JPH08205298A - Sound image localization controller - Google Patents

Abstract

PURPOSE: To provide a sound image localization controller which corrects sound image localization so as to obtain a sense of most suitable sound image localization for each listening position. CONSTITUTION: A sound image localization controller 1 reproduces the sound signal, which is obtained by convolutional operation of the signal outputted from a sound source X, from speakers sp1 and sp2 and controls sound image localization so as to orient the acoustic image in a position different from speakers sp1 and sp2 and is provided with a pair of convolvers 2 to which the signal outputted from the sound source 2 is supplied, a coefficient memory 3 from which convolver coefficient data cfLx and cfRx to orient the sound image in the position other than speakers sp1 and sp2 are supplied to convolvers 2, and a conversion table 4 which can correct the sound image localized at the rear of the listening position.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sound image localization control device for making a user feel that a sound image is localized at a desired arbitrary position different from the actual installation position of a transducer (speaker), and more particularly to an amusement game. IMAGE LOCALIZATION CONTROLLER, WHICH IS EXCELLENT IN A SOUND IMAGE LOCALIZATION FEEL AND CAN BE REDUCED IN A CIRCUIT, COMPUTER TERMINAL, AND THE SAME, AND IN REDUCED CIRCUIT SIZE

[0002]

2. Description of the Related Art Conventionally, there is a sound image localization device that makes a sound source feel at a specific position (direction) by a level difference and a phase difference (time difference) of signals in both ears. For example, Japanese Unexamined Patent Publication
The "sound image forming method and apparatus" described in Japanese Patent Publication No. 298200, which controls the localization of a sound image by giving a level difference and a phase difference depending on the frequency to both left and right channel signals emitted from two speakers. is there. In addition, the applicant of the present invention is Japanese Patent Application No. 4-343459 and Japanese Patent Application No. 4-356.
No. 358 was proposed. This is HRTF (head related transfer function)
The sound image is localized by determining the characteristics based on the actual measurement by.

In either case, the amplitude frequency characteristic and the phase frequency characteristic are digitally processed by signal processing to localize the sound image.

[0004]

However, the localization position of the sound image is dispersed at the intended position (the sound image is blurred),
There was a problem that it shifted from the intended position. Therefore, the present applicant proposed Japanese Patent Application No. 5-34946 to improve the localization of the sound image by incorporating the psychological element of HRTF.
This is to improve the localization of the sound image by using the correction process for each band. The present invention is based on the improvement of the positional deviation of the localization of the sound image, and further, the localization of the sound image can always be optimally obtained even by an individual environment due to a noisy external environment or a difference in the distance between the ears. An object is to provide a sound image localization control device.

[0005]

In order to solve the above problems, the present invention provides a sound image localization control device having the following configurations (1) and (2).

(1) What is the arrangement of the pair of transducers (speakers sp1, sp2) by reproducing the audio signal obtained by convolving the signal output from the sound source (X) from the pair of transducers (speakers sp1, sp2). A sound image localization control device 1 for performing sound image localization so that sound images are localized at different positions, and a pair of convolution operation processing circuits (convolvers) 2 to which signals output from the same sound source (X) are supplied.
And the pair of transducers (speakers sp1, s
sound image localization position coefficient (a pair of convolver coefficient data (cf
Lx, cfRx)) to the convolution operation processing circuit (convolver) 2 and a listening position (listener M).
Signal at a horizontal angle or elevation angle (horizontal angle from diagonally forward to diagonally rearward of the listener M (horizontal direction), elevation angle from right behind the listener M to vertical top (vertical direction)) at A sound image localization control device comprising: a correction unit (conversion table) 4 for correcting the sound image (a sound image localized by a pair of convolver coefficient data (cfLx, cfRx)) based on the above.

(2) What is the arrangement of the pair of transducers (speakers sp1, sp2) by reproducing the audio signal obtained by convolving the signal output from the sound source (X) from the pair of transducers (speakers sp1, sp2). A sound image localization control device 1 for performing sound image localization so that sound images are localized at different positions, and a pair of convolution operation processing circuits (convolvers) 2 to which signals output from the same sound source (X) are supplied.
And the pair of transducers (speakers sp1, s
sound image localization position coefficient (a pair of convolver coefficient data (cf
Lx, cfRx)) to the convolution operation processing circuit (convolver) 2 and a listening position (listener M).
Position), the angular position of the sound image localized is periodically changed (for example, the sound image is shaken by ± 10 degrees every 3 to 5 seconds) to ensure recognition of the localized sound image. A sound image localization control device characterized by.

