JP2010171785A - Coefficient calculation device for head-related transfer function interpolation, sound localizer, coefficient calculation method for head-related transfer function interpolation and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a coefficient calculation device for head-related transfer function interpolation for calculating an interpolation coefficient for interpolation of head-related transfer function. <P>SOLUTION: The device has a Fourier coefficient storage part 13 for interpolation wherein a complex Fourier coefficient of head-related transfer function measured at an azimuth angle for each first angle unit to a dummy of test subject is stored; a coefficient calculation part 14 for interpolation for calculating an interpolation coefficient for interpolating a complex Fourier coefficient corresponding to an interpolation azimuth angle which is a multiple of the first angle unit and is not an angle for each second angle unit which is at least twice the first angle unit, so that a complex Fourier coefficient corresponding to an interpolation azimuth angle stored in the Fourier coefficient storage part 13 for interpolation is interpolated by multiplying a complex Fourier coefficient corresponding to three or more azimuth angles for each second angle unit stored in the Fourier coefficient storage part 13 for interpolation by an interpolation coefficient; and a coefficient accumulation part 15 for interpolation for accumulating a coefficient for interpolation. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a head transfer function interpolation coefficient calculation apparatus and the like that perform processing for calculating a coefficient for interpolation related to a head transfer function.

  The head-related transfer function and the head impulse response that expresses the head-related transfer function in the time domain are important for adding spatial information to sound signals output from headphones and earphones in a virtual auditory display system. It is. However, measuring head-related transfer functions in all directions is a cumbersome and time consuming task. Furthermore, since the head-related transfer function varies depending on the individual, it is preferable to measure for each subject. However, in order to measure the head-related transfer function, the subject must be stationary for a long time. The burden on the subject was also very large.

  Therefore, it is necessary to interpolate the head-related transfer function. In the measurement of the head-related transfer function for the subject, the measurement is performed at a somewhat rough angle. Then, by interpolating the head-related transfer function measured at the rough angle, a fine-angle head-related transfer function used in the virtual auditory display system can be obtained. In addition, the interpolation method of the head related transfer function has been developed in the last ten years, and for example, the interpolation methods of the following non-patent documents 1 to 4 are known.

T.A. Djelani, C.I. Porschmann, J. et al. Sahrhage, J. et al. Blauert, “An interactive virtual-environmental generator for psychological research and algebraic research and algebraic research. II. 86, no. 6, p. 1046-1053, November 2000 M.M. A. Blommer, G.M. H. Wakefield, “Pole-zero application for head-related transfer functions using a logical error criterion”, IEEE Trans. on Speech and Audio Processing, vol. 5, no. 3, p. 278-287, May 1997 K. Watanabe, S.M. Takane, Y. et al. Suzuki, "A novel interpolation method of HRTFs based on the common-acoustic-pole and zero model", Acta Acoustica united Acoustic. 91, no. 6, p. 958-966, November, December 2005 J. et al. Chen, B.M. D. V. Veen, K.M. E. Hecox, “A spatial feature extraction and legalization model for the head-related transfer function”, J. Am. Acoustic. Soc. Am. , Vol. 97, no. 1, p. 439-452, January 1995

It has been desired to develop a highly accurate method for the interpolation method of the actually measured head-related transfer function.
The present invention has been made in the above situation, and an object thereof is to provide a head transfer function interpolation coefficient calculation device and the like that perform processing for calculating a coefficient for interpolation related to the head transfer function.

  In order to achieve the above object, the head-related transfer function interpolation coefficient calculation apparatus according to the present invention is directed to the subject's dummy in each first angle unit with reference to a reference azimuth that is a reference azimuth. An interpolation Fourier coefficient storage unit that stores a complex Fourier coefficient of a head-related transfer function measured by an angle, and an angle that is an integral multiple of the first angle unit with respect to the reference azimuth angle, One or more interpolated azimuth angles that are not angles for each second angle unit, which is an angle unit of M times the first angle unit (M is an integer of 2 or more) with respect to the azimuth angle. An interpolation coefficient which is a coefficient for interpolating a complex Fourier coefficient corresponding to is stored in the interpolation Fourier coefficient storage unit, and a complex Fourier coefficient corresponding to the interpolation azimuth is stored in the interpolation Fourier coefficient storage. Stored in the department An interpolation coefficient calculation unit that calculates an interpolation by multiplying a complex Fourier coefficient corresponding to three or more azimuth angles for each second angle unit with the reference azimuth angle as a reference. And an interpolation coefficient storage section that stores the interpolation coefficient calculated by the interpolation coefficient calculation section.

  With such a configuration, an interpolation coefficient used for interpolation of the head related transfer function can be calculated. By using this interpolation coefficient, it is possible to accurately interpolate the subject's head-related transfer function measured in units of coarse azimuth. In addition, since the head-related transfer function of the subject can be measured in units of coarse azimuth, the time for restraining the subject when measuring the head-related transfer function is shortened, and the load on the subject is reduced. The head-related transfer function of the subject can be obtained in a short time.

In the head transfer function interpolation coefficient calculation apparatus according to the present invention, the interpolation coefficient has the same value even if the interpolation azimuth corresponding to the interpolation coefficient changes by an integral multiple of the second angle unit. May be a coefficient.
With such a configuration, it is possible to appropriately calculate an interpolation coefficient for performing interpolation in consideration of periodicity in the azimuth direction of the head-related transfer function.

In the head transfer function interpolation coefficient calculation apparatus according to the present invention, an interpolation for receiving a head impulse response corresponding to an azimuth for each first angle unit with respect to the dummy of the subject as a reference with respect to the reference azimuth. An impulse response receiving unit, and a Fourier transform unit for Fourier transforming the head impulse response received by the interpolation impulse response receiving unit and storing the Fourier transform complex Fourier coefficient in the interpolation Fourier coefficient storage unit; , May be further provided.
With such a configuration, the complex Fourier coefficient of the head-related transfer function can be obtained from the head impulse response to the dummy.

  In the head transfer function interpolation coefficient calculation apparatus according to the present invention, the complex Fourier of the head related transfer function measured at the azimuth angle for each second angle unit with respect to the reference azimuth angle with respect to the subject. A Fourier coefficient storage unit storing coefficients, an interpolation coefficient corresponding to one interpolation azimuth accumulated by the interpolation coefficient storage unit, and the interpolation coefficient stored in the Fourier coefficient storage unit. An interpolation unit that calculates a complex Fourier coefficient corresponding to the one interpolation azimuth by performing interpolation using a complex Fourier coefficient corresponding to three or more azimuths corresponding to the complex Fourier coefficient used in the calculation; An interpolation Fourier coefficient storage unit that stores the complex Fourier coefficients calculated by the interpolation unit may be further included.

  With this configuration, the complex Fourier coefficient of the head-related transfer function for the subject can be interpolated using the interpolation coefficient. Therefore, the head-related transfer function in units of fine azimuth can be obtained using the head-related transfer function of the subject measured in units of coarse azimuth.

In the head transfer function interpolation coefficient calculation apparatus according to the present invention, an impulse response reception unit that receives a head impulse response corresponding to an azimuth for each second angle unit with respect to the reference relative to the reference azimuth. And a Fourier transform unit that Fourier-transforms the head impulse response received by the impulse response reception unit and accumulates the Fourier transform complex Fourier coefficient in the Fourier coefficient storage unit.
With such a configuration, the complex Fourier coefficient of the head-related transfer function can be obtained from the head impulse response to the subject.

The sound image localization apparatus according to the present invention receives the head-related transfer function interpolation coefficient calculation device, an azimuth angle accepting unit that accepts a sound image localization azimuth that is an azimuth angle for performing sound image localization, and a sound signal for performing sound image localization. A complex Fourier coefficient accumulated by the sound signal accepting unit and the interpolated Fourier coefficient accumulating unit, and using the complex Fourier coefficient corresponding to the sound image localization azimuth, the sound signal accepted by the sound signal accepting unit A sound image localization processing unit that performs sound image localization processing; and an output unit that outputs a sound signal subjected to sound image localization processing by the sound image localization processing unit.
With such a configuration, the sound image localization process can be performed using the head transfer function of the subject interpolated in the head transfer function interpolation coefficient calculation device.

In the sound image localization apparatus according to the present invention, the interpolation unit may perform an interpolation process corresponding to an interpolation azimuth corresponding to the sound image localization azimuth received by the azimuth angle reception unit.
With such a configuration, a minimum amount of interpolation processing is performed as compared with a case where interpolation processing corresponding to many interpolation azimuth angles is performed, and the load of the interpolation processing can be reduced. Further, since it is not necessary to store the interpolated head related transfer functions corresponding to various interpolation azimuth angles, the storage area can be saved.

  According to the head transfer function interpolation coefficient calculation device and the like according to the present invention, it is possible to calculate an interpolation coefficient used when performing interpolation related to the head transfer function. In addition, when performing interpolation related to the head-related transfer function using the interpolation coefficient, it is possible to realize highly accurate interpolation.

