JPH0937397A - Method and device for localization of sound image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a sound image localizing method and device capable of reducing a time delay from the input of an original sound up to its output and localizing an always stable sound image. SOLUTION: Arithmetic processing based upon the filter characteristics X, Y of prescribed filters FX, FY is applied to sound signals from plural sound sources AX, AY based upon transmission functions gL, gR, hL, hR existing between plural sound sources AX, AY and a listener P and transmission functions zL, zR existing between a virtual sound source AZ existing on an optional virtual position and the listener P. Thereby the sound image localizing method allows the listener P to localize sound waves as if the sound waves are radiated from the virtual sound source. The filter characteristics X, Y are used for executing operation by using an all-band pass component related to a division component out of an equation for regulating the filter characteristics as a fixed value. Since the influence of the all-band pass component upon waveform characteristics is small, the output of an impulse signal can be obtained quickly from its input and the localization of a sound image can always be stabilized.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a so-called sound image localization device, and more particularly to a sound field processing technique for allowing a listener to hear sound from a sound source that does not actually exist.

[0002]

2. Description of the Related Art In recent years, due to the progress of voice processing technology, sound localization for causing a listener to recognize a sound image so that the sound source actually exists at a position where the sound source does not exist. Technology is being developed.

The sound image localization technique includes a head-related transfer function (head-r
Measurement of elated transfer function) is important. For a paper on the head-related transfer function, see, for example, "Binaural Sound Field Reproduction System" (Mr. Haruo Hamada), published in Journal of Acoustical Society of Japan, Vol. 48, No. 4 (1992). An invention to which this sound image localization technique is applied is disclosed in Japanese Patent Laid-Open No. 3-270400.
JP-A-5-252597, JP-A-5-25259
No. 8, JP-A-6-47170 or JP-A-6-5175.
It is published in each gazette of No. 9.

First, a conventionally known sound image localization technique will be described with reference to FIG. In a normal use state, the listener can recognize the position of the sound source by listening to the sound emitted from the sound source such as the speaker.
The sound field localization technology allows a listener to hear a sound wave radiated from a sound source that actually exists, but a sound source that exists at a position different from the position of the actual sound source (referred to as a “virtual sound source”) to the listener. The sound waves are emitted from the device. In order for the listener to recognize the virtual sound source, a predetermined sound input as an audio signal is converted by a filter and then emitted from the sound source.

In FIG. 8, input terminals L and R are input.
Original sound AX, AY (for simplicity, the unit impulse
You. ) Is the filter F_X, F_YSound by going through
Waveform BX, BY converted into a speaker SP_X, SP_Y
Emitted from. Filter F _XAnd F_YFor input
And perform a convolution operation based on a predetermined transfer function FX, FY.
Convert the waveform. For obtaining transfer functions FX and FY
This will be described later.

Speaker SP_X, SP_YIs in front of listener P
It is located in. The listener P has two speakers SP_XOver
And SP_YAnd left and right ears_L, E_R4 propagations between
Road g _L, G_R, H_L, H_RListen to sound waves through. sound
Waveforms BX and BY are the propagation paths g_L, G_R, H_L, H_R
Sound field conditions (particularly due to the shape of the listener's head and outer ears)
The waveform is transformed under the influence of the following. Sound waves in both ears
E_hl, E_hrAnd the listener resynthesizes it in the brain.
As a result, the sound waves are recognized.

[0007] Such listeners and speakers SP _X , SP
_{In order} to realize sound image localization in the sound field with _Y , when the listener recognizes the sound waves emitted from the speakers SP _X and SP _Y , the sound waves are transmitted from the virtual sound source S to the original sound AX to the listener.
It must be recognized that AY was emitted.