[0008]

According to the above sound image localization control device, the HRTF (head related transfer function) at each sound image localization position is obtained as IR (impulse response) after accurate approximation processing, and the coefficient data of a pair of convolvers is obtained. (CfLx, cfR
x). The signal from the sound source (X) is subjected to convolutional arithmetic processing on the time axis by the pair of convolvers, and is reproduced from the pair of transducers (speakers sp1 and sp2) arranged apart from each other. For the sound reproduced from the pair of transducers, the HRTFs from the transducers to both ears are canceled, and the HRTFs from the desired arbitrary position (x) to both ears are reproduced,
The sound image is localized so that the sound source exists at a desired arbitrary position (x), and the listener (for example, game operator) M hears the sound image. In addition, even if you do not obtain the optimal sound image localization even if you do so (for example, when installed in a noisy external environment,
Alternatively, in the case of individual differences due to differences in the interaural distance of listeners who operate the device), the sound image localization position once set by the conversion table 4 can be corrected so that each listener can obtain an optimal sound image localization feeling. And

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A sound image localization control apparatus according to the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram of an embodiment of a sound image localization control device according to the present invention, FIG. 2 is a diagram illustrating mapping for correcting a deviation in sound image localization recognition, and FIG. 3 is a diagram illustrating how a sound image localization point is deflected. 4 is a configuration diagram of a general sound image localization device, FIG. 5 is a configuration diagram showing a basic principle of sound image localization control, and FIG.
8 is a block diagram showing an HRTF (head-related transfer function) measurement system, FIG. 7 is a diagram for explaining the points of HRTF measurement, and FIG.
Is a diagram illustrating an example of calculation of a cancellation filter, and FIG. 9 is H
FIG. 10 is a diagram showing a concrete example of IR (impulse response) of RTF, and FIG. 10 is a diagram showing a concrete example of coefficients of a cancel filter.

First, the basic principle of sound image localization control will be described. This is a technique for localizing a sound image at an arbitrary position in space by using a pair of transducers (which will be described below by taking a speaker as an example) arranged apart from each other.

In FIG. 5, which is a block diagram showing the basic principle of sound image localization control, sp1 and sp2 are speakers arranged in front of and to the left of the listener (namely, the listener M), and sp1.
From the head to the listener's left ear (impulse response)
Is h1L, the head-related transfer characteristics to the right ear are h1R, and the head-related transfer characteristics from sp2 to the left and right ears are h2L and h2R.
In addition, the head-related transfer characteristics to the listener's left and right ears when an actual speaker is placed at the intended localization position x are pLx, p
Rx. Here, each transfer characteristic is obtained by actually measuring with a speaker in a sound space and a microphone at both ears of a dummy head (or human head), and subjected to appropriate waveform processing.

Next, signals obtained by passing the sound source X to be localized through signal converters cfLx and cfRx (transfer characteristics by a convolver) are sp1 and sp, respectively.
Consider playing in 2. At this time, when the signals obtained in the left and right ears of the listener are eL and eR, eL = h1L · cfLx · X + h2L · cfRx · X eR = h1R · cfLx · X + h2R · cfRx · X 1).

On the other hand, when the signals obtained in the left and right ears of the listener when the source X is reproduced from the target localization position are dL and dR, dL = pLxX dR = pRxX where both equations are combined. , (Formula 2).

Here, if the signals obtained in the left and right ears of the listener by reproduction of sp1 and sp2 match the signals when the source is reproduced from the target position, the listener is as if the speaker exists at the target position. The sound image will be recognized. From this condition eL = dL, eR = dR and (Equation 1) and (Equation 2), X is deleted and h1L · cfLx + h2L · cfRx = pLx h1R · cfRx + h2R · cfRx = pRx Here, combining both equations, Let equation 3). When cfLx and cfRx are calculated from (Equation 3), cfLx = (h2R · pLx−h2L · pRx) / H cfRx = (− h1R · pLx + h1L · pRx) / H where both equations are combined to obtain (Equation 4a). To do. However, H = h1L · h2R−h2L · h1R ... (Equation 4b)

Therefore, if a signal to be localized is processed by a convolver (convolution operation processing circuit) or the like using the transfer characteristics cfLx and cfRx calculated by (Equation 4a) and (Equation 4b), the sound image at the target position x is obtained. Can be localized. Although there are various concrete methods for realizing the signal converter, an asymmetric FIR digital filter (convolver) is used.
Can be realized by using. The final transfer characteristic when used in the FIR digital filter is a time response function.

That is, as the transfer characteristics cfLx, cfRx at the required localization position x, (Equation 4a), (Equation 4b)
The coefficients cfLx and cfRx are created in advance as the coefficients for realizing the one obtained in step 1 by the FIR filter processing, and are prepared as ROM data.
If the coefficient of the required sound image localization position is transferred from the ROM to the FIR digital filter and the signal from the sound source is subjected to the convolutional arithmetic processing and reproduced from the pair of speakers, the sound image is localized at any desired position.

The real sound image localization control based on the above principle will be described in the order of to.

Head related transfer function
Function; hereinafter referred to as HRTF)

As shown in FIG. 6, a pair of microphones ML and MR are installed in both ears of a dummy head (or human head) DM, a measurement sound from a speaker SP is received, and a source sound (reference data) is recorded in a recorder DAT. The refL and refR and the to-be-measured sound (measurement data) L and R are recorded in synchronization.