1 is a block diagram showing the configuration of a sound image localization apparatus according to Embodiment 1 of the present invention. The figure for demonstrating the interpolation in the same embodiment Flowchart showing the operation of the sound image localization apparatus according to the embodiment The flowchart which shows the detail of operation | movement of the sound image localization apparatus by the embodiment The figure for demonstrating the measurement of the head impulse response in the same embodiment The figure which shows an example of the head-related transfer function in the embodiment The figure for demonstrating the periodicity of the head-related transfer function in the embodiment The figure which shows an example of the comparison result in the same embodiment Schematic diagram showing an example of the appearance of the computer system in the embodiment The figure which shows an example of a structure of the computer system in the embodiment

  Hereinafter, a head transfer function interpolation coefficient calculation apparatus according to the present invention will be described using embodiments. In the following embodiments, components and steps denoted by the same reference numerals are the same or equivalent, and repetitive description may be omitted.

(Embodiment 1)
A head transfer function interpolation coefficient calculation device 1 and a sound image localization device 2 according to Embodiment 1 of the present invention will be described with reference to the drawings. The head transfer function interpolation coefficient calculation apparatus 1 according to the present embodiment performs a process of calculating an interpolation coefficient used when performing interpolation related to a head-related transfer function (HRTF). . The sound image localization apparatus 2 according to the present embodiment performs sound image localization processing using the head-related transfer function interpolated by the head-related transfer function interpolation coefficient calculation apparatus 1 using the interpolation coefficient.

  FIG. 1 is a block diagram showing a configuration of a sound image localization apparatus 2 according to this embodiment. The sound image localization apparatus 2 according to the present embodiment includes a head transfer function interpolation coefficient calculation apparatus 1, an azimuth angle reception unit 21, a sound signal reception unit 22, a sound image localization processing unit 23, and an output unit 24. .

  The head transfer function interpolation coefficient calculation apparatus 1 according to the present embodiment includes an interpolation impulse response reception unit 11, an interpolation Fourier transform unit 12, an interpolation Fourier coefficient storage unit 13, and an interpolation coefficient calculation unit 14. , An interpolation coefficient storage unit 15, an impulse response reception unit 16, a Fourier transform unit 17, a Fourier coefficient storage unit 18, an interpolation unit 19, and an interpolation Fourier coefficient storage unit 20.

  First, a method for performing head-related transfer function interpolation in the head-related transfer function interpolation coefficient calculation apparatus 1 according to the present embodiment will be briefly described. Here, the head-related transfer function for the subject's dummy and the head-related transfer function for the subject are measured. The test subject dummy is a general dummy used in the measurement of the head-related transfer function, for example, a dummy composed of a head and a torso. The subject is a person whose head transfer function is to be measured. Regarding the measurement of the head-related transfer function for the subject's dummy, the head-related transfer function is measured for each fine azimuth, that is, for each first angle unit. And the coefficient for interpolation is calculated beforehand using the measured head related transfer function. On the other hand, for the measurement of the head-related transfer function for the subject, the head-related transfer function is specified for each coarse azimuth, that is, for each second angle unit. The second angle unit is an angle unit that is M times the first angle unit (M is an integer of 2 or more). Then, by interpolating the subject's head-related transfer function, the subject's head-related transfer function whose azimuth angle is the first angular unit can be obtained. Note that the azimuth angle used in the present embodiment is the azimuth angle of the subject to be measured for the head-related transfer function and the subject's dummy, and is the azimuth angle around the subject in a plane centered on the subject.

  The interpolating impulse response receiving unit 11 receives a head-related impulse response (HRIR) corresponding to an azimuth angle for each first angle unit with respect to the subject's dummy. The reference azimuth is a reference azimuth that can be arbitrarily set. In the present embodiment, the reference azimuth is the front of the subject or the subject's dummy. When the impulse signal is given to the test subject dummy, the head impulse response that is the sound signal itself collected by the microphone at the position of the test subject's dummy ear is received by the interpolation impulse response receiving unit 11. Also good. On the other hand, when a time extension pulse is given to the subject's dummy, a predetermined analysis is performed on the sound signal collected by the microphone at the position of the subject's dummy ear, and an impulse signal is sent to the subject's dummy. The head impulse response obtained when given may be reproduced, and the reproduced impulse signal may be received by the interpolation impulse response receiving unit 11. In the present embodiment, the latter case will be described.

  For example, the interpolation impulse response receiving unit 11 may receive a head impulse response input from a microphone, may receive a head impulse response input from another device, and may connect a wired or wireless communication line. The head impulse response transmitted via the head may be received, or the head impulse response read from a predetermined recording medium (for example, an optical disk, a magnetic disk, or a semiconductor memory) may be received. Note that the interpolation impulse response receiving unit 11 may or may not include a device for receiving (for example, a modem or a network card). The interpolation impulse response receiving unit 11 may be realized by hardware, or may be realized by software such as a driver that drives a predetermined device.

  The interpolating Fourier transform unit 12 performs Fourier transform on the head impulse response received by the interpolating impulse response receiving unit 11 and accumulates the Fourier transform complex Fourier coefficient in the interpolating Fourier coefficient storage unit 13. This Fourier transform is already known as data processing, a head-related transfer function measurement method, and the like, and detailed description thereof will be omitted.

  In the interpolation Fourier coefficient storage unit 13, the complex Fourier of the head-related transfer function measured with respect to the subject's dummy at the azimuth for each first angle unit with reference to the reference azimuth, which is the reference azimuth. The coefficient is stored. The complex Fourier coefficient is converted by the interpolation Fourier transform unit 12. Storage in the interpolation Fourier coefficient storage unit 13 may be temporary storage in a RAM or the like, or may be long-term storage. The interpolation Fourier coefficient storage unit 13 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, an optical disk, or the like).

  The interpolation coefficient calculation unit 14 calculates an interpolation coefficient that is a coefficient for interpolating a complex Fourier coefficient corresponding to one or more interpolation azimuth angles. The interpolation coefficient calculation unit 14 stores the interpolation coefficient stored in the interpolation Fourier coefficient storage unit 13, and stores the complex Fourier coefficient corresponding to the interpolation azimuth in the interpolation Fourier coefficient storage unit 13. The calculation is performed such that interpolation is performed by multiplying a complex Fourier coefficient corresponding to three or more consecutive azimuth angles for each second angle unit with reference to the reference azimuth angle by an interpolation coefficient. Here, the interpolation azimuth is an angle that is an integer multiple (a positive integer multiple or not) of the first angle unit with the reference azimuth as a reference. However, the interpolation azimuth is an angle that is not an angle for each second angle unit with the reference azimuth as a reference. Further, as described above, the second angle unit is an angle unit that is M times the first angle unit (M is an integer of 2 or more). In addition, the interpolation coefficient calculation unit 14 calculates an interpolation coefficient using complex Fourier coefficients corresponding to three or more consecutive azimuths whose azimuths separated from the interpolation azimuth by a certain azimuth are the azimuths of the end points. It may be calculated. Further, the interpolation coefficient may be a coefficient that has the same value even when the interpolation azimuth corresponding to the interpolation coefficient changes by an integral multiple of the second angle unit. Strictly speaking, the interpolation azimuth corresponding to the interpolation coefficient is the interpolation azimuth corresponding to the complex Fourier coefficient interpolated using the interpolation coefficient.

The calculation of the interpolation coefficient will be described in more detail. First, the complex Fourier coefficient of the head-related transfer function corresponding to the azimuth angle θ and the frequency ω is c (ω, θ). The first angle unit is a. The second angle unit is k. Therefore, k = M × a. However, M is an integer of 2 or more. The reference azimuth angle is φ. Further, it is assumed that the azimuth angle θ _j (j = 0, 1,..., K−1) is an azimuth angle for each second angle unit with reference to the reference azimuth angle. That is, the following formula is assumed.

θ _j = φ + k × j
However, j is an integer from 0 to K-1. K is a value represented by the following equation.

When k is a factor of 360, 360 / k is an integer, so K = 360 / k. In the present embodiment, it is assumed that k is a factor of 360 for convenience of explanation. Also, d _i (ω) is defined as follows:

However, i = 1, 2,..., M−1. The azimuth angle in the complex Fourier coefficient of each element of d _i (ω) represented by the equation (1) is the interpolation azimuth angle. Therefore, each element of d _i (ω) is a complex Fourier coefficient corresponding to the interpolation azimuth angle.

Also, C (ω) is defined as follows:

However, N is a tap length and is an integer of 3 or more. _NS is an arbitrary integer. Therefore, each element of C (ω) represented by the equation (2) is a complex Fourier coefficient corresponding to the azimuth for each second angle unit “k” with the reference azimuth as a reference.

The interpolation coefficient is w _i (ω). Strictly speaking, w _i (ω) is a set of interpolation coefficients each having an interpolation coefficient as an element, but the set of interpolation coefficients is also simply referred to as an interpolation coefficient for convenience of explanation. To. d _i (ω), C (ω), and the interpolation coefficient w _i (ω) have the following relationship.

Therefore, the interpolation coefficient calculation unit 14 substitutes the complex Fourier coefficients stored in the interpolation Fourier coefficient storage unit 13 for d _i (ω) on the right side of equation (3) and C (ω) on the left side. The interpolation coefficient w _i (ω) can be calculated by solving simultaneous equations for the interpolation coefficient w _i (ω). Specifically, by using the least square method, an interpolation coefficient can be calculated as in the following equation. However, ^* indicates a complex conjugate.

Using the interpolation coefficient w _i (ω) calculated in this way, the complex Fourier coefficient of the head-related transfer function measured for each second angle unit for the subject to be described later is used as each element of C (ω). By substituting into, and calculating the left side of equation (3), each component of equation (1) can be calculated, and the head-related transfer function can be interpolated. This process will be described later.

