circuit is implemented as a direct component p ｄand cross component p ｃ The arrival time difference between ｉ ) is uniformly provided for each frequency band. This allows the sound reproduction system 1 to significantly reduce the calculation cost.

[0054] The third implementation method M3 uses the inverse filter matrix H ｉ An approximate inverse matrix is ​​used, which is approximated as the matrix (20). As shown in matrix (20) and in FIG. 7, the gain correction term α ｉ is implemented in a feedback circuit, and the coefficient correction term G ｉ is implemented in an amplifier circuit, and the antiphase term R ｉ and delay correction term D ｉ is implemented by a delay circuit. Each coefficient scalar and matrix is ​​implemented in the order of a feedback circuit, an amplifier circuit, and a delay circuit. Compared with the first implementation method M1, the gain correction term α ｉ In the third implementation method M3, the feedback circuit is implemented with a frequency-dependent gain correction term α ｉ As a result, the sound reproduction system 1 calculates the gain correction term α ｉ This allows for more accurate crosstalk cancellation than when is a constant.

[0055] The fourth implementation method M4 uses the inverse filter matrix H ｉ An approximate inverse matrix is ​​used, which is approximated as the matrix (21). As shown in matrix (21) and in FIG. 7, the gain correction term α ｉ is implemented in a feedback circuit, and the coefficient correction term G ｉ is implemented in an amplifier circuit, and the antiphase term R ｉ is implemented as an inverting circuit, and the delay correction term D ｉ is implemented by a delay circuit. Each coefficient scalar and matrix is ​​implemented in the order of a feedback circuit, an inverting circuit, an amplifier circuit, and a delay circuit. Compared with the first implementation method M1, the gain correction term α ｉ and the antiphase term R ｉ In the fourth implementation method M4, the delay circuit is implemented as a direct component p ｄ and cross component p ｃ The arrival time difference between ｉ) is provided uniformly for each frequency band. The feedback circuit uses a gain correction term α ｉ As a result, the sound reproduction system 1 calculates the gain correction term α ｉ This allows for more accurate crosstalk cancellation than when is a constant.

[0056] [Simplification of the Design Method of the Inverse System] In conventional systems that employ the principle of optimal sound source distribution, it is necessary to design an inverse filter matrix for each frequency. In contrast, the sound reproduction system 1 simplifies the design method of the inverse filter matrix by designing the direct component p ｄ and cross component p ｃ Delay time Δt ｉ and the transmission gain ratio g ｉ In other words, when using 2m speakers, two parameters are set for each pair of left and right speakers, so a total of 2m parameters need to be set. In this way, the sound reproduction system 1 can also simplify the design method for an inverse system.

[0057] [Setting the arrival time difference (delay time) between the direct component and the cross component] The difference in propagation distance Δl is approximated by the above-mentioned formula (9), so the delay time Δt ｉ can be set by the following formula (22). In reality, it is necessary to consider the time it takes to bypass the head and the ears. ｄ and the cross-component p ｃ The time difference at which the correlation between the impulse responses is maximized is adopted. The time difference at which the waveforms of the impulse responses overlap most when shifted along the time axis. The time difference may be determined by shifting the waveforms, or may be calculated using the cross-correlation function shown in the following equation (23). Here, h ｄ，ｉ is the i-th direct component p ｄ，ｉ is the impulse response of h ｃ，ｉ is the i-th cross component p ｃ，ｉ The impulse response of arg is zero-filled for non-causal components t<0. ｘmaxf(x) is an operator that returns x when the function f(x) gives the maximum value.

[0058] [Setting the transfer gain ratio between the direct component and the cross component] Transfer gain ratio g ｉ In general, only the distance attenuation of the point sound source needs to be considered, and the direct component p ｄ Transmission gain g ｄ，ｉ and cross component p ｃ Transmission gain g ｃ，ｉ is expressed by the following equation (24). Transmission distance l ｄ， l ｃ Since is approximated by equations (7) and (8), the transmission gain ratio g ｉ can be approximated by the following equation (25). Furthermore, when the transmission distance l is much larger than the equivalent distance Δr between the two ears, the transmission gain ratio g ｉ It can be set to 1.

[0059] Transfer gain ratio g ｉ is the delay time Δt ｉ Similarly, in reality, it is necessary to consider the distance around the head and the auricles. ｉ There are various methods for evaluating the energy density, but here, a method using an effective value and a method using a maximum energy value are exemplified.