The source sound XH may be impulse sound, white noise, or other noise. In particular, from the viewpoint of statistical processing, white noise is
Since it is a continuous sound and the energy distribution is constant over the audio band, it is possible to use SN by using white noise.
The ratio is improved. The position of the speaker SP is set at a plurality of angles θ (for example, 12 points for every 30 degrees as shown in FIG. 7) in the space arranged with the front as 0 degree, and each speaker is continuously operated for a predetermined time. To record.

HRTF impulse response (Impulse Re
Calculation of sponse; hereinafter referred to as IR) Source sounds (reference data) refL and refR and sound to be measured (measurement data) L and R, which are synchronously recorded in the above-described HRTF measurement, are stored in a workstation (Fig. Process (not shown).

Letting X (S) be the frequency response of the source sound (reference data), Y (S) be the frequency response of the sound to be measured (measurement data), and IR (S) be the frequency response of the HRTF at the measurement position. There is an input / output relationship as shown in (Equation 5). Y (S) = IR (S) * X (S) (Equation 5) Therefore, the frequency response of the HRTF is calculated as IR (S) = Y (S) / X (S). (Equation 6) Therefore, the frequency response X (S) of the reference,
The frequency response Y (S) of the measurement data is obtained in step 10 above.
The data obtained in 1 is cut out in a time-synchronized window, and a finite Fourier series expansion is performed by FFT conversion to calculate discrete frequencies. From (Equation 6), the frequency response IR (S) of the HRTF is calculated by a known calculation method. Required by.

In this case, the accuracy of IR (S) is increased (S
In order to improve the N ratio), IR (S) may be calculated for several hundred windows that are temporally different, and they may be averaged. Then, the calculated HRTF frequency response IR (S)
Is subjected to inverse FFT conversion to obtain a time axis response (impulse response) IR (first IR) of HRTF.

IR (Impulse Response) Shaping Processing Here, the IR obtained by the HRTF impulse response is shaped. First, for example, by FFT conversion, the above-mentioned first IR is expanded at a discrete frequency over the audio spectrum, and an unnecessary band (a large dip occurs in the high range, but this is an unnecessary band that does not significantly affect the sound image localization). Are removed by a BPF (band pass filter). By limiting the band in this way, unnecessary peaks and dips on the frequency axis are removed and unnecessary coefficients do not occur in the cancel filter, so that the convergence is improved and the coefficients can be shortened.

Then, the band-limited IR (S) is inverted F
FT conversion is performed, and IR (impulse response) is cut out on the time axis by a window (for example, a cosine function window) and subjected to window processing (second IR). By performing the window processing, the effective length of the IR is not long, the convergence of the cancel filter is improved, and the sound quality is not deteriorated. FIG. 9 shows a specific example of IR (impulse response) of HRTF. The horizontal axis is time (sample clock is 4
The clock unit is 8 kHz), and the vertical axis is the amplitude level. The chain double-dashed line shows the window.

Cancellation filters cfLx, cfRx
The cancellation filters cfLx and cfRx, which are the convolvers (convolution operation processing circuits), have the values shown in (Expression 4a) and (Expression 4b) described above.

Here, the speakers sp1 and sp to be arranged
Head related transfer characteristics h1L, h1R, h2L, h2R
And, as the head-related transfer characteristics pLx and pRx when the actual speaker is arranged at the target localization position x, the second IR (impulse response) that has been subjected to the shaping process for each angle θ obtained by ) Is substituted.

Head-related transfer characteristics h1L and h1R are L in FIG.
Corresponding to the position of the channel speaker, if it is installed at 30 degrees (θ = 330 degrees) from the front to the left, an IR of θ = 330 degrees is used. Head-related transfer characteristic h2
R and h2L correspond to the position of the R channel speaker in the figure, and are, for example, 30 degrees (θ = 30) from the front to the right.
, The IR of θ = 30 degrees is used (that is, the system at the time of actual sound image reproduction (see, for example, FIG. 4).
Select one that is close to)).

The head-related transfer characteristics pLx and pRx are 90 degrees from the front, which is the target sound source localization position.
The cfLx of the entire space corresponding to that by substituting IR for every 30 degrees in a wide space (overall space) beyond that of 180 °
cfRx, that is, 12 sets of cancellation filters cfLx and cfRx are obtained every 30 degrees (in FIG. 8,
The 240 ° position is taken as an example). The cancel filters cfLx and cfRx group are finally obtained as IR (impulse response) which is a response on the time axis.

The calculation of the cancel filters cfLx and cfRx according to (Equation 4a) is as follows. First, H ^-1 , which is a kind of inverse filter for H in (Equation 4b), is obtained by the least squares method, and this is subjected to inverse FFT conversion to obtain a time function h (t). Also, the terms h1L and h1 of (Equation 4a)
By expressing R, h2L, pRx, pLx, and h2R by a time function, respectively, the following equation holds. cfLx (t) = (h2R · pLx−h2L · pRx) ·
h (t) cfRx (t) = (-h1R · pLx + h1L · pRx)
-H (t) Here, both equations are combined to obtain (Equation 7).