As is clear from the above equations, the interpolation coefficient calculation unit 14 calculates the interpolation coefficient w _i (ω) for each i (i = 1, 2,..., M−1) and each frequency ω. It will be. Further, as shown in FIG. 2, the interpolation coefficient w _i (ω) for performing the interpolation of the complex Fourier coefficient of the interpolation azimuth angle θ _{0 + i × a} is the azimuth angle θ _0-Ns , θ _1-Ns,. , Θ _N-1-Ns is a coefficient to be multiplied to each of the complex Fourier coefficients. Also, the interpolation coefficient w _i (ω) for performing the interpolation of the complex Fourier coefficient of the interpolation azimuth angle θ _{1 + i × a} corresponds to the azimuth angles θ _{1 -Ns} , θ _{2 -Ns} ,..., Θ _N-Ns . A coefficient to be multiplied to each of the complex Fourier coefficients. Therefore, when “i” is the same, that is, when the relative positional relationship of the interpolation azimuth in the second angle unit range with respect to the reference azimuth is the same, the fixed azimuth “ Interpolation is performed using complex Fourier coefficients corresponding to three or more consecutive azimuths whose azimuths separated by “b” are the azimuths of the endpoints. Note that the value of “b” in FIG. 2 is different when “i” is different, that is, when the relative positional relationship of the interpolation azimuth in the second angle unit range is different. As for the interpolation coefficient w _i (ω), it can be seen that the same interpolation coefficient is used for each element of the equation (1) in which the interpolation azimuth angle changes by an integral multiple of the second angle unit.

The interpolation azimuth angle is preferably located between the minimum value and the maximum value of each azimuth angle of three or more complex Fourier coefficients used in the interpolation. For example, in FIG. 2, the interpolation azimuth angle θ _{0 + i × a} is preferably included between the azimuth angles θ _0-Ns and θ _N-1-Ns of the complex Fourier coefficient used in the interpolation. Further, the interpolation azimuth angle is more preferably located in the vicinity of the center between the minimum value and the maximum value of each azimuth angle of three or more complex Fourier coefficients used in the interpolation. For this purpose, for example, N _S may be set as follows. Needless to say, the relationship between the interpolation azimuth angle and each azimuth angle of three or more complex Fourier coefficients used in the interpolation may be other than that described in this paragraph. Therefore, here, the extrapolation is also referred to as interpolation.

The interpolation coefficient accumulation unit 15 accumulates the interpolation coefficient calculated by the interpolation coefficient calculation unit 14 in a predetermined recording medium. The recording medium is, for example, a semiconductor memory, an optical disk, a magnetic disk, or the like, and may be included in the interpolation coefficient storage unit 15 or may exist outside the interpolation coefficient storage unit 15. In addition, this recording medium may or may not temporarily store interpolation coefficients. Further, the interpolation coefficient storage unit 15 preferably stores the interpolation coefficient w _i (ω) so that the corresponding angle (i) and frequency (ω) can be known. That is, the interpolation coefficient storage unit 15 may store the interpolation coefficient in association with the angle or the frequency.

  The impulse response reception unit 16 receives a head impulse response (HRIR) corresponding to the azimuth angle for each second angle unit with respect to the reference azimuth angle with respect to the subject. The head impulse response received by the impulse response receiving unit 16 is an interpolation impulse response reception except that the subject's dummy is the subject and the azimuth angle for each first angle unit is the azimuth angle for each second angle unit. This is the same as the part 11, and the description thereof is omitted.

  The Fourier transform unit 17 performs Fourier transform on the head impulse response received by the impulse response reception unit 16 and accumulates the complex Fourier coefficient obtained by the Fourier transform in the Fourier coefficient storage unit 18. The Fourier transform unit 17 is also the same as the interpolation Fourier transform unit 12 except that the head impulse response to be processed is different, and the description thereof is omitted.

  The Fourier coefficient storage unit 18 stores the complex Fourier coefficient of the head-related transfer function measured at the azimuth for each second angle unit with respect to the reference azimuth as a reference. The complex Fourier coefficient is converted by the Fourier transform unit 17. The storage in the Fourier coefficient storage unit 18 may be temporary storage in a RAM or the like, or may be long-term storage. The Fourier coefficient storage unit 18 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, an optical disk, etc.).

The interpolating unit 19 stores the interpolating coefficient corresponding to one interpolation azimuth accumulated by the interpolating coefficient accumulating unit 15 and the complex used for calculating the interpolating coefficient stored in the Fourier coefficient storing unit 18. By performing interpolation using complex Fourier coefficients corresponding to three or more consecutive azimuths corresponding to the Fourier coefficients, a complex Fourier coefficient corresponding to the one interpolation azimuth is calculated. Specifically, the interpolation unit 19 reads out the complex Fourier coefficient stored in the Fourier coefficient storage unit 18 and substitutes it for each element of C (ω) in the above equation (2), thereby interpolating the coefficient storage unit 15 for interpolation. Is read out, substituted for each element of w _i (ω) in the above equation (3), and multiplied to calculate d _i (ω). D _i (ω) calculated in this way is the interpolated complex Fourier coefficient. As is clear from the above equation (1), each element of d _i (ω) corresponds to a different azimuth angle. Therefore, when it is desired to calculate a complex Fourier coefficient corresponding to a desired azimuth angle, it is only necessary to calculate the element of d _i (ω) corresponding to the desired azimuth angle. In that case, the complex Fourier coefficient corresponding to the desired azimuth angle is calculated by multiplying each element of the row of C (ω) corresponding to the desired component and each element of w _i (ω). (Interpolation). On the other hand, it goes without saying that the interpolation unit 19 may calculate complex Fourier coefficients corresponding to a plurality of azimuth angles for each first angle unit with reference to the reference azimuth angle. Further, the interpolation unit 19 calculates a complex Fourier coefficient for each frequency ω.

  The interpolation unit 19 may perform an interpolation process on an interpolation azimuth corresponding to a sound image localization azimuth received by an azimuth receiving unit 21 described later. The sound image localization azimuth is an azimuth for which sound image localization processing to be described later is desired. The “interpolation azimuth corresponding to the sound image localization azimuth” is, for example, when the sound image localization azimuth coincides with one of the azimuths for each first angle unit with reference to the reference azimuth. The orientation azimuth angle itself may be sufficient. On the other hand, when the sound image localization azimuth does not coincide with any of the azimuths for each first angle unit with reference to the reference azimuth, the “interpolation azimuth corresponding to the sound image localization azimuth” is the sound image localization azimuth. It may be an azimuth angle that is the nearest azimuth angle across the azimuth angle, and may be an azimuth angle for each first angle unit with reference to the reference azimuth angle (in this case, the interpolation azimuth angle corresponding to the sound image localization azimuth is 2).

  The interpolation Fourier coefficient storage unit 20 stores the complex Fourier coefficient calculated by the interpolation unit 19 in a predetermined recording medium. The recording medium is, for example, a semiconductor memory, an optical disk, a magnetic disk, or the like, and may be included in the interpolation Fourier coefficient storage unit 20 or may exist outside the interpolation Fourier coefficient storage unit 20. The recording medium may or may not temporarily store the interpolated complex Fourier coefficients. The interpolated Fourier coefficient accumulation unit 20 preferably accumulates the interpolated complex Fourier coefficients so that the corresponding azimuth angle and frequency can be understood. That is, the interpolation Fourier coefficient storage unit 20 may store the interpolated complex Fourier coefficient in association with the azimuth angle or the frequency.

  The azimuth acceptance unit 21 accepts a sound image localization azimuth that is an azimuth for performing sound image localization. The azimuth angle accepting unit 21 may accept a sound image localization azimuth angle input from an input device (for example, a keyboard, a mouse, a touch panel, etc.), and the sound image localization azimuth angle transmitted via a wired or wireless communication line. A sound image localization azimuth angle read from a predetermined recording medium (for example, an optical disk, a magnetic disk, or a semiconductor memory) may be received. Note that the azimuth angle acceptance unit 21 may or may not include a device (for example, a modem or a network card) for acceptance. Further, the azimuth acceptance unit 21 may be realized by hardware, or may be realized by software such as a driver that drives a predetermined device. Further, the sound image localization azimuth angle received by the azimuth angle receiving unit 21 may be temporarily stored in a recording medium (not shown).

  The sound signal receiving unit 22 receives a sound signal for performing sound image localization. For example, the sound signal receiving unit 22 may receive a sound signal input from a microphone, may receive a sound signal transmitted via a wired or wireless communication line, and may be a predetermined recording medium (for example, An audio signal read from an optical disk, a magnetic disk, a semiconductor memory, or the like may be received. Note that the sound signal reception unit 22 may or may not include a device (for example, a modem or a network card) for reception. Further, the sound signal receiving unit 22 may be realized by hardware, or may be realized by software such as a driver that drives a predetermined device. Further, the sound signal received by the sound signal receiving unit 22 may be temporarily stored in a recording medium (not shown).