For this reason, the original sound radiated from the virtual sound source
AX and AY are the propagation paths z_R, Z _LPropagate this
Sowing route z_R, Z_LSound field conditions (especially on the listener's head or outer ear)
Waveform converted by the influence of
E_ZR, E_ZLHowever, the left and right ears of the listener e_L, E_RTo reach
Suppose Waveforms that are assumed to have reached left and right
E_ZR, E_ZLIs the actual speaker SP_X, SP_YReached from
Waveform E_gl, E _gr, E_hl, E_hrAnd a combination of
Must be equal. This conditional expression is expressed by the expressions (1) and (2)
Shown in

XG_L+ YH_L= Z_L… (1) XG_R+ YH_R= Z_R(2) In the above equation, X and Y are filters F_X, F_YPerformed at
This is the transfer function that is arithmetically processed. G_L, H_L, Z_L,
G_R, H_R, Z_RIs the speaker SP_X, SP_YAnd left and right
Ear e_L, E_REach path between_L, G_R, H_L, H_RIn
The transfer function is determined by the sound field condition. these
The transfer function is a requirement for determining the transfer function.
Since the influence of the shape of the head of a person is large,
It is called "transfer function". Z_L, Z_RIs a virtual sound source S
And left and right ears_L, E_REach path z between _R, Z_LIn
These are transfer functions that are determined by the sound field conditions.
The partial transfer function.

Based on equations (1) and (2), equations (3) and (4)
Can lead. _{X = (Z R H L -Z} L H R) / (G R H L -G L H R) ... (3) Y = (Z L H R -Z R H L) / (G R H L -G _L H _R) ... (4) HRTF _{G L, H L, Z L} , G R, H R, because Z _R is can be determined by measurement, by calculation of the right side of equation (3) (4) , The transfer functions X and Y to be set in the filters F _X and F _Y can be obtained.

The transfer functions X and Y thus obtained are
By using the filters F _X and F _Y shown in FIG. 8, the listener P listens to the sound waves emitted from the speakers SP _X and SP _Y, and thus listens to the original sounds AX and AY emitted from the virtual sound source. Can be recognized as if.

FIG. 9 shows a method of measuring the head related transfer function.
As shown in FIG. 9, the microphones M _L and M _R are attached near the listener P, and the sound wave waveforms near both ears can be measured. The sound field surrounding the listener, the speaker S ₁ ~S ₃ is installed as a sound source. Since the sound waves emitted from the sound sources S _{1 to} S ₃ are known in advance, the head-related transfer function between the sound source and the ear of the listener is determined from the original sound wave form of these sound waves and the sound wave waveform measured near both ears. Can be asked.

For example, when the unit impulse waveform W _S1 is radiated from the sound source S ₁ , and the propagation waveform in the vicinity of the left ear e _L is W _S1e , the head between the sound source S ₁ and the left ear e _L Since the _partial transfer function G is the impulse response itself, it can be seen that the transfer function represented by the propagation waveform W _S1e is the head related transfer function. In this way, the head related transfer function of each path existing between the sound source (virtual sound source) and both ears can be sequentially obtained.

By the way, the transfer functions X and Y of the filters F _X and F _Y are obtained from the equations (3) and (4). However, it is difficult to calculate the transfer functions in the time domain, and therefore in the frequency domain. It is preferable to calculate. Therefore, the head related transfer function is converted into a frequency domain expression, and the filter characteristic is determined by a predetermined calculation. In addition, FIR (Finite Impulse Respo
Since it is preferable to perform the waveform transformation by a digital filter such as an nse) filter in the time domain, the obtained filter characteristic has a discrete inverse Fourier transform (Inverse DiscreteFourier).
Transform) and filter in the time domain.

However, the transfer functions X and Y of the filters F _X and F _Y obtained by performing the operation on the right side of the equations (3) and (4) in the frequency domain and converting the operation result into the time domain are:
There was a problem that the causality was impaired and it became unrealistic.

FIG. 10 (a) shows a unit impulse signal input at time t = 0, and FIG. 10 (b) shows a response waveform to the unit impulse obtained based on equations (3) and (4). Is an example. As shown in FIG. 6B, the response waveform starts operating from a negative time. However, in practice, the filter cannot start the operation before the input time, so it is impossible to set the transfer functions X and Y obtained by this calculation as the filter characteristics as they are.