[0060] First, the transfer gain ratio g ｉ This method involves taking the square root of the square integral of the impulse response taking the head into account, divided by the integral time length, and using this as the transfer gain for each of the direct and cross components. For example, the direct component p ｄ，ｉ Transmission gain g ｄ，ｉ and cross component p ｃ，ｉ Transmission gain g ｃ，ｉ can be expressed as the following equation (26). where T is the time it takes for the impulse response to decay sufficiently. ｉ is the direct component p ｄ，ｉ The square root of the time-averaged integral of the squared value of the impulse response at each time and the cross component pｃ，ｉ is the ratio of the square root of the time-averaged integral of the squared value of the impulse response of

[0061] Next, the transfer gain ratio g ｉ The method for evaluating the maximum energy component p is explained below. It can be said that there is a high possibility that there is a correlation between the maximum energy components. ｄ，ｉ Transmission gain g ｄ，ｉ and cross component p ｃ，ｉ Transmission gain g ｃ，ｉ can be expressed as the following equation (27). Transfer gain ratio g ｉ is the direct component p ｄ，ｉ The maximum absolute value of the impulse response at each time and the cross component p ｃ，ｉ The maximum absolute value of the impulse response of each time can be calculated as the ratio of

[0062] [Evaluation of Each Implementation Method] Each of the above-described implementation methods was evaluated together with a comparative example. The results are shown in FIG. 8. FIG. 8 is a diagram explaining the evaluation results of the implementation method shown in FIG. 7. The comparative example is a system that implements an inverse filter matrix, and is a conventional system that employs the principle of optimal sound source distribution. The evaluation is a relative evaluation, and the ranking is determined in descending order of evaluation from A to C. The evaluation items were "simplicity of theory," "accuracy of the inverse system," "low circuit implementation cost," and "simplicity of inverse system design."

[0063] Regarding "logical simplicity," the comparative example was rated "A," and the first implementation method M1 to the fourth implementation method M4 were rated "B." The first implementation method M1 to the fourth implementation method M4 were rated as being more complex than the comparative example because they break down the formulas of the comparative example into elements.

[0064] [Accuracy of the inverse system] In the comparative example, the product of the plant matrix and the inverse filter matrix is ​​theoretically an identity matrix, so the comparative example was rated "A." In contrast, the first implementation method M1 to the fourth implementation method M4 approximate the elements of the inverse filter matrix, so they were rated "B" or lower. Note that, among the first implementation methods M1 to M4, the first implementation method M1 has the most similar items, so only the first implementation method M1 was rated "C," and the second implementation methods M2 to M4 were rated "B."

[0065] [Low circuit implementation cost] The comparative example requires a convolution circuit, so it was rated lower than the first to fourth implementation methods M1 to M4. Furthermore, the third implementation method M3 and the fourth implementation method M4 have relatively complex feedback circuits, so they were rated lower than the first implementation method M1 and the second implementation method M2. Based on the above, the comparative example was rated "C," the first implementation method M1 and the second implementation method M2 were rated "A," and the third implementation method M3 and the fourth implementation method M4 were rated "B."

[0066] [Ease of inverse system design] The comparative example was rated lower than the first to fourth implementation methods M1 to M4 because it required designing an inverse filter matrix for each frequency. The third implementation method M3 and the fourth implementation method M4 were rated lower than the first implementation method M1 and the second implementation method M2 because they required setting stabilization parameters for the feedback circuit. Based on the above, the comparative example was rated "C," the first implementation method M1 and the second implementation method M2 were rated "A," and the third implementation method M3 and the fourth implementation method M4 were rated "B."

[0067] [Summary of the embodiment] According to the sound reproduction system 1, the inverse filter matrix H ｉ is approximated by an approximate inverse matrix. The approximate inverse matrix is ​​applied to the left recording signal 41 and the right recording signal 42, thereby causing the multi-way speaker unit 2 to reproduce the sound. The approximate inverse matrix is ｉ , coefficient correction term G ｉ , antiphase term R ｉ , and the delay correction term D ｉand . That is, the inverse filter matrix that performs the convolution operation of a long filter coefficient is approximately divided into four elements that perform simpler operations, such as gain correction, coefficient correction, phase reversal, and delay correction. The functions related to the four elements are then shared and implemented by multiple circuits. Therefore, the sound reproduction system 1 can reduce the implementation cost related to independent control of both ears compared to when the inverse filter matrix is ​​implemented as is.

[0068] [Modifications] Although various exemplary embodiments have been described above, the present invention is not limited to the above exemplary embodiments, and various omissions, substitutions, and modifications may be made.

[0069] The numbers of right speakers and left speakers described in the above embodiments may be changed as appropriate, and the numbers of right speakers and left speakers do not have to be the same.