  The sound image localization processing unit 23 uses the complex Fourier coefficients accumulated by the interpolation Fourier coefficient accumulation unit 20 and corresponding to the sound image localization azimuth to the sound signal received by the sound signal receiving unit 22. Performs sound image localization processing. Specifically, the sound image localization processing unit 23 adds a head-related transfer function indicated by a complex Fourier coefficient corresponding to the sound image localization azimuth received by the azimuth angle receiving unit 21 to the sound signal received by the sound signal receiving unit 22. Sound image localization processing is performed by convolution. Note that the sound image localization processing by the sound image localization processing unit 23 is already known and will not be described in detail. By performing this sound image localization process, a sound signal is generated in which the sound signal received by the sound signal receiving unit 22 is heard from the direction of the sound image localization azimuth received by the azimuth angle receiving unit 21. The “complex Fourier coefficient according to the sound image localization azimuth” is, for example, when the sound image localization azimuth coincides with any one of the first azimuth angles with respect to the reference azimuth. It may be a complex Fourier coefficient (strictly, a plurality of complex Fourier coefficients corresponding to each frequency) corresponding to an interpolation azimuth that is a sound image localization azimuth. On the other hand, when the sound image localization azimuth does not coincide with any of the azimuths for each of the first angle units with reference to the reference azimuth, the “complex Fourier coefficient according to the sound image localization azimuth” is the sound image localization azimuth. Complex Fourier coefficients respectively corresponding to complex azimuth angles that are azimuth angles of the first angle unit with reference to the reference azimuth angle as the nearest azimuth angle across the azimuth angle (in this case) The complex Fourier coefficient corresponding to the sound image localization azimuth becomes two sets). When there are two complex Fourier coefficients corresponding to the sound image localization azimuth, the complex Fourier coefficient of the sound image localization azimuth is calculated by performing interpolation using the two sets of complex Fourier coefficients, and the calculated Sound image localization processing may be performed using complex Fourier coefficients. For the interpolation using the two sets of complex Fourier coefficients, a known linear interpolation or the like can be used. For example, linear interpolation described later or corrected linear interpolation may be used.

  As described above, when the interpolation unit 19 performs only the interpolation processing related to the interpolation azimuth corresponding to the sound image localization azimuth received by the azimuth angle receiving unit 21, the sound image localization processing unit 23 performs interpolation. The sound image localization process may be performed using the complex Fourier coefficients accumulated by the Fourier coefficient accumulation unit 20. On the other hand, when the interpolation unit 19 is performing interpolation processing regarding a large number of interpolation azimuth angles regardless of the sound image localization azimuth angle, the sound image localization processing unit 23 includes the complex Fourier coefficients accumulated by the interpolation Fourier coefficient accumulation unit 20. The complex Fourier coefficient corresponding to the sound image localization azimuth received by the azimuth angle receiving unit 21 is selected, and the sound image localization processing is performed using the selected complex Fourier coefficient. Note that when the interpolation unit 19 performs an interpolation process regarding a large number of interpolation azimuth angles regardless of the sound image localization azimuth angle, the azimuth angle reception unit 21 receives a dynamically changing sound image localization azimuth angle. However, sound image localization corresponding to the sound image localization azimuth can be performed in real time or with a small delay.

  When the azimuth angle receiving unit 21 receives a sound image localization azimuth that is an azimuth angle for each second angle unit with reference to the reference azimuth angle, the sound image localization processing unit 23 passes through a route (not shown). Then, the complex Fourier coefficient corresponding to the sound image localization azimuth angle may be read from the Fourier coefficient storage unit 18 to perform the sound image localization process. Alternatively, the complex Fourier coefficient of the head-related transfer function corresponding to the azimuth angle for each second angle unit stored in the Fourier coefficient storage unit 18 with respect to the reference azimuth angle is also complex by the interpolation Fourier coefficient storage unit 20. You may make it memorize | store in the recording medium which accumulate | stores a Fourier coefficient. When the sound image localization processing unit 23 receives a sound image localization azimuth that is an azimuth for each second angle unit with reference to the reference azimuth, the sound image localization processing unit 23 calculates the complex Fourier coefficient stored in the recording medium. The sound image localization processing may be performed by using this.

  The output unit 24 outputs the sound signal that has been subjected to the sound image localization processing by the sound image localization processing unit 23. Here, this output may be, for example, audio output to headphones or earphones, transmission via a communication line to a predetermined device, accumulation in a recording medium, or delivery to another component. Good. Even if it is transmitted, stored in a recording medium, or delivered to another component, the sound signal after the sound image localization process is finally transmitted via headphones, earphones, etc. And being listened to by the user. The output unit 24 may or may not include a device that performs output. The output unit 24 may be realized by hardware, or may be realized by software such as a driver that drives these devices.

  The interpolation Fourier coefficient storage unit 13, the recording medium in which the interpolation coefficient storage unit 15 stores the interpolation coefficients, the Fourier coefficient storage unit 18, and the interpolation Fourier coefficient storage unit 20 store the interpolated complex Fourier coefficients. Of the recording media to be recorded, any two or more storage units and recording media may be realized by the same recording medium, or may be realized by separate recording media. In the former case, for example, the region storing the complex Fourier coefficient related to the subject's dummy is the interpolation Fourier coefficient storage unit 13, and the region storing the complex Fourier coefficient related to the subject is the Fourier coefficient storage unit 18. .

  Next, operations of the head-related transfer function interpolation coefficient calculation device 1 and the sound image localization device 2 according to the present embodiment will be described with reference to the flowchart of FIG. In this flowchart, the case where the interpolation unit 19 performs an interpolation process related to the interpolation azimuth corresponding to the sound image localization azimuth received by the azimuth acceptance unit 21 will be described.

  (Step S101) The interpolation impulse response receiving unit 11 determines whether a head impulse response to the dummy of the subject has been received. If accepted, the process proceeds to step S102; otherwise, the process proceeds to step S104.

  (Step S <b> 102) The interpolation Fourier transform unit 12 performs a Fourier transform on the head impulse response received by the interpolation impulse response reception unit 11.

  (Step S <b> 103) The interpolation Fourier transform unit 12 stores the complex Fourier coefficients after the Fourier transform in the interpolation Fourier coefficient storage unit 13. Then, the process returns to step S101.

  (Step S104) The impulse response reception unit 16 determines whether a head impulse response to the subject has been received. If it is accepted, the process proceeds to step S105, and if not, the process proceeds to step S107.

  (Step S105) The Fourier transform unit 17 performs Fourier transform on the head impulse response received by the impulse response reception unit 16.

  (Step S106) The Fourier transform unit 17 stores the complex Fourier coefficients after the Fourier transform in the Fourier coefficient storage unit 18. Then, the process returns to step S101.

  (Step S107) The interpolation coefficient calculation unit 14 determines whether or not to perform processing for calculating the interpolation coefficient. And when performing the process, it progresses to step S108, and when that is not right, it progresses to step S109. The interpolation coefficient calculation unit 14 may determine to calculate the interpolation coefficient when receiving an instruction to calculate the interpolation coefficient from the user, for example. If sufficient complex Fourier coefficients are stored to calculate the interpolation coefficient, it may be determined to calculate the interpolation coefficient, or may be determined to calculate the interpolation coefficient at other timings. Good.

  (Step S108) The interpolation coefficient calculation unit 14 and the like perform processing for calculating an interpolation coefficient. Then, the process returns to step S101. Details of this processing will be described later with reference to the flowchart of FIG.

  (Step S109) The azimuth acceptance unit 21 determines whether a sound image localization azimuth has been accepted. And when it receives, it progresses to step S110, and when that is not right, it returns to step S101.

  (Step S110) The interpolation unit 19 performs an interpolation process related to the interpolation azimuth corresponding to the sound image localization azimuth accepted by the azimuth acceptance unit 21. That is, the interpolation unit 19 calculates a complex Fourier coefficient corresponding to the interpolation azimuth using the interpolation coefficient corresponding to the interpolation azimuth and the complex Fourier coefficient stored in the Fourier coefficient storage unit 18. . The calculation of the complex Fourier coefficient is performed for each frequency.

  (Step S111) The interpolation Fourier coefficient storage unit 20 stores the complex Fourier coefficients interpolated by the interpolation unit 19 in a predetermined recording medium.

  (Step S112) The sound signal receiving unit 22 determines whether a sound signal has been received. If a sound signal is received, the process proceeds to step S113. If not, the process proceeds to step S115.

  (Step S113) The sound image localization processing unit 23 performs sound image localization processing on the sound signal received by the sound signal receiving unit 22 using a complex Fourier coefficient corresponding to the sound image localization azimuth received by the azimuth angle receiving unit 21. Do.

  (Step S114) The output unit 24 outputs the sound signal that has been subjected to the sound image localization processing by the sound image localization processing unit 23. Then, the process returns to step S112.

  (Step S115) The sound image localization processing unit 23 determines whether or not to end the sound image localization processing. If the process is to end, the process returns to step S101; otherwise, the process returns to step S112. Note that the sound image localization processing unit 23 may determine that the sound image localization processing is to be ended when, for example, an instruction to end the sound image localization processing from the user is received, When the sound signal is not received by the sound signal receiving unit 22, it may be determined that the sound image localization process is finished, or it may be determined that the sound image localization process is finished at other timing.

  In the flowchart of FIG. 3, the process ends when the power is turned off or the process is terminated. Further, in place of the processing shown in the flowchart of FIG. 3, interpolation processing relating to all possible interpolation azimuth angles is performed in advance, and complex Fourier coefficients corresponding to the sound image localization azimuth angle received by the azimuth angle reception unit 21 are read out. Interpolation may be performed. In that case, in step S113, the complex Fourier coefficient corresponding to the sound image localization azimuth accepted by the azimuth acceptance unit 21 is read out and used.