Therefore, the coefficients of the cancel filters cfLx and cfRx can be obtained from these (formula 7). As is clear from (Equation 7), in order to shorten the coefficients of the cancel filters cfLx and cfRx, the head-related transfer characteristics h1L, h1R, h2L, pRx, pLx,
It is extremely important to shorten each h2R. Therefore, as described above, various processes such as the window process and the shaping process are performed in to each head-related transfer characteristic h1L,
h1R, h2L, pRx, pLx, h2R are shortened.

FIG. 10 shows the cancel filter coefficient cfL.
A specific coefficient sequence of x and cfRx is shown. The horizontal axis is time (time in clock units where the sample clock is 48kHz),
The vertical axis is the amplitude level. The chain double-dashed line shows the window. The cancel filter coefficients cfLx and cfR
Alternatively, x may be FFT-transformed to obtain a frequency response, which may be moving-averaged within a certain width, and inverse FFT-transformed to obtain the final time response of the cancellation filter. By performing moving averaging in this way, unnecessary peaks and dips can be removed, the time response to be realized can be converged quickly, and the scale of the cancel filter can be reduced.

Scaling of Cancellation Filter at Each Localization Point x Further, the spectrum distribution of the sound source (source sound) actually processed by the convolver (cancellation filter) is statistically distributed like pink noise, Alternatively, there is something that gently drops in the high range, and in any case, since the sound source is different from a single sound, there is a risk that the convolution operation (integration) overflows and distortion occurs. Therefore, in order to prevent the overflow, the maximum gain (for example, the cancel filter cfL) among the coefficients of the cancel filters cfLx and cfRx.
x, cfRx), and all coefficients are scaled so that overflow does not occur when the coefficient and the white noise of 0 db are convoluted. Practically, the maximum absolute value is about 0.1 to 0.4 (for example, 0 ..
It is good to attenuate so that it becomes 2).

Then, by using the window (cosine window) shown in FIG. 10, according to the number of coefficients of the actual convolver, the window processing is performed so that both ends become 0, and the effective length of the coefficient is shortened. In this way, the data group (in this example, 12 sets of convolver coefficient groups capable of sound localization for every 30 degrees) cfLx and cfRx that are subjected to scaling processing and finally supplied to the convolver as coefficients are obtained.

A signal from a sound source is convolutionally calculated and reproduced. For example, as a sound reproducing device of a game machine, as shown in FIG. 4, a game operator (listener) M is left and right 30
A pair of loudspeakers sp1 and sp2 are arranged so as to be spaced apart from each other, and the pair of loudspeakers sp1 and sp2 is configured to reproduce an acoustic signal processed by a pair of convolvers (convolution operation processing circuits). . For a pair of convolvers,
A signal from the same sound source X (for example, a flight sound from a game synthesizer) is supplied, and the IR coefficients cfLx and cfRx created in step 105 are supplied.
(For example, if you want to localize the flight sound to the left 120 ° (θ = 240 °) position, a coefficient of θ = 240 °)
Is selected and set as the convolver. For example, based on a sound image localization command from a main CPU (central processing unit) of a game machine or the like, the control sub CPU has a coefficient R.
The coefficient of the desired localization position is transferred from the OM to the pair of convolvers.

In this way, the signal from the sound source X is subjected to the convolution operation processing on the time axis by the pair of convolvers, and the pair of speakers sp1 and sp1 arranged apart from each other.
Played from 2. The sound reproduced from the pair of speakers sp1 and sp2 cancels the HRTFs from the speakers sp1 and sp2 to both ears, and reproduces the HRTF from any desired position to both ears, thereby producing a sound source at a desired position. The sound image is localized as described above, and is heard by the listener M, and is reproduced as a sound that is extremely realistic. With respect to the coefficient of the convolver, the optimum sound image position is sequentially selected and switched with the transition of the movement of the airplane according to the operation of the listener M. Further, when the flight sound is changed to, for example, a missile sound, the source sound from the sound source X is changed from the flight sound to the missile sound. In this way, the sound image can be freely localized at any position.

As a transducer for reproduction, a headphone can be used instead of the pair of speakers sp1 and sp2. In this case, since the HRTF measurement conditions are different, it is advisable to prepare the coefficient separately and switch it according to the reproduction situation.

The IR (impulse response) shaping process is not always necessary, and the sound image localization can be controlled even if omitted.

Further, in the present sound image localization control, the configuration for reproducing the signal processed by the pair of convolvers to which the same sound source is supplied from the pair of transducers arranged apart is the minimum for obtaining the effect of the present application. This shows the configuration of the limit. Therefore, if necessary, a pair, that is, 2
Of course, one or more transducers and convolvers may be additionally configured, and when the coefficient of the convolver is long, the coefficient may be divided into a plurality of convolvers.