[0070] In the following, examples carried out by the present inventors will be described to explain the above effects.

[0071] The sound reproduction system 1 shown in Figure 1 was used. The inverse filter matrix was implemented using the second implementation method M2. The frequency band was divided into four parts. Sound was output to only one ear, and the frequency response at the other ear (control point) was simulated. The results are shown in Figure 9. Figure 9 is a graph simulating the frequency response at the control point. The horizontal axis is frequency [Hz], and the vertical axis is sound pressure level [dB]. With the second implementation method M2, crosstalk is completely canceled in numerical calculations. If the recorded signal is accurately reproduced at the control point, the vertical axis will be 0 dB. As shown in Figure 9, although the sound pressure level varies depending on the frequency, it was confirmed that it was generally within -5 dB.

[0072] A sound pressure distribution was simulated in the listening space 10 described above, with the listener 11 positioned at the origin. The sound reproduction system 1 was the same as the system used in the evaluation of FIG. 9 . The results are shown in FIG. 10 . FIG. 10 shows the results of simulating the sound pressure distribution using a sound reproduction system realized using an approximate inverse matrix. FIG. 10 is a diagram in which the sound pressure level [dB] is expressed by shading. As shown in FIG. 10 , the left ear of the listener (at the origin) was approximately 0 dB, and the right ear was approximately −20 dB, meaning that sufficient sound reduction was confirmed in the right ear. In this way, it was confirmed that independent control of both ears can be achieved using an approximate inverse matrix.

[0073] Next, an example of parameter setting will be described. An example of parameter calculation will be shown when a pair of left and right speakers are placed at positions where the opening angle θ of the speakers is 29 degrees. First, the delay time Δt ｉ is calculated. FIG. 11 shows the results of simulating the impulse response of sound reaching the left ear from a pair of left and right speakers. The horizontal axis represents time [ms], and the vertical axis represents relative sound pressure. The relative sound pressure is normalized so that the peak value of the impulse response of the direct component reaching the left ear from the left speaker (head-related impulse response taking the head into account) is 1. The solid line data shown in FIG. 11 is the impulse response of the direct component reaching the left ear from the left speaker, and the dashed dotted line data is the impulse response of the cross component reaching the left ear from the right speaker. A time difference (shift time) is applied between the impulse response of the direct component and the impulse response of the cross component, and a cross-correlation function is prepared to return the magnitude of overlap of the impulse responses at the applied time difference as a correlation value. FIG. 12 shows the calculation results of the cross-correlation function between the impulse response of the direct component and the impulse response of the cross component. The horizontal axis represents the shift time, and the vertical axis represents the correlation value. As shown in Figure 12, the shift time at which the correlation value reached its maximum was -0.29 ms. This indicates that the direct component arrived 0.29 ms earlier than the cross component. Therefore, in the frequency bands assigned to the pair of left and right speakers, the delay time Δt ｉ may be set to 0.29 ms.

[0074] Next, the transmission gain ratio gｉ The transfer gain ratio g is calculated based on the effective value. ｉ When calculating the transfer gain of the direct component and the cross component, the square integral value of the impulse response considering the head shown in FIG. 11 is divided by the integral time length, and the square root is taken to determine the transfer gain of each of the direct component and the cross component (see Equation (26)). Then, the ratio of the transfer gain of the direct component to the cross component is calculated. This gives the transfer gain ratio g ｉ becomes 2.04.

[0075] The maximum energy value determines the transfer gain ratio g ｉ When calculating the transfer gain ratio g, the ratio between the maximum absolute value of the impulse response of the direct component shown in FIG. 11 and the maximum absolute value of the impulse response of the cross component is calculated. ｉ is 2.44. As described above, it has been confirmed that the parameters can be calculated clearly and simply.

[0076] 1...sound reproduction system, 2...multi-way speaker unit (an example of multiple speakers), 3...controller, 10...listening space, 11...listener, 41...left recording signal, 42...right recording signal, LW...multiple left speakers, RW...multiple right speakers.