  FIG. 4 is a flowchart showing details of processing (step S108) such as calculation of an interpolation coefficient in the flowchart of FIG.

  (Step S201) The interpolation coefficient calculation unit 14 sets the counter i to 1.

  (Step S202) The interpolation coefficient calculation unit 14 sets the frequency ω to an initial value.

(Step S <b> 203) The interpolation coefficient calculation unit 14 reads out a complex Fourier coefficient that is each element of d _i (ω) from the interpolation Fourier coefficient storage unit 13.

  (Step S <b> 204) The interpolation coefficient calculation unit 14 reads out a complex Fourier coefficient that is each element of C (ω) from the interpolation Fourier coefficient storage unit 13.

(Step S205) The interpolation coefficient calculation unit 14 calculates the interpolation coefficient w _i (ω) using the complex Fourier coefficients read out in steps S203 and S204. This calculation may be performed using, for example, the aforementioned equation (4).

(Step S206) The interpolation coefficient accumulation unit 15 accumulates the interpolation coefficient w _i (ω) calculated by the interpolation coefficient calculation unit 14 in a recording medium.

  (Step S207) The interpolation coefficient calculation unit 14 updates the frequency ω to the next frequency. This update may be performed, for example, by adding or subtracting a certain value to the current frequency ω.

  (Step S208) The interpolation coefficient calculation unit 14 determines whether or not the updated frequency ω exists. And when it exists, it returns to step S203, and when it does not exist, it progresses to step S209. Whether or not the updated frequency ω exists may be determined by whether or not the complex Fourier coefficient stored in the interpolation Fourier coefficient storage unit 13 includes the coefficient of the frequency ω, for example.

  (Step S209) The interpolation coefficient calculation unit 14 increments the counter i by 1.

  (Step S210) The interpolation coefficient calculation unit 14 determines whether the counter i is equal to or greater than M. If it is greater than or equal to M, interpolation coefficients have been calculated for all azimuth angles, so the process returns to the flowchart of FIG. 3, and if not, the process returns to step S202.

  Next, the reason for performing the interpolation processing using the above-described equations (1) to (4) will be briefly described. First, the inventors measured the head-related transfer function at azimuths of 1 ° in order to analyze spatial frequency characteristics. This measurement was performed using an anechoic chamber shown in FIG. The anechoic chamber is provided with a head-related transfer function measurement system. An arcuate arm provided with a speaker is provided so as to be able to turn in the horizontal direction. The azimuth angle and elevation angle of the speaker are automatically controlled by a personal computer (PC). In order to obtain a head impulse response, an Optimized Aoki's Time-Stretched Pulse (OATSP) signal having a length of 16384 points is used with a 16-bit resolution and a sampling rate of 48 kHz. Moreover, the test subject's dummy used as the measurement object was the head and torso (HATS: B & K 4128). Then, a head impulse response was specified with a resolution of 16 bits and a sampling rate of 48 kHz using a microphone (Sound Professionals SP-TFB-2) provided at the entrance of the dummy both ears. The OATSP signal was output 40 times for each speaker position, and the head impulse response was averaged to improve the S / N ratio. FIG. 6 shows a head-related transfer function obtained by Fourier-transforming the head impulse response measured at the position of the left ear. In FIG. 6, the forward direction of the dummy corresponds to the azimuth angle “0 °”. The azimuth angle increases clockwise. The head-related transfer function measured at the position of the right ear also had a similar target pattern. In FIG. 6, there is no data with an azimuth angle between 161 ° and 164 ° because the arcuate arm cannot move to that azimuth range due to mechanical limitations.

  FIG. 7 is a diagram obtained by further Fourier transforming the power spectrum of FIG. 6 in the azimuth angle direction. In FIG. 7, the horizontal axis represents each frequency in the head-related transfer function. The vertical axis corresponds to the periodic repetition of the power spectrum level along the azimuth direction. For example, a repetition period of 5 ° indicates that the wavelength is close to a periodicity corresponding to 5 °. Before performing the Fourier transform on the power spectrum of FIG. 6, by moving the region where the missing azimuth angle is between 161 ° and 164 ° to both ends of the data, The area was removed. This is done in order to prevent the periodicity of the head-related transfer function from becoming unclear due to loss of data continuity. In FIG. 7, the periodicity shown by the white part is remarkable. Its periodicity is up to about 15 kHz, and the repetition period is about 6-7 °. According to Nyquist's theorem, in order to avoid the influence of aliasing due to this periodicity, it is preferable to have a resolution of 3 ° or less in order to measure the head-related transfer function.

  This periodicity in the azimuth direction may be due to interference of sound diffracted around the head. This is because the periodicity is remarkable on the opposite side of the speaker. Therefore, it can be predicted that the same periodicity is observed for heads of the same size. In fact, this also applies to the measurement of the head-related transfer function using a dummy head produced based on precise estimation of the physical characteristics of a person using MRI data.

  Thus, since there is periodicity in the azimuth direction, it is expected that better interpolation can be realized by using the head-related transfer functions measured at a plurality of positions in the head-related transfer function interpolation. it can. In particular, rather than performing interpolation using head-related transfer functions measured at two positions as in conventional linear interpolation, interpolation is performed using head-related transfer functions measured at three or more positions. Therefore, it is considered that the interpolation becomes more accurate. In FIG. 7, the periodicity is observed for the power spectrum, but this periodicity is considered to apply to the complex spectrum of the head-related transfer function, and the above-described equations (1) to (4) are used. This formulates the interpolation processing.

  Next, operations of the head-related transfer function interpolation coefficient calculation device 1 and the sound image localization device 2 according to the present embodiment will be described using specific examples. In this specific example, the reference azimuth angle is assumed to be 0 ° in front of the dummy or the subject.

  First, a case where the head-related transfer function for the subject's dummy is measured will be described. In the anechoic chamber shown in FIG. 5, the head impulse response is measured for each azimuth angle of 1 ° using a dummy (HATS: B & K 4128) composed of a head and a torso. Therefore, in this example, the first angle unit is 1 °. In this specific example, it is assumed that a head impulse response of approximately 360 ° is measured for each azimuth angle except for an angle that cannot be measured. Also in the measurement of the head impulse response, the position of each speaker is measured 40 times using the OATSP signal, and an average value is taken to improve the S / N ratio. When the head impulse response thus measured is input to the sound image localization apparatus 2, the head impulse response is received by the interpolation impulse response receiving unit 11 (step S101). Then, the Fourier transform is performed by the interpolation Fourier transform unit 12 (step S102), and the complex Fourier coefficients after the Fourier transform are accumulated in the interpolation Fourier coefficient storage unit 13 (step S103).

  Next, a case where the head-related transfer function for the subject is measured will be described. In the anechoic chamber shown in FIG. 5 described above, the head impulse response is measured for each azimuth of 5 ° with respect to the subject. Therefore, in this specific example, the second angle unit is 5 °. In this specific example, it is assumed that a head impulse response of approximately 360 ° is measured for each azimuth angle except for an angle that cannot be measured. Also, in the measurement of the head impulse response for this subject, the measurement is performed 40 times for each speaker position using the OATSP signal in the same manner as the measurement of the dummy head impulse response described above, and an average value is obtained. When the head impulse response thus measured is input to the sound image localization apparatus 2, the head impulse response is received by the impulse response receiving unit 16 (step S104). Then, the Fourier transform is performed by the Fourier transform unit 17 (step S105), and the complex Fourier coefficients after the Fourier transform are accumulated in the Fourier coefficient storage unit 18 (step S106).

After that, when it is detected that the complex Fourier coefficient of the dummy head-related transfer function and the complex Fourier coefficient of the subject's head-related transfer function are accumulated for all azimuth angles, the interpolation coefficient calculating unit 14 performs the interpolation coefficient Is calculated (step S107), and an interpolation coefficient is calculated (step S108). Specifically, as described above, interpolation coefficients w _i (ω) are calculated and accumulated for all i (in this example, i = 1 to 4) and all frequencies ω. (Steps S201 to S210).

  In calculating the interpolation coefficient, the interpolation coefficient is calculated without using the complex Fourier coefficient of the head-related transfer function having an azimuth angle in a range where the head-related transfer function cannot be measured. As is apparent from the above equation, the number of equations (K) in the simultaneous equations is generally larger than the number of coefficients to be obtained (N). The interpolation coefficient can be calculated without using the complex Fourier coefficient of the head related transfer function. In the measurement of the head-related transfer function, when there is no azimuth range that cannot be measured, it is not necessary to consider this.

Next, it is assumed that the user inputs a sound image localization azimuth “32 °” by operating an input device such as a mouse or a keyboard, and the sound image localization azimuth is received by the azimuth acceptance unit 21 (step S109). . Then, the interpolation part 19 performs the interpolation regarding the angle (step S110). Specifically, in the above-described notation, in this specific example, φ = 0 °, k = 5 °, and a = 1 °, so that 32 ° = θ ₆ + 2 × a. However, θ ₆ = 30 °. Therefore, the seventh element of d ₂ (ω) may be calculated. Therefore, the interpolation unit 19 reads out the complex Fourier coefficient related to each column of the seventh row of C (ω) from the Fourier coefficient storage unit 18, and the interpolation coefficient storage unit 15 interpolates w ₂ (ω) as the interpolation coefficient. C (ω, θ ₆ + 2 × a = 32 °) is calculated by reading from the recording medium in which the coefficients are stored and multiplying them. The interpolation unit 19 performs this calculation for all frequencies ω.