Further, the coefficient of the convolver is the opening angle of the speaker (that is, the angle sp1-M-sp in FIG. 4).
Since it depends on 2), it may be obtained for each opening angle of the speaker so that it can be selectively used according to the actual reproduction system. That is, in the above description, the convolver coefficient is calculated when the pair of speakers sp1 and sp2 are arranged with the opening angle of 60 degrees and the left and right sides of the game operator (listener) separated by 30 degrees. , When calculating the above cancellation filter, the speaker sp
Head-related transmission characteristics h1L, h1R, h2 due to 1, sp2
As L and h2R, values corresponding to other opening angles, for example, 45 degrees and 30 degrees may be substituted and obtained.

Since the convolver coefficient differs depending on the HRTF measurement conditions, this point may be taken into consideration. That is, since there are individual differences in the size of the human head, H
At the time of RTF measurement, the size of the dummy head (or human head) is changed to obtain several types, and selectively used according to the viewer (for example, adults with large heads and children with small heads). You may allow it.

Further, a position not actually measured, for example, θ =
When it is desired to more precisely localize the sound image every 15 degrees, the intermediate value may be calculated by the above or based on the actually measured coefficients cfLx and cfRx. At this time, instead of simply arithmetically averaging the two actual measurement values (for example, in order to obtain the intermediate value at θ = 15 degrees, the measurement values at θ = 0 degree and θ = 30 degrees), the convolver based on the actual measurement is used. After the FFT conversion of the coefficients cfLx and cfRx of 1 to obtain the frequency response, the frequency-amplitude characteristic is obtained as the geometric mean of the amplitude characteristics of the reference transfer characteristics, and the frequency-phase characteristic is the vector average of the frequency complex vector of the reference transfer characteristics. It is good to obtain it as a phase component. When the intermediate value is obtained in this way, an intermediate transfer characteristic that closely approximates the actually measured value is obtained, and there is almost no deterioration in localization and sound quality.

In measuring the HRTF in the above, only the semicircular portion from θ = 0 ° to 180 ° is actually measured, and the remaining semicircular portions are measured values by utilizing the symmetry of the coefficient. You may divert. In this case, the measurement becomes easy, and the calculation of IR and the calculation of the coefficient are not unnecessarily fine, and a coefficient excellent in sound image localization may be obtained.

As described in detail above, according to the above-described sound image localization control technique, the signal from the sound source is processed on the time axis by the pair of convolvers to localize the sound image. As described above, only one set of convolvers on the time axis is required as shown in the above, and the circuit scale is very small and inexpensive.

Further, since the coefficient data of the sound image localization processing of the convolver is finally supplied as IR (impulse response) data on the time axis, in order to further reduce the circuit scale, the convolver is reduced. It suffices to reduce the number of coefficients.

Further, since the IR (impulse response) data is supplied as a coefficient to the convolver for processing, the IR coefficient is uniquely obtained in a simple and adaptive manner without being adaptive in the time domain. Further, since the delay time of the response waveform on the time axis can be clearly defined, the mutual time relationship between the response waveforms at a plurality of points can be accurately controlled. Further, the convolver coefficient can be accurately measured for the phase and amplitude for each frequency based on the actual measurement, and the sound image can be localized in a wide space exceeding the range of 180 degrees.

Further, the IR shaping processing according to the above,
That is, the first IR corresponding to the measured HRTF is obtained, predetermined processing (band limitation) is performed in the discrete frequency band extending from the first IR to the audio spectrum, and the cutout window (for example, cosine) is again performed on the time axis. Function window)
If the second IR whose length is converged to a predetermined value is obtained and the coefficients of the pair of cancel filters are obtained from the second IR, the cancel filter is calculated as described above. The accuracy of the calculation in the calculation process is improved.

Further, in the above, if the white noise is used to measure the HRTF in each sound image localization, the SN ratio is improved by using the white noise, and the HRTF (the IR and the coefficient based on it) is accurately measured. You can ask. Further, in the above, a plurality of IRs obtained corresponding to each HRTF are averaged, that is, as described above, IR (S) is calculated for each of several hundred windows that are temporally different, and Is averaged, the SN ratio is improved and the accuracy is improved.

Further, according to the above, if the maximum value of the coefficient size is scaled so as to be the maximum level of 0.1 to 0.4 and attenuated, overflow does not occur and the reproduced sound is not distorted. .

In the sound image localization control technique described above, in order to localize a sound image at a position different from the position of the pair of speakers sp1 and sp2, the coefficient data (c
fLx, cfRx) is supplied to the pair of convolvers 2 to control the horizontal angle of 360 degrees around the listener M and the distance from the listener M. Elevation control from right beside to right above is not included.

Therefore, the present inventor has studied to incorporate elevation angle control in this sound image localization control technique. As a result,
Only by incorporating uniform elevation angle control data into the convolver coefficient data (cfLx, cfRx) relating to horizontal angle and distance control, the listener M reproduces the sound source (source) with respect to the realization probability of sound localization in the elevation direction, It was noted that there is a large variation in the accuracy of localization and that the sound image is likely to be reversed back and forth when the sound image is stationary on the rear surface of the listener M even on the horizontal plane.