Claims

1. 1. A sound reproduction system for reproducing sound in a listening space in which a listener is present, comprising: a plurality of speakers arranged in the listening space; a controller that reproduces two-channel sounds, i.e., a left channel and a right channel, from the plurality of speakers based on a left recording signal corresponding to the left ear of the listener and a right recording signal corresponding to the right ear of the listener; Equipped with The plurality of speakers include: a plurality of left speakers corresponding to the left channel; a plurality of right speakers corresponding to the right channel; and a discrete frequency bandwidth is preset for each of the plurality of left speakers and the plurality of right speakers so that the frequency bandwidths of sounds that can be output are different from each other; the controller applies an approximate inverse matrix, which is an approximation of an inverse matrix of a plant matrix that indicates a transfer system from the plurality of speakers to both ears, to the left recording signal and the right recording signal, thereby reproducing sound from the plurality of speakers; the sounds output from the plurality of speakers include a direct component, which is sounds output from a plurality of left speakers to the left ear of the listener and a plurality of right speakers to the right ear of the listener, and a cross component, which is sounds output from a plurality of left speakers to the right ear of the listener and a plurality of right speakers to the left ear of the listener; The approximate inverse matrix is is expressed as a product of a gain correction term that corrects the signal strength of the direct component according to a frequency band, a coefficient correction term that corrects the difference in strength between the direct component and the cross component according to a frequency band, an anti-phase term that inverts the phase of the cross component, and a delay correction term that provides a time difference according to the frequency band of the cross component, the controller has a plurality of circuits that implement the gain correction term, the coefficient correction term, the anti-phase term, and the delay correction term; Sound reproduction system.

2. the controller includes an amplifier circuit that implements the gain correction term and the coefficient correction term, and a delay circuit that implements the antiphase term and the delay correction term; the delay circuit uniformly sets a sum of a time difference at which a phase difference between the direct component and the cross component becomes 180 degrees at a center frequency of the frequency bandwidth and an arrival time difference between the direct component and the cross component for each discretized frequency band.

2. The sound reproduction system according to claim 1.

3. the controller includes an amplifier circuit that implements the gain correction term and the coefficient correction term, an inverting circuit that implements the antiphase term, and a delay circuit that implements the delay correction term; The inverting circuit inverts the polarity of the signal; the delay circuit sets a uniform arrival time difference between the direct component and the cross component for each discretized frequency band; 2. The sound reproduction system according to claim 1.

4. the controller includes a feedback circuit that implements the gain correction term, an amplifier circuit that implements the coefficient correction term, and a delay circuit that implements the antiphase term and the delay correction term; the feedback circuit calculates the gain correction term according to frequency; the delay circuit uniformly sets a sum of a time difference at which a phase difference between the direct component and the cross component becomes 180 degrees at a center frequency of the frequency bandwidth and an arrival time difference between the direct component and the cross component for each discretized frequency band.

2. The sound reproduction system according to claim 1.

5. the controller includes a feedback circuit that implements the gain correction term, an amplifier circuit that implements the coefficient correction term, an inverting circuit that implements the antiphase term, and a delay circuit that implements the delay correction term; the feedback circuit calculates the gain correction term according to frequency; The inverting circuit inverts the polarity of the signal; the delay circuit sets a uniform arrival time difference between the direct component and the cross component for each discretized frequency band; 2. The sound reproduction system according to claim 1.

6. 6. The sound reproduction system according to claim 2, wherein the arrival time difference is a time difference at which a correlation between the impulse response of the direct component and the impulse response of the cross component is maximized.

7. 4. A sound reproduction system according to claim 2 or 3, wherein the gain correction term is a number in the range of 0.1 to 10.

8. The gain correction term is γ / (1+g ｉ ２ ) where γ is a number ranging from 0.1 to 10, and g ｉ 4. The sound reproduction system according to claim 2, wherein i is a transmission gain ratio, which is a ratio of signal strengths resulting from a path difference between a direct component and a cross component of a speaker in the i-th frequency band.

9. The sound reproduction system of claim 8, wherein the transmission gain ratio is the ratio of the square root of the value obtained by integrating the squared value of the impulse response of the direct component at each time point with the time average, to the square root of the value obtained by integrating the squared value of the impulse response of the cross component at each time point with the time average.

10. The sound reproduction system according to claim 8, wherein the transmission gain ratio is the ratio of the maximum absolute value of the impulse response of the direct component at each time to the maximum absolute value of the impulse response of the cross component at each time.

11. The gain correction term and the coefficient correction term are calculated based on a transmission gain ratio g ｉ 6. The sound reproduction system according to claim 4, wherein

12. The sound reproduction system of claim 11, wherein the transmission gain ratio is the ratio of the square root of the value obtained by integrating the squared value of the impulse response of the direct component at each time point with the time average, to the square root of the value obtained by integrating the squared value of the impulse response of the cross component at each time point with the time average.

13. The sound reproduction system according to claim 11, wherein the transmission gain ratio is the ratio of the maximum absolute value of the impulse response of the direct component at each time to the maximum absolute value of the impulse response of the cross component at each time.

Images