The interpolated complex Fourier coefficient c (ω, θ ₆ + 2 × a = 32 °) is stored in a predetermined recording medium by the interpolation Fourier coefficient storage unit 20 (step S111). Thereafter, when the sound signal receiving unit 22 receives the sound signal (step S112), the sound image localization processing unit 23 calculates the head-related transfer function of the interpolated Fourier coefficient c (ω, 32 °) accumulated by the interpolated Fourier coefficient accumulating unit 20. Sound image localization processing is performed by convolution (step S113). Then, the output unit 24 outputs the sound signal subjected to the sound image localization processing to headphones or the like (step S114). Then, the sound image localization corresponding to the desired sound image localization azimuth is executed by repeating the reception of the sound signal, the sound image localization process, and the output of the sound signal subjected to the sound image localization process. .

  In this specific example, the case where the unit of the azimuth angle for measuring the dummy head-related transfer function, that is, the first angle unit is 1 ° has been described, but this need not be the case. However, it is preferable to measure in the first angle unit that can measure the periodicity depending on the azimuth angle. Further, in this specific example, the case where the unit of the azimuth angle for measuring the head-related transfer function of the subject, that is, the second angle unit is five times the first angle unit has been described. Good. However, it is preferable to use the second angle unit within a range that can be appropriately interpolated.

  Next, a comparison between the interpolation method according to the present embodiment and another conventional interpolation method will be described. As conventional interpolation methods, (1) linear interpolation and (2) corrected linear interpolation are used. Each will be briefly described.

(1) Linear interpolation A method of linearly interpolating the head impulse response h (n, θ) at the azimuth angle θ using the head impulse responses measured at the azimuth angles θ _p and θ _q is obtained by the following equation. . Note that n indicates the sample time.

However, (alpha) is as following Formula.

(2) Corrected linear interpolation In this interpolation method, first, the initial delay of the head impulse response depending on the azimuth angle is removed. Then, the removed initial delay and the remaining part are interpolated separately, and both are connected again after the interpolation. This makes it possible to perform more appropriate interpolation than in the case of simple linear interpolation. For details of the corrected linear interpolation, refer to the following document.

  Literature: M.M. Matsumoto, S .; Yamanaka, M .; Toyama, H .; Nomura, “Effect of arrival time correction on the accuracy of binaural impulse response interpolation”, J. Am. Audio Eng. Soc. , Vol. 52, no. 1/2, p. 56-61, January, February 2004

In this comparison, the performance of each interpolation method is evaluated by the least square error (MSE) expressed by the following equation.

Here, j = L and R are subscripts indicating the left ear and the right ear. S is the number of azimuths at which the head impulse response is measured. Here, S = 284 is obtained because the range that cannot be measured out of 360 ° and the azimuth angle every 5 ° used for interpolation are removed. Further, h _j (n, θ) is a head impulse response to the subject corresponding to the azimuth angle θ, and a hat (^) of h _j (n, θ) is an interpolated head impulse response. Therefore, in this evaluation, the head-related transfer function for each first angle unit was also measured for the subject in order to calculate the error.

  In this comparison, in the interpolation method according to the present embodiment, as described in the present embodiment, interpolation using an interpolation coefficient is performed. On the other hand, in (1) linear interpolation and (2) corrected linear interpolation to be compared, interpolation was performed by the above-described method. Further, the complex Fourier coefficients of the subject to be interpolated are the same for all. Then, the performance of each interpolation method was evaluated by taking the difference between the measurement result of the subject and the interpolated result, as in the above-described least square error equation.

  FIG. 8 is a graph showing a comparison result of the performance of the interpolation method. In FIG. 8, the tap length indicates the number of complex Fourier coefficients used for interpolation. Subject 1 and subject 2 are two types of dummy heads produced using real human MRI data. The subject 3 is the same HATS (B & K 4128) as the subject's dummy. Therefore, in the case of the test subject 3, both the test subject dummy and the test subject are the same. Moreover, in this graph, the horizontal axis direction of each plotted point is intentionally shifted a little for ease of viewing. Also, this comparison result shows only when the tap length is an even number. When the tap length is “2”, there is almost no difference between the conventional interpolation method and the interpolation method of the head transfer function interpolation coefficient calculation apparatus 1 according to the present embodiment. On the other hand, when the tap length exceeds 2, it can be seen that the interpolation method of the head transfer function interpolation coefficient calculation apparatus 1 according to the present embodiment is more accurate than the conventional interpolation method. Therefore, when the tap length is 3 or more, that is, when the number of complex Fourier coefficients used in the interpolation is 3 or more, it can be understood that interpolation can be performed with higher accuracy than the conventional interpolation method. .

  As described above, according to the head transfer function interpolation coefficient calculation apparatus 1 according to the present embodiment, the subject's head transfer function measured in a coarse azimuth unit is measured in a fine azimuth unit. Interpolation can be performed using the interpolation coefficient calculated from the dummy head-related transfer function. Therefore, in the measurement of the head-related transfer function for the subject, the time for restraining the subject is shorter than in the case of measuring in units of fine azimuth, and the burden on the subject is reduced. It is also possible to measure the subject's head-related transfer function in a short time. Further, the sound image localization can be performed in the sound image localization apparatus 2 by using the head-related transfer function thus interpolated.

  In addition, when performing only interpolation according to the sound image localization azimuth received by the azimuth acceptance unit 21, a necessary and sufficient range of head-related transfer functions is held, and the heads corresponding to all azimuths. Compared to the case where the transfer function is stored from the beginning, the necessary memory amount can be reduced.

  In the present embodiment, the case where the complex Fourier coefficients stored in the interpolation Fourier coefficient storage unit 13 are accumulated by the interpolation Fourier transform unit 12 has been described, but this need not be the case. For example, a complex Fourier coefficient obtained by Fourier transforming the head impulse response to the subject's dummy in another apparatus or the like may be stored in the interpolation Fourier coefficient storage unit 13. In this case, the process of storing the complex Fourier coefficient in the interpolation Fourier coefficient storage unit 13 does not matter. For example, the complex Fourier coefficient may be stored in the interpolation Fourier coefficient storage unit 13 via a recording medium, and the complex Fourier coefficient transmitted via a communication line or the like may be stored in the interpolation Fourier coefficient storage unit 13. Alternatively, the complex Fourier coefficients input via the input device may be stored in the interpolation Fourier coefficient storage unit 13.

  In the present embodiment, the case where the complex Fourier coefficients stored in the Fourier coefficient storage unit 18 are accumulated by the Fourier transform unit 17 has been described, but this need not be the case. For example, a complex Fourier coefficient obtained by subjecting the head impulse response to the subject to Fourier transform in another device or the like may be stored in the Fourier coefficient storage unit 18. In that case, the process in which the complex Fourier coefficient is stored in the Fourier coefficient storage unit 18 does not matter. For example, the complex Fourier coefficient may be stored in the Fourier coefficient storage unit 18 via a recording medium, and the complex Fourier coefficient transmitted via a communication line or the like may be stored in the Fourier coefficient storage unit 18. Alternatively, the complex Fourier coefficient input via the input device may be stored in the Fourier coefficient storage unit 18.

  Moreover, although the case where the sound image localization device 2 includes the head transfer function interpolation coefficient calculation device 1 has been described in the present embodiment, this need not be the case. Only the head transfer function interpolation coefficient calculation device 1 may be used separately from the sound image localization device 2. In that case, the head-related transfer function interpolation coefficient calculation device 1 outputs an output unit that outputs the complex Fourier coefficient stored in the interpolation Fourier coefficient storage unit 20 or the complex Fourier coefficient stored in the Fourier coefficient storage unit 18 ( (Not shown) may be provided. Here, this output may be transmitted via a communication line to a predetermined device, stored in a recording medium, or delivered to another component.

  Further, in this embodiment, the case where the interpolation processing is performed in the head-related transfer function interpolation coefficient calculation device 1 has been described. However, the head-related transfer function interpolation coefficient calculation device 1 performs the processing up to the calculation of the interpolation coefficient. And the interpolation processing using the interpolation coefficient may be performed in another apparatus. In that case, the head-related transfer function interpolation coefficient calculation device 1 does not include the impulse response reception unit 16, the Fourier transform unit 17, the Fourier coefficient storage unit 18, the interpolation unit 19, and the interpolation Fourier coefficient storage unit 20. Also good. Further, an output unit (not shown) that outputs the interpolation coefficient accumulated by the interpolation coefficient accumulation unit 15 may be provided. Here, this output may be transmitted via a communication line to a predetermined device, stored in a recording medium, or delivered to another component.

  In the present embodiment, for example, as shown in FIG. 5, a case where the head-related transfer function is measured by changing the azimuth at a position where the elevation angle is approximately 0 ° (horizontal position) will be described. But it doesn't have to be. A head-related transfer function measured at a position where the elevation angle is other than 0 ° may be measured, and the interpolation processing described in the present embodiment may be performed on the measured head-related transfer function. However, it is preferable that the elevation angle when measuring the head-related transfer function for the subject's dummy and the elevation angle when measuring the head-related transfer function for the subject are the same.