That is, in order to optimally perform the sound image localization control in the elevation angle direction in the sound image localization control, for each listener M,
It is necessary to finely correct the sound image localization for each source to be played back.

Therefore, by using the correction data based on mapping, which will be described later, it is possible to improve the rearward localization and the vertical plane localization, and it is possible to perform optimum sound image localization for each listener M and each source to be reproduced. it can.

(1) Mapping of localization filter for correcting deviation of auditory localization recognition The present inventor conducted a spatial sound image localization experiment using an actual sound. As a result, the position at which the listener M recognizes the sound image does not necessarily match the position at which the sound image is presented, and the difference between the presentation position and the localization recognition position depends on the presentation position regardless of the listener M. , (B) showed almost the same tendency.

(A) The deviation of the horizontal angle is likely to occur in front of the listener M as a whole, and this tendency is particularly strong from right beside the listener M to obliquely rearward. That is, as shown in FIG. 2 (A), the presentation position a (directly in front of the listener M), b, c, d (listening position) that is mapped at a horizontal position centered at the listener M and separated from the listener M by a certain distance. (Directly behind the person M). After that, the localization recognition position where the listener M actually recognizes the sound image is mapped. As a result, the presentation positions a and d do not deviate from the localization recognition position. On the other hand, the presentation position c deviates from the localization recognition position. For the presentation position c, a localization recognition positional deviation occurs as if it were for the presentation position b. In particular, the listener M
A gap between the presentation position and the localization recognition position occurs from right next to the rear.

By positively utilizing the difference between the presenting position and the localization recognition position, in this example, one example is as follows:
When the sound image is localized at the presentation position b, a pair of convolver coefficient data (cfLx, c) corresponding to the presentation position c.
If fRx) is corrected and set, the localization recognition position can be optimally obtained.

(B) The deviation of the elevation angle is likely to occur above the listener M as a whole, and this tendency is particularly strong around the listener M. That is, as shown in FIG. 2B, the presentation position e (directly behind the listener M), f, which is mapped at an elevation angle position centered on the listener M from a position directly behind the listener M toward the crown and a certain distance apart. Set g and h (crown of listener M). After that, the localization recognition position where the listener M actually recognizes the sound image is mapped. As a result, the presentation positions e and h do not deviate from the localization recognition position. On the other hand, the presentation position f is displaced from the localization recognition position. For the presentation position f, it is as if the presentation position g
As a result, the localization recognition position shift as if it is the one occurs.
In particular, a deviation between the presentation position and the localization recognition position occurs in the upper rear part of the listener M except right behind and the crown.

By positively utilizing such a difference between the presentation position and the localization recognition position, in this example, one example is as follows:
When the sound image is localized at the presentation position g, a pair of convolver coefficient data (cfLx, c) corresponding to the presentation position f.
If fRx) is corrected and set, the localization recognition position can be optimally obtained.

By the way, although there is a degree of difference depending on the listener M, the sound subjected to sound image localization tends to be localized in the motion, but the sound image is reversed in the front and back when the sound image is stationary on the rear surface. It can be easy. When the sound image was kept still, we examined by experiments whether it is possible to clarify the localization by swaying the localization point.

As a result, with respect to the swing width, when the swing cycle was short, the quiver of the playback source was felt. That is, FIG.
As shown in (A), within the range of angle + θ to angle −θ, the localization point was continuously wiggled at intervals of 1 second (angle −θ → 0 → + θ → −θ 1 second rest −θ → 0 → + Θ
→ -θ ...). Incidentally, the angle + θ is a required angle diagonally to the right of the localization point, and the angle −θ is a required angle diagonally to the left of the localization point.

When the shaking cycle was long, the localization movement itself was felt. That is, as shown in FIG. 3B, within the range of angle + θ to angle −θ, + θ (1 second) → 0 (2 seconds) → −θ (1 second) → 0 (2 seconds) →
... and swayed the localization point in a long cycle.

Now, as shown in FIG. 1, the sound image localization control device for improving the rearward localization and the vertical plane localization by using the correction data based on the above-mentioned mapping is as follows.
The audio signal obtained by convoluting the audio signal output from the sound source (X) is reproduced from the speakers sp1 and sp2, and the speaker s
A device 1 that performs sound image localization so that a sound image is localized at a position different from the positions of p1 and sp2, and includes a pair of convolvers 2 to which a signal output from a sound source (X) is supplied, and an arrangement of speakers sp1 and sp2. Is a coefficient memory 3 for supplying the convolver 2 with coefficient data cfLx, cfRx for localizing a sound image at different positions, and a listening position (listener (for example,
The conversion table 4 for correcting the sound image localized by the coefficient data cfLx, cfRx localized in the sound image is provided behind the position (game operator) M). The same components as those described above are designated by the same reference numerals and the description thereof will be omitted.