Therefore, conventionally, the transfer functions X and Y have been processed so as to maintain causality. For example, as shown in (c) of the figure, the waveform of the negative time is omitted, or as shown in (d) of the figure, the original sound is processed to delay the calculation start time, and then the filtering process is performed. It was

[0018]

However, such an operation means that a transfer function (for example, X ', Y') different from X, Y is used as the transfer function of the filters F _X , F _Y. Therefore, the above equations (1) and (2) do not hold, and there arises a problem that the listener cannot recognize the desired virtual sound source.

In other words, the listener is supposed to be XG with his left ear._L+ Y
H_LAnd listen to XG with the right ear_R+ YH_RBy listening to
Is the unit impulse radiated from the desired virtual sound source?
However, instead of these waveforms,
X'G_L+ Y'H_LAnd X'G _R+ Y'H_RListened to
So, with the desired virtual sound source position and the tone of the unit impulse
Cannot be accurately recognized. Therefore, the desired sound
Image localization cannot be realized.

Therefore, it is an object of the present invention to provide a sound image localization method and apparatus which can obtain a causal transfer function of a filter and faithfully reproduce a desired virtual sound source.

[0021]

FIG. 1 shows an explanatory view of the principle of the present invention. The invention according to claim 1 is a transfer function (g _L , g _R , h _L , h _R ) existing between each of a plurality of sound sources (A _X , A _Y ) and a listener P, and an arbitrary transfer function (z _L, z _R) existing between the virtual sound source SP _V and the listener P existing in the virtual position based on a predetermined filter to the audio signals of a plurality of sound sources (a _X, a _Y) A sound image localization method that causes a listener to recognize that sound waves are emitted from a virtual sound source by performing calculation processing based on the characteristic (X, Y). The all-pass component related to the division component of the prescribed arithmetic expression is a fixed value.

In the present specification, when the expression "transfer function existing between A and B" is used, a mathematical expression existing between a sound wave output from A and a sound wave transferred to B. A relationship is meant and a function for defining the state of the sound wave transmitted from A to B is transmitted.

The invention according to claim 2 is a plurality of sound sources (A
_X , A _Y ) and the transfer function (g _L , g _R , h _L , h _R ) respectively existing between the listener P and the virtual sound source SP _V existing at an arbitrary virtual position and the listener P. Based on the transfer functions (z _L , z _R ) existing between and, the audio signals of a plurality of sound sources (A _X , A _Y ) have predetermined filter characteristics (X,
Y) is a sound image localization method that causes a listener to recognize that sound waves are emitted from a virtual sound source, and the filter characteristic (X, Y) is an arithmetic expression that defines the filter characteristic. The all-pass component and the minimum phase component related to the division component of the above are both fixed values.

The invention according to claim 3 is the first sound source A _X.
And the transfer function existing between the left and right ears of the listener _{P (G L, G R)} , the transfer function (H _L that exists between the left and right ears of the second sound source A _Y and the listener P, H _R) and a transfer function that is present between the left and right ears of the virtual sound source a _V and the listener P at the position of an arbitrary virtual (Z _L, based on the Z _R),
By calculating based on the filter characteristics X to the audio signal of the first sound source A _X, is calculated based on the filter characteristics Y to the audio signal of the second sound source A _Y, from the virtual sound source A _V to the listener P In a sound image localization apparatus that recognizes sound waves as being emitted, the filter characteristics X and Y have a division component D (= G
When the _R H _L -G _L H _R) minimum phase related to the _{θ min, X = (Z R} H L -Z L H R) / (| D | exp (-jθ
_min )) Y = (Z _L G _R -Z _R G _L ) / (| D | exp (-jθ
_min )).

According to a fourth aspect of the invention, the first sound source A _X
And the transfer function existing between the left and right ears of the listener _{P (G L, G R)} , the transfer function (H _L that exists between the left and right ears of the second sound source A _Y and the listener P, H _R ) and a transfer function (Z _L , Z _R ) existing between the virtual sound source A _V existing at an arbitrary virtual position and the left and right ears of the listener P,
By calculating the sound signal of the first sound source A _X based on the filter characteristic X and the sound signal of the second sound source A _Y based on the filter characteristic Y, the virtual sound source A _V can be changed to the listener P. In a sound image localization apparatus that recognizes sound waves as being emitted, the filter characteristics X and Y have a division component of D (=
When a _{G R H L -G L H R} ), X = (Z R H L -Z L H R) / | D | Y = (Z L G R -Z R G L) / | D | and the Based on arithmetic expressions.