  Further, as described above, since it can be predicted that the same periodicity is observed for the heads of the same size, for example, the interpolation coefficient is calculated for the dummy of the subject of various sizes, and the subject When performing the interpolation of the head related transfer function, the interpolation may be performed using a dummy interpolation coefficient closest to the size of the subject's head. By doing so, it is considered that more accurate interpolation can be realized.

  In the present embodiment, the case where the subject is a human is mainly described, but this need not be the case. The subject may be a human or a dummy as long as it is a target for measuring the head-related transfer function. When the subject is a dummy, the head-related transfer function as a subject can be interpolated by measuring the head-related transfer function with another dummy in advance and calculating the interpolation coefficient. There is an advantage that the measurement time of the head-related transfer function for the dummy as the subject can be obtained.

  In the present embodiment, the complex Fourier coefficients stored in the interpolation Fourier coefficient storage unit 13 and the Fourier coefficient storage unit 18 correspond to the calculation of the interpolation coefficient and the azimuth of the range necessary for the interpolation. Any complex Fourier coefficient may be used. Therefore, the complex Fourier coefficients stored in those storage units do not have to correspond to all azimuth angles. For example, when the range of the interpolation azimuth is determined, it is sufficient that at least the storage unit stores the complex Fourier coefficients necessary for performing the interpolation according to the interpolation azimuth. For example, the complex Fourier coefficient stored in the interpolation Fourier coefficient storage unit 13 is a complex of the head-related transfer function corresponding to a continuous azimuth for each first angle unit at a position separated by about 180 ° from the reference azimuth. It may be a Fourier coefficient.

  Further, in the present embodiment, the case where k is a factor of 360 ° has been described, but as described above, this need not be the case. However, if k is not a factor of 360 °, even if the head-related transfer function is measured for each second angle unit, there is a part of the azimuth angle interval that is not for each second angle unit. End up. For example, if φ = 0 ° and k = 7 °, the head-related transfer function of the subject is measured at 0 ° and 357 °, but the azimuth interval (= 3 °) is This is different from the second angular unit (= 7 °). In such a case, it is preferable to perform the interpolation so as not to span the range of the azimuth interval different from the second angle unit. That is, in the above equation (2), the range of the azimuth interval different from the second angle unit between the minimum value and the maximum value of the azimuth angle of the complex Fourier coefficient of each column (in the case of the above example) Is preferably interpolated so that 357 ° to 360 ° (= 0 °) is not included.

  Further, in the present embodiment, when the interpolation coefficient calculation unit 14 calculates the interpolation coefficient, the interpolation coefficient, which is a coefficient for interpolating the complex Fourier coefficient corresponding to one or more interpolation azimuth angles, is interpolated. The second angular unit based on the reference azimuth angle, which is stored in the Fourier coefficient storage section 13 for the complex Fourier coefficient corresponding to the interpolation azimuth angle and stored in the Fourier coefficient storage section 13 for interpolation. The case where calculation is performed so as to be interpolated by multiplying a complex Fourier coefficient corresponding to three or more consecutive azimuth angles by an interpolation coefficient has been described. However, the three or more azimuth angles are not continuous. Also good. That is, the interpolation coefficient calculation unit 14 determines that the complex Fourier coefficient corresponding to the interpolation azimuth angle stored in the interpolation Fourier coefficient storage unit 13 is 3 for each second angle unit with reference to the reference azimuth angle. The interpolation coefficient may be calculated so as to be interpolated by multiplying the complex Fourier coefficient corresponding to the above azimuth by the interpolation coefficient. In other words, in the above equation (2), the absolute value of the difference between the azimuth angles of adjacent complex Fourier coefficients in a certain row is a value obtained by multiplying the second angle unit by an integer of 2 or more. It will be good. However, even in that case, in Equation (2), the difference in azimuth between the complex Fourier coefficient of the i-th column and the complex Fourier coefficient of the (i + 1) -th column in a certain row is the same in any other row. It is assumed that the difference in azimuth between the complex Fourier coefficient in the i-th column and the complex Fourier coefficient in the i + 1-th column is the same. However, i is an arbitrary integer from 1 to N-1. As described above, N is a tap length.

  In the above embodiment, the head transfer function interpolation coefficient calculation device 1 and the sound image localization device 2 are stand-alone. However, the head transfer function interpolation coefficient calculation device 1 and the sound image localization device 2 are It may be a stand-alone device or a server device in a server / client system. In the latter case, the output unit or the reception unit may receive input or output information via a communication line.

  In the above embodiment, each process or each function may be realized by centralized processing by a single device or a single system, or may be distributedly processed by a plurality of devices or a plurality of systems. It may be realized by doing.

  In the above embodiment, information related to processing executed by each component, for example, information received, acquired, selected, generated, transmitted, or received by each component In addition, information such as threshold values, mathematical formulas, addresses, etc. used by each component in processing is retained temporarily or over a long period of time on a recording medium (not shown) even when not explicitly stated in the above description. It may be. Further, the storage of information in the recording medium (not shown) may be performed by each component or a storage unit (not shown). Further, reading of information from the recording medium (not shown) may be performed by each component or a reading unit (not shown).

  In the above embodiment, when information used by each component, for example, information such as a threshold value, an address, and various setting values used by each component may be changed by the user Even if it is not specified in the above description, the user may be able to change the information as appropriate, or it may not be. If the information can be changed by the user, the change is realized by, for example, a not-shown receiving unit that receives a change instruction from the user and a changing unit (not shown) that changes the information in accordance with the change instruction. May be. The change instruction received by the receiving unit (not shown) may be received from an input device, information received via a communication line, or information read from a predetermined recording medium, for example. .

  In the above embodiment, when two or more components included in the head transfer function interpolation coefficient calculation device 1 and the sound image localization device 2 have communication devices, input devices, and the like, the two or more components are physical. May have a single device or separate devices. For example, the interpolation impulse response reception unit 11 and the impulse response reception unit 16 may be physically the same.

  In the above embodiment, each component may be configured by dedicated hardware, or a component that can be realized by software may be realized by executing a program. For example, each component can be realized by a program execution unit such as a CPU reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory. In addition, the software which implement | achieves the coefficient calculation apparatus 1 for head related transfer function interpolation in the said embodiment is the following programs. In other words, this program is an angle that is an integral multiple of the first angle unit with reference to the reference azimuth, which is a reference azimuth for the subject's dummy, and is based on the reference azimuth. Head related transfer function corresponding to one or more interpolated azimuth angles that are not angles for each second angle unit that is M times the first angle unit (M is an integer of 2 or more). A head-related transfer function measured at an azimuth angle for each of the first angle units with respect to the subject's dummy as an interpolation coefficient, which is a coefficient for interpolating the complex Fourier coefficient of The complex Fourier coefficient corresponding to the interpolation azimuth angle stored in the interpolation Fourier coefficient storage unit in which the complex Fourier coefficient is stored is stored in the interpolation Fourier coefficient storage unit, and the reference azimuth angle is Standard An interpolation coefficient calculation unit for calculating so as to be interpolated by multiplying a complex Fourier coefficient corresponding to three or more azimuth angles for each second angle unit by the interpolation coefficient, and the interpolation coefficient calculation unit It is a program for functioning as an interpolation coefficient storage unit that stores calculated interpolation coefficients.

  Further, in this program, the interpolation coefficient corresponding to one interpolation azimuth angle accumulated by the interpolation coefficient accumulation unit and the second angle unit with respect to the reference azimuth angle are stored for the subject. Stored in the Fourier coefficient storage unit in which the complex Fourier coefficient of the head-related transfer function measured at the azimuth angle is stored, and at three or more azimuth angles corresponding to the complex Fourier coefficient used in the calculation of the interpolation coefficient. An interpolation unit that calculates a complex Fourier coefficient corresponding to the one interpolation azimuth by performing interpolation using the corresponding complex Fourier coefficient, and an interpolation Fourier coefficient storage unit that stores the complex Fourier coefficient calculated by the interpolation unit Further, a program for causing a computer to function may be used.

  Moreover, the software which implement | achieves the sound image localization apparatus 2 in the said embodiment is the following programs. That is, this program causes the computer to receive the sound image localization azimuth, which is an azimuth angle for performing sound image localization, the interpolation coefficient calculation unit, the interpolation coefficient storage unit, the interpolation unit, the interpolation Fourier coefficient storage unit, A sound signal receiving unit for receiving a sound signal for performing sound image localization, a complex Fourier coefficient stored by the interpolation Fourier coefficient storage unit, and using the complex Fourier coefficient corresponding to the sound image localization azimuth, the sound signal It is a program for functioning as a sound image localization processing unit that performs sound image localization processing on a sound signal received by the reception unit, and an output unit that outputs a sound signal subjected to sound image localization processing by the sound image localization processing unit.

  In the program, the functions realized by the program do not include functions that can be realized only by hardware. For example, functions that can be realized only by hardware such as a modem and an interface card in a reception unit that receives information and an output unit that outputs information are not included in at least the functions realized by the program.

  Further, this program may be executed by being downloaded from a server or the like, and a program recorded on a predetermined recording medium (for example, an optical disc such as a CD-ROM, a magnetic disc, a semiconductor memory, etc.) is read out. May be executed. Further, this program may be used as a program constituting a program product.