As a result, the listening position (the position of the listener M)
If the optimum sound image localization is not obtained in
Specifies a signal corresponding to a position where optimum sound image localization cannot be obtained by an input device (not shown) and inputs it. When the main CPU recognizes this input state, the control sub CPU
, The coefficient data cfLx, cfRx corresponding to the above-mentioned position supplied from the coefficient memory 3 to the convolver 2.
Instead of the coefficient data cfLx, c corresponding to the above position
A control signal for switching and outputting fRx (correction data) from the conversion table 4 to the coefficient memory 3 is output to the coefficient memory 3 and the conversion table 4. Such correction data is data based on the mapping of the localization filter that corrects the deviation of the above-mentioned auditory localization recognition, and this correction data is stored in the conversion table 4 in advance.

In this way, when a specific position where optimum sound image localization is not obtained at the listening position occurs, correction data relating to the specific position is switched and output,
Rear localization and vertical surface localization can be improved, and optimal sound image localization can be performed for each listener M and each source to be reproduced.

For example, when the listener M cannot optimally obtain the rearward localization due to the deviation of the horizontal angle as described above, FIG.
As shown in (A), the sound image localization recognition positional deviation occurs so that the sound image at the presentation position c is localized at an angular position corresponding to the presentation position b. The pair of convolver coefficient data (cfLx, cfRx) output from the coefficient ROM 3 is not the coefficient data obtained from the HRTF from the localization target point to both ears,
This is the coefficient data that is considered to be localized in consideration of the deviation of localization recognition (that is, the coefficient to be localized at the presentation position b is a coefficient created from the HRTF from the presentation position c).

Therefore, when the pair of convolver coefficient data (cfLx, cfRx) output from the coefficient ROM 3 indicates the presentation position b, the correction data for outputting the coefficient data at the presentation position c is converted instead. All that is necessary is to output from the table 4 to the coefficient memory 3. The correction data is read from the conversion table 4 under the control of the control sub CPU.

Similarly, in the case where the listener M cannot optimally obtain the rearward localization due to the above-described elevation angle deviation, FIG.
As shown in (B), the sound image localization recognition positional deviation occurs so that the sound image at the presentation position f is localized at an angular position corresponding to the presentation position g. The pair of convolver coefficient data (cfLx, cfRx) output from the coefficient ROM 3 is not the coefficient data obtained from the HRTF from the localization target point to both ears,
This is the coefficient data in which it is felt that the localization is recognized in consideration of the deviation of the localization recognition (that is, the coefficient to be localized at the presentation position g is a coefficient created from the HRTF from the presentation position f).

Accordingly, when the pair of convolver coefficient data (cfLx, cfRx) output from the coefficient ROM 3 indicates the presentation position g, the correction data for outputting the coefficient data at the presentation position f is converted instead. All that is necessary is to output from the table 4 to the coefficient memory 3.

Further, in the case where the listener M cannot optimally obtain the rearward localization due to the above-mentioned reversal of the sound image, FIG.
As shown in (A), the period in which the sound image is shaken is fluctuated in small increments (angle −θ → 0 → + θ → −θ 1 second pause −θ → 0 → + θ
→ -θ ...) makes it possible to recognize a stationary sound image in a stable and reliable manner without inversion. That is, the coefficient R
A pair of convolver coefficient data (cfLx, cfRx) corresponding to −θ, 0, + θ is stored in advance in OM3. As an example, the convolver coefficient data (cfLx, cfRx) at the presentation position e is −θ (for example,
Coefficient data of the presentation position e at the center 0 to about −10 degrees), coefficient data of the presentation position e at the center 0, coefficient data of the presentation position e at + θ (for example, an angular position from the center 0 to +10 degrees). Store respectively.

When the stable localization of the sound image that is stationary at the listening position (the position of the listener M) cannot be obtained, the listener M does not obtain a stable localization of the sound image by the input device (not shown). Specify and input a signal according to.
When this input state is recognized by the main CPU, the coefficient data of a pair of convolvers corresponding to the above-mentioned coefficient data -θ, 0, + θ corresponding to the specific position of the control sub CPU is -θ → 0 → + Θ → −θ pause for 1 second −θ → 0 → + θ → −θ ... The control signal indicating the sequential output from the coefficient ROM 3 is output to the coefficient ROM 3.

In this way, the sound image corresponding to the above-mentioned specific position can be periodically swayed in small steps, so that at the listening position (the position of the listener M), a stationary sound image is stably and reliably reversed without being reversed. Can be recognized.

Similarly, in the case where the listener M cannot optimally obtain the rearward localization due to the above-described reversal of the sound image, FIG.
As shown in (B), the sound image is shaken with a long cycle (+ θ (1 second) → 0 (2 seconds) → −θ (1 second) → 0.
By ((2 seconds) → ...), a stationary sound image can be stably and surely recognized without reversal. That is,
The coefficient ROM 3 stores a pair of convolver coefficient data (cfLx, cfRx) corresponding to −θ, 0, + θ in advance. As an example, as the convolver coefficient data (cfLx, cfRx) of the presentation position e, the coefficient data of the presentation position e at −θ (for example, an angular position of about −10 degrees from the center 0) and the presentation position e at the center 0 are shown. The coefficient data and the coefficient data of the presentation position e at + θ (for example, an angular position of about +10 degrees from the center 0) are stored.