[0026]

The present invention will be described with reference to FIG. In the following description, the transfer function is expressed in lower case alphabets when expressed in the time domain, and in upper case alphabets when expressed in the frequency domain. Therefore, the transfer functions g _L , g _R , h _L , h _R , z _L , and z measured in the time domain
_R in the frequency _{domain, G L, G R, H} L, H R,
It is expressed as Z _L and Z _R.

As described in the section of the prior art, the following relationships are established from FIG. XG _L + YH _L = Z _L (1) XG _R + YH _R = Z _R (2) When formulas (1) and (2) are modified, formulas (3) and (4) are obtained.
Is obtained. _{X = (Z R H L -Z} L H R) / (G R H L -G L H R) ... (3) Y = (Z L H R -Z R H L) / (G R H L -G _L H _R) ... (4) conventionally converts equation (3) and (4) a representation of the time domain, has been set to the digital filter such as a FIR filter.

The present invention further modifies the following equation. Usually, the predetermined transfer function A (z) represented in the frequency domain is
It is divided into a minimum phase component A ₁ (z) and an all-pass component A ₂ (z). The all-pass component is a component whose amplitude characteristic is constant in all frequency bands, and is a component which adds only a phase without affecting the amplitude characteristic. The minimum phase component is a component that has the smallest phase transition and the smallest delay time of the response to the input among the transfer functions that realize the same amplitude characteristics as the transfer function A (z).

The all-pass component of the transfer function has a constant amplitude characteristic. Therefore, even if the all-pass component of another transfer function is added to the transfer function existing between the virtual sound source and the listener, the amplitude characteristic itself is not deformed. Therefore, it is presumed that there is no difference in the auditory perception of the sound image regardless of the presence or absence of the all-pass component. For the effect of all-pass components, see the paper “Perception of Ph
ase Distortion in All-Pass Filters "(by JADeer
and PJBloom; J.Audio Eng.Soc., Vol.33, No.10,1985 O
ctober), paper "Group delay distortions in electro
acoustical systems "(by J. Blauert; J. Acoust, Soc.A
m.63 (5), May 1978).

Based on these considerations, the present invention has been made paying attention to the fact that the all-pass component has almost no effect on the auditory perception in the calculation of the transfer function. Therefore, the denominator of the equations (3) and (4) is divided into the all-pass component and the minimum phase component, and the transfer function from the virtual sound source to the listener is obtained.

The right sides of the equations (3) and (4) are expressed by X = Xn / D ... (3) 'Y = Yn / D ... (4)' (Xn = Z_RH_L-Z_LH_R, Yn = Z_LH_R-Z_RH
_L, D = G_RH_L-G _LH_R). Where D is the all-pass component and minimum phase formation
Expressed by minutes and, D = | D | exp (-jθ_min) ・ Exp (-jθ_ap)… (5) Furthermore, the formulas (1) and (2) are expressed as (Xn / D) G_L+ (Yn / D) H_L= Z_L... (1) '(Xn / D) G_R+ (Yn / D) H_R= Z_R... (2) 'can be converted.

When both sides of the equations (1) 'and (2)' are multiplied by the all-pass component of D (equation (5)), the equations (1) 'and (2)' are

[0033]

[Equation 1] Becomes Equations (1) "and (2)" are essentially Xn as an alternative to the filter transfer functions X, Y in FIG.
/ | D | exp (-jθ _min ), Yn / | D | exp
This means that (-jθ _min ) is used. At this time, if a unit impulse is input, the listener will use the expression (1) ”in his left ear.
, And the waveform represented by the formula (2) "in the right ear. However, as described above, the all-pass component has almost no effect on the auditory sense. At Z _L exp (-θ _ap ), at right ear Z _R exp
You will recognize it as if you heard (-θ _ap ).