  Further, the computer that executes this program may be singular or plural. That is, centralized processing may be performed, or distributed processing may be performed.

  FIG. 9 is a schematic diagram showing an example of the appearance of a computer that executes the program and realizes the head transfer function interpolation coefficient calculation device 1 and the sound image localization device 2 according to the embodiment. The above-described embodiment can be realized by computer hardware and a computer program executed on the computer hardware.

  9, a computer system 900 includes a computer 901 including a CD-ROM (Compact Disk Read Only Memory) drive 905 and an FD (Floppy (registered trademark) Disk) drive 906, a keyboard 902, a mouse 903, a monitor 904, and the like. Is provided.

  FIG. 10 is a diagram showing an internal configuration of the computer system 900. In FIG. 10, in addition to the CD-ROM drive 905 and the FD drive 906, a computer 901 is connected to an MPU (Micro Processing Unit) 911, a ROM 912 for storing a program such as a bootup program, and the MPU 911. A RAM (Random Access Memory) 913 that temporarily stores program instructions and provides a temporary storage space, a hard disk 914 that stores application programs, system programs, and data, and an MPU 911 and a ROM 912 are interconnected. And a bus 915. The computer 901 may include a network card (not shown) that provides connection to the LAN.

  A program for causing the computer system 900 to execute the functions of the head transfer function interpolation coefficient calculation device 1 and the sound image localization device 2 according to the above-described embodiment is stored in the CD-ROM 921 or the FD 922, and the CD-ROM drive 905, Alternatively, it may be inserted into the FD drive 906 and transferred to the hard disk 914. Instead, the program may be transmitted to the computer 901 via a network (not shown) and stored in the hard disk 914. The program is loaded into the RAM 913 when executed. The program may be loaded directly from the CD-ROM 921, the FD 922, or the network.

  The program does not necessarily include an operating system (OS) or a third-party program that causes the computer 901 to execute the functions of the head-related transfer function interpolation coefficient calculation device 1 and the sound image localization device 2 according to the above embodiment. Also good. The program may include only a part of an instruction that calls an appropriate function (module) in a controlled manner and obtains a desired result. How the computer system 900 operates is well known and will not be described in detail.

  Further, the present invention is not limited to the above-described embodiments, and various modifications are possible, and it goes without saying that these are also included in the scope of the present invention.

  As described above, according to the head transfer function interpolation coefficient calculation apparatus and the like according to the present invention, an effect that the head transfer function can be interpolated can be obtained, which is useful, for example, in an apparatus for performing sound image localization.

DESCRIPTION OF SYMBOLS 1 Head transfer function interpolation coefficient calculation device 2 Sound image localization device 11 Interpolation impulse response reception unit 12 Interpolation Fourier transform unit 13 Interpolation Fourier coefficient storage unit 14 Interpolation coefficient calculation unit 15 Interpolation coefficient storage unit 16 Impulse response reception Unit 17 Fourier transform unit 18 Fourier coefficient storage unit 19 interpolation unit 20 interpolation Fourier coefficient storage unit 21 azimuth angle reception unit 22 sound signal reception unit 23 sound image localization processing unit 24 output unit

Claims (9)

Interpolating Fourier coefficient in which the complex Fourier coefficient of the head-related transfer function measured at the azimuth angle for each first angle unit with respect to the reference azimuth angle, which is the reference azimuth angle, is stored for the subject dummy A storage unit;
An angle that is an integer multiple of the first angle unit with respect to the reference azimuth angle, and is M times the first angle unit with respect to the reference azimuth angle (M is an integer of 2 or more) ) Is an interpolation coefficient which is a coefficient for interpolating a complex Fourier coefficient corresponding to one or more interpolation azimuth angles which is not an angle for each second angle unit which is an angle unit). The complex Fourier coefficient corresponding to the interpolation azimuth angle stored in the unit is stored in the Fourier coefficient storage unit for interpolation, and 3 for each second angle unit based on the reference azimuth angle. An interpolation coefficient calculation unit for calculating to be interpolated by multiplying the complex Fourier coefficient corresponding to the above azimuth by the interpolation coefficient;
A head-related transfer function interpolation coefficient calculation apparatus comprising: an interpolation coefficient storage section that stores the interpolation coefficient calculated by the interpolation coefficient calculation section. 2. The head-related transfer function interpolation according to claim 1, wherein the interpolation coefficient is a coefficient that has the same value even when an interpolation azimuth corresponding to the interpolation coefficient changes by an integral multiple of the second angle unit. Coefficient calculation device. An interpolation impulse response reception unit for receiving a head impulse response corresponding to an azimuth angle for each of the first angle units with respect to the dummy of the subject as a reference;
An interpolation Fourier transform unit that Fourier-transforms the head impulse response received by the interpolation impulse response reception unit and stores the Fourier transform complex Fourier coefficient in the Fourier coefficient storage unit for interpolation. The coefficient calculation apparatus for head related transfer function interpolation of Claim 1 or Claim 2. A Fourier coefficient storage unit for storing a complex Fourier coefficient of a head-related transfer function measured at an azimuth angle for each second angle unit with respect to the reference azimuth angle with respect to the subject;
Corresponds to the interpolation coefficient corresponding to one interpolation azimuth accumulated by the interpolation coefficient accumulating section and the complex Fourier coefficient stored in the Fourier coefficient storage section and used in the calculation of the interpolation coefficient. An interpolation unit that calculates a complex Fourier coefficient corresponding to the one interpolation azimuth by performing interpolation using a complex Fourier coefficient corresponding to three or more azimuths;
The head transfer function interpolation coefficient calculation device according to any one of claims 1 to 3, further comprising an interpolation Fourier coefficient storage unit that stores the complex Fourier coefficients calculated by the interpolation unit. An impulse response receiving unit that receives a head impulse response corresponding to an azimuth angle for each second angle unit with respect to the reference azimuth angle with respect to the subject;
The head according to claim 4, further comprising: a Fourier transform unit that Fourier-transforms the head impulse response received by the impulse response reception unit and accumulates the Fourier transform complex Fourier coefficient in the Fourier coefficient storage unit. Coefficient calculation device for transfer function interpolation. A coefficient calculation apparatus for head related transfer function interpolation according to claim 4 or 5,
An azimuth angle acceptance unit for accepting a sound image localization azimuth that is an azimuth angle for performing sound image localization;
A sound signal receiving unit for receiving a sound signal for performing sound image localization;
A sound image that is a complex Fourier coefficient accumulated by the interpolating Fourier coefficient accumulation unit and that performs sound image localization processing on the sound signal received by the sound signal reception unit using a complex Fourier coefficient corresponding to the sound image localization azimuth. A localization processing unit;
A sound image localization apparatus comprising: an output unit that outputs a sound signal obtained by the sound image localization processing unit performing sound image localization processing. The sound image localization apparatus according to claim 6, wherein the interpolation unit performs an interpolation process corresponding to an interpolation azimuth angle corresponding to the sound image localization azimuth angle received by the azimuth angle reception unit. Interpolating Fourier coefficient in which the complex Fourier coefficient of the head-related transfer function measured at the azimuth angle for each first angle unit with respect to the reference azimuth angle, which is the reference azimuth angle, is stored for the subject dummy A head-related transfer function interpolation coefficient calculation method processed using a storage unit, an interpolation coefficient calculation unit, and an interpolation coefficient storage unit,
The interpolation coefficient calculation unit is an angle that is an integer multiple of the first angle unit with respect to the reference azimuth angle, and is M times the first angle unit with respect to the reference azimuth angle ( M is an integer greater than or equal to 2) An interpolation coefficient that is a coefficient for interpolating a complex Fourier coefficient corresponding to one or more interpolation azimuth angles that are not angles for each second angle unit that is an angle unit The complex Fourier coefficient corresponding to the interpolation azimuth angle stored in the interpolation Fourier coefficient storage unit is stored in the interpolation Fourier coefficient storage unit, and the reference azimuth angle is used as a reference. An interpolation coefficient calculation step for calculating to be interpolated by multiplying a complex Fourier coefficient corresponding to three or more azimuth angles for each second angle unit by the interpolation coefficient;
An interpolating coefficient accumulating step in which the interpolating coefficient accumulating unit accumulates the interpolating coefficient calculated in the interpolating coefficient calculating step. Computer
An angle that is an integral multiple of the first angle unit with reference to the reference azimuth angle, which is the reference azimuth angle for the subject's dummy, and is M times the first angle unit with reference to the reference azimuth angle For interpolating a complex Fourier coefficient of a head related transfer function corresponding to one or more interpolated azimuth angles that are not angles for each second angle unit (M is an integer of 2 or more). Interpolation that stores the complex Fourier coefficient of the head-related transfer function measured at the azimuth angle for each first angle unit with respect to the reference azimuth angle as a coefficient for interpolation, which is a coefficient The second angle based on the reference azimuth angle, which is stored in the Fourier coefficient storage unit for the complex Fourier coefficient corresponding to the interpolation azimuth angle and stored in the interpolation Fourier coefficient storage unit. 3 or more per unit Interpolation coefficient calculation section that calculates as interpolated by the complex Fourier coefficient corresponding to the angular subjecting the interpolation coefficients,
A program for functioning as an interpolation coefficient accumulation unit that accumulates the interpolation coefficient calculated by the interpolation coefficient calculation unit.