When the stable localization of the sound image that is stationary at the listening position (the position of the listener M) cannot be obtained, the listener M does not obtain a stable sound image localization by the input device (not shown). Specify and input a signal according to.
When this input state is recognized by the main CPU, the pair of convolver coefficient data corresponding to the above-mentioned coefficient data -θ, 0, + θ corresponding to the specific position of the control sub CPU is + θ (for 1 second). → 0 (2 seconds) →
A control signal for sequentially outputting from the coefficient ROM 3 is output to the coefficient ROM 3 in the order of −θ (1 second) → 0 (2 seconds) → ....

In this way, the sound image corresponding to the above-mentioned specific position can be periodically swayed for a long time, so that the stationary sound image can be recognized reliably and reliably at the listening position (the position of the listener M) without back-and-forth reversal. can do.

The above-mentioned displacement of the sound image localization recognition position or the reversal of the sound image in the front-rear direction is caused by an environment of a listening position where external noise is large.
Since the difference greatly depends on the individual difference of the listener or the music source, the sound pattern localization recognition position shift or the front-back inversion phenomenon of the sound image is detected in advance by the test pattern, and the above correction data based on this detection is converted into the conversion table 4. By storing in the memory, the above correction can be efficiently performed.

[0076]

As described in detail above, according to the sound image localization control apparatus of the present invention, the signal from the sound source is processed on the time axis by the pair of convolvers to localize the sound image. As a circuit for performing sound image processing, only a convolution operation processing circuit on the time axis is required, and the circuit scale is very small and inexpensive. Further, since the coefficient data of the sound image localization processing of the convolver is finally supplied as IR (impulse response) data on the time axis, in order to further reduce the circuit scale,
It is sufficient to reduce the number of convolver coefficients.
As a result, the HRTF can be accurately and effectively approximated as compared with the conventional approximation of the data on the frequency axis, and from this point as well, the circuit scale can be reduced without impairing the sound image localization feeling. . Furthermore, especially when a uniform sound image localization control does not provide a good sound image localization feeling, such as when the listener is listening in an environment with a lot of external noise, or when the listener has a difference that is greater than the standard distance between both ears. By correcting the sound image localization position behind the listening position, an optimal sound image localization feeling can be obtained.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment of a sound image localization control device according to the present invention.

FIG. 2 is a diagram illustrating mapping for correcting a deviation in sound image localization recognition.

FIG. 3 is a diagram illustrating how a sound image localization point swings.

FIG. 4 is a configuration diagram of a general sound image localization device.

FIG. 5 is a configuration diagram showing a basic principle of sound image localization control.

FIG. 6 is a configuration diagram showing an HRTF (head related transfer function) measurement system.

FIG. 7 is a diagram illustrating the points of HRTF measurement.

FIG. 8 is a diagram illustrating an example of calculating a cancel filter.

FIG. 9 is a diagram showing a specific example of IR (impulse response) of HRTF.

FIG. 10 is a diagram showing a specific example of coefficients of a cancel filter.

[Explanation of symbols]

1 sound image localization control device 2 convolution arithmetic processing circuit, convolver 3 coefficient ROM (coefficient memory) 4 conversion table (correction means) cfLx, cfRx cancellation filter (convolver) and its coefficient M listener sp1, sp2 speaker (transducer) X sound source

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical display location H04R 1/40 310 H04S 7/00 F

Claims (2)

[Claims] 1. A sound image localization control for reproducing a sound signal obtained by convoluting a signal output from a sound source from a pair of transducers, and performing sound image localization so that a sound image is localized at a position different from the arrangement of the pair of transducers. In the apparatus, a pair of convolution calculation processing circuits to which signals output from the same sound source are supplied, and a sound image localization position coefficient for localizing a sound image at a position different from the arrangement of the pair of transducers are convoluted. A sound image localization control device comprising: a coefficient memory supplied to a processing circuit; and a correction unit that corrects a sound image based on a correction signal regarding a sound image localization of a horizontal angle or an elevation angle at a listening position. 2. A sound image localization control for reproducing a sound signal obtained by performing convolution calculation of a signal output from a sound source from a pair of transducers and performing sound image localization so that a sound image is localized at a position different from the arrangement of the pair of transducers. In the apparatus, a pair of convolution calculation processing circuits to which signals output from the same sound source are supplied, and a sound image localization position coefficient for localizing a sound image at a position different from the arrangement of the pair of transducers are convoluted. With a coefficient memory supplied to the processing circuit, it is possible to ensure the recognition of the localized sound image by using the sound image localization position coefficient for periodically changing the angular position of the sound image localized at the listening position. Characteristic sound image localization control device.