The transfer functions Xn / | D | exp (-jθ _min ) and Yn / | D | e obtained as described above.
Since it is known from the results of experiments that xp (-jθ _min ) has causality, it is possible to set and use these transfer functions in the filter without performing any processing operation. it can. The results of these experiments will be described later in Examples.

It is also known from the experimental results that the minimum phase component of the denominator on the left side of the equations (1) "and (2)" has no influence on the causality of the transfer function of the filter. Therefore, the equations (1) "and (2)" are modified to obtain the equations (1) and
^ (2) ^, and this Xn is used as the transfer function of the filter.
It is also possible to input / | D | and Yn / | D |.

[0036]

[Equation 2] However, in this case, the left and right ears of the listener, each Z _L ex
_{p (-j (θ ap + θ} mi n)), Z R exp (-j (θ ap
+ Θ _min )) is equivalent to listening to the sound wave waveform. For this reason, the sound field environment is limited so that Z _L and Z _R can be recognized in the same way as when they are heard. Therefore, in order to put the sound image localization into practical use by using the equations (1) ^ and (2) ^, a prior hearing test is required.

As described above, according to the first aspect of the present invention, the arithmetic processing for setting the all-pass component regarding the division component to a fixed value in the arithmetic expression defining the filter characteristic is performed. According to the second aspect of the present invention, the arithmetic processing is performed in which both the all-pass component and the minimum phase component related to the division component in the arithmetic expression defining the filter specification are fixed values.

According to the invention described in claim 3, the filter
Specific X and Y are divided components D (= G _RH_L-G_LH_R)
The minimum phase of θ_minThen, X = (Z_RH_L-Z_LH_R) / (| D | exp (-jθ_min)) ... (6) Y = (Z_LH_R-Z_RH_L) / (| D | exp (-jθ_min))

[0039] According to the invention described in claim 4, filter characteristics X and Y, the division component D (= G _R H _L -G
When the _{L H R), X = (} Z R H L -Z L H R) / | D | ... (7) Y = (Z L H R -Z R H L) / | D | and the arithmetic expression based on.

[0040]

EXAMPLE A preferred example of the sound image localization technique of the present invention will be described. The device according to the sound image localization technique of the present invention can basically apply the same device as the device used in the conventional sound image localization technique. That is, a desired head related transfer function is measured by using a device as shown in FIG. 9, the transfer function of the filter is obtained by the calculation according to the present invention, and this transfer function is converted into the sound image localization shown in FIG. 1 and FIG. It may be put in the filter of the device. In this embodiment, a device for reproducing a sound image localization immediately after the measurement of the head related transfer function will be described.

FIG. 2 shows the configuration of the sound image localization apparatus of the embodiment. In FIG. 2, the configuration indicated by the broken line relates to the configuration for measuring the transfer function of the space from the virtual sound source to the listener P.

The virtual sound source measuring circuit 2 supplies a predetermined signal (impulse signal) to the speakers SP _V1 , SP _V2 , SP _L and SP _R at the time of measuring the transfer function of the virtual sound source, and the microphones M _R and M _L Based on the input detection signals S _DR and S _DL , the instruction signals S _CX and S of the specified transfer function
_CY (transfer function _{G R, G L, H R} , signal instructing H _L)
To the filter circuits F _X and F _Y.

The microphones M _R and M _L are installed in the left and right ears of the listener P when measuring the transfer function of the virtual sound source, and detect sound waves from the speakers SP _V1 , SP _V2 , SP _L and SP _R.

The speaker SP _V1 or SP _V2 is installed at the localization of the virtual sound source required when reproducing the virtual sound image. Therefore, it is installed at an arbitrary position in the sound field surrounding the listener P. The speakers SP _L and SP _L are installed at positions where the original sound is actually reproduced.

The original sound generation circuit 1 outputs the original sound signal S which becomes the original sound.
Generate _L and S _R. The filter circuits F _X and F _Y are
Based on the transfer function of the virtual sound source specified by the virtual sound source measurement circuit 2, each of the original sound signals S _L and S _R is filtered and supplied to the speakers SP _L and SP _R.

The transfer function is measured by a dummy head imitating a typical human head in addition to being detected by a microphone provided in the vicinity of both ears of an actual listener as shown in the figure. You may detect by the microphone provided in the back. Since the transfer function is also affected by the human external ear and other organs, the dummy head enables highly accurate measurement.

In the above configuration, transmission related to virtual sound source
When measuring the function, the virtual sound source measurement circuit 2
Then, at the position where the listener P wants to recognize the speaker SP_V1Or
SP _V2Is installed to supply the unit impulse signal to the sound field.
You. The unit impulse signal is affected by the reverberation characteristics of the space.
Received a sound and performed waveform conversion with a predetermined transfer function, etc.
It is detected by the microphone as a valuable signal. For example, a speaker
SP_V1For the transfer function z_R1And z_L1, Speaker SP
_V2For the transfer function z_R1And z_L1Affected by. Temporary
The ideal sound source measurement circuit 2 uses the obtained transfer function in a predetermined format.
Filter circuit F_XOr F_YTo supply.

At the time of reproducing the original sound, the filter circuits F _X and F _Y perform the filter processing described in the section of the operation on the original sound signals S _L and S _R supplied from the original sound generation circuit 1. That is, each transfer function measured in the time domain is converted into an expression in the frequency domain and filtered. At this time, equation (6)
May be converted into the time domain to perform the filtering process based on the equations (1) ″ and (2) ″. Further, the filter processing based on the equation (1) ^ and the equation (2) ^ may be performed by using the filter coefficient obtained by transforming the equation (7) into the time domain.

3 to 7 show the measurement results by the sound image localization apparatus of this embodiment. FIG. 3 shows the amplitude characteristic of the right channel (system of the filter circuit F _Y ) and FIG. 4 shows the amplitude characteristic of the left channel (system of the filter circuit F _X ). In both figures, (a) is the characteristic by the conventional filter processing including the all-pass component, and (b) is the characteristic by the present invention in which the all-pass component is a fixed value.

As can be seen by comparing (a) and (b) of each figure, there is almost no change in the amplitude characteristic regardless of the presence or absence of the all-pass component. In FIG. 5, the ratio of the amplitude values and the ratio of the phase angles of the left and right channels are calculated.

As can be seen from FIG. 5, the amplitude characteristics and the phase characteristics do not change depending on whether or not the all-pass component is calculated, even when viewed from the output ratios of both channels. 6 and 7
Is a comparison between the conventional impulse response and the impulse response according to the present invention. By the way, the impulse response waveform in the figure shows an output waveform when a unit impulse is input at t = 0.00. 6 and 7
Is an impulse response of each of the left and right channels, which is obtained by using different input signals. In both figures, (a) is the conventional right channel (Xn / D),
(B) is the right channel (Xn / D _min ) of the present invention,
(C) shows the characteristics of the conventional left channel (Yn / D), and (d) shows the characteristics of the left channel (Yn / D _min ) of the present invention.

As can be seen from the respective figures, in the conventional filter operation including the all-pass component, the output operates from a negative time, and the impulse response waveform is non-causal (each figure (a), (C)). On the other hand, in this example, it is understood that the causality is maintained ((b), (d)).

As described above, according to this embodiment, the causality is maintained for the input and the desired sound image localization is maintained without essentially changing the transfer function obtained by actual measurement. It is possible to obtain a filter that can realize In particular, the transfer function of the filter used in this embodiment can maintain a constant amplitude ratio and phase ratio received by the left and right ears, so that sound waves are emitted from a desired virtual sound source,
The listener can surely recognize a stable sound image.

[0054]

According to the invention described in claim 1 or 3, in order to realize a desired sound image localization, a filter that is causal to the input and does not perform any processing operation is used. Since it can be easily obtained by calculation with the partial transfer function, a desired sound image localization can be realized easily and more reliably than in the conventional case.

According to the invention described in claim 2 or claim 4, in addition to the effect of claim 1 or claim 3, the effect of further simplifying the calculation can be obtained, so that the sound image localization is realized. It is possible to provide a sound field reproduction technology that emphasizes cost and labor reduction.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a configuration diagram of a sound image localization device according to an embodiment.

FIG. 3 is an amplitude characteristic (right channel) of the example.

FIG. 4 is an amplitude characteristic (left channel) of the embodiment.

FIG. 5 is a ratio of amplitude and phase angle between left and right channels.

FIG. 6 is an impulse response characteristic (No. 1).

FIG. 7 is an impulse response characteristic (No. 2).

FIG. 8 is an explanatory diagram showing waveforms of a conventional sound image localization device and each part.

FIG. 9 is an explanatory diagram of a method of measuring a head related transfer function.

FIG. 10 is an explanatory diagram of a problem of the conventional sound image localization device.

[Explanation of symbols]

A _X, A _Y ... source A _Z ... virtual sound source F _X, F _Y ... filter circuit _{(means) SP L, SP R, SP} V1, SP V2 ... speaker 1 ... original sound generator 2 ... virtual sound measurement circuit

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Mikio Higashiyama 2665-1 Nakano-cho, Hachioji-shi, Tokyo Inside Kogakuin University Hachioji Campus (72) Inventor Norihisa Hirata 2568-9 Ishikawa-cho, Hachioji-shi, Tokyo (72) Inventor Eitoshi Nakajima 2665-1 Nakano-cho, Hachioji-shi, Tokyo Inside Kogakuin University Hachioji campus (72) Inventor Kiyoshi Tanijima 6-1-1 Fujimi, Tsurugashima City, Saitama Pioneer Corporation

Claims (4)

[Claims] 1. A transfer function existing between each of a plurality of sound sources and a listener and a transfer function existing between a virtual sound source existing at an arbitrary virtual position and the listener, based on the transfer function. A sound image localization method that causes the listener to recognize that sound waves are emitted from the virtual sound source by performing a calculation process based on a predetermined filter characteristic on audio signals regarding a plurality of sound sources, wherein the filter characteristic is A sound image localization method, wherein an all-pass component related to a division component of an arithmetic expression defining a filter characteristic is set to a fixed value. 2. A transfer function existing between each of a plurality of sound sources and a listener and a transfer function existing between a virtual sound source existing at an arbitrary virtual position and the listener, based on the transfer function. A sound image localization method that causes the listener to recognize that sound waves are emitted from the virtual sound source by performing arithmetic processing based on a predetermined filter characteristic on original sound signals of a plurality of sound sources, wherein the filter characteristic is A sound image localization method characterized in that both the all-pass component and the minimum phase component relating to the division component in the arithmetic expression defining the filter characteristic are fixed values. 3. A transfer function existing between the first sound source and the listener's left and right ears (G _L, G _R), exists between the second sound source and the listener's left and right ears transfer function (H _L, H _R) and a transfer function that is present between the left and right ears of the virtual sound source and the listener at the position of an arbitrary virtual (Z _L, Z _R) on the basis of said first The sound signal related to the sound source is calculated based on the filter characteristic X, and the sound signal related to the second sound source is calculated based on the filter characteristic Y, so that the listener can recognize that sound waves are emitted from the virtual sound source. In the sound image localization device, the filter characteristics X and Y have a division component D (= G _R H _L
When the minimum-phase component regarding -G _L H _R) and _{θ min, X = (Z R} H L -Z L H R) / (| D | exp (-jθ
_min )) Y = (Z _L G _R -Z _R G _L ) / (| D | exp (-jθ
_min )) based on the calculation formula. 4. A transfer function existing between the first sound source and the listener's left and right ears (G _L, G _R), exists between the second sound source and the listener's left and right ears transfer function (H _L, H _R) and a transfer function that is present between the left and right ears of the virtual sound source and the listener at the position of an arbitrary virtual (Z _L, Z _R) on the basis of said first The sound signal related to the sound source is calculated based on the filter characteristic X, and the sound signal related to the second sound source is calculated based on the filter characteristic Y, so that the listener can recognize that sound waves are emitted from the virtual sound source. In the sound image localization device, the filter characteristics X and Y have a division component of D (= G _R H
When the _{L -G L H R), X} = (Z R H L -Z L H R) / | to the the arithmetic expression _{| D | Y = (Z L} G R -Z R G L) / | D A sound image localization device characterized by being based.