JPH0681360B2 - Sound reproduction device - Google Patents

Description

The present invention relates to a sound reproducing device that allows a plurality of listeners to perceive a sound image at an arbitrary position.

A conventional sound reproducing device for localizing a sound image at an arbitrary position for a listener will be described with reference to FIGS. 1 and 2. From the speaker (1a) (1b) to the auricle (2) of the listener (2)
a) Transfer functions up to (2b) are set to H _Rr , H _Rl , H _LR , and H _Ll, and the virtual specific speaker (3) at the position where the sound image is localized is detected from the auricle (2a) (2b) of the listener (2). ) until the transfer functions H _{[phi] r,} if _H φl, loudspeaker input signal (1a) (1
By applying the same sound pressure to both ears of the listener (2) when applied to the input terminal (4) of b) and when applied to the input terminal (5) of the specific speaker (3). Is derived, and if this equation is solved for E _R and E _L , Is uniquely determined.

However, such a conventional device has the most fundamental and decisive drawback that the listener (2) is limited to one. Further, the circuit that realizes E _R and E _L satisfying the above equation becomes very complicated and cannot be specified precisely. For example E from the equation _R, seeking E _L, be derived the circuit configuration shown in FIG. 2 as an equalizer (6a) (6b) to organize, to each of the block diagram (7) - (10) At least one delay circuit and at least one frequency equalizer are required, and the setting of the delay time and the frequency equalizer of each delay circuit becomes complicated and the accuracy is limited.

The present invention solves the above drawbacks, and one embodiment thereof will be described below with reference to the drawings.

In FIG. 3, (8a) to (8d) are speakers, (9) is an input terminal of the speakers (8a) to (8d), and (10a) to (10d).
Is an equalizer, (11) is a virtual specific speaker at a position for sound image localization, (12) is an input terminal of the specific speaker (11), (13) and (14) are listeners, (13a) and (13b) are listeners. Pinna of (13), pinna (14a) (14b) are pinna of the listener (14). From each speaker (8a) to (8d) to each listener (13)
The transfer function up to the pinna (13a) (13b) (14a) (14b) of (14) is determined as follows. For example, the transfer function from the speaker (8a) to the right ear of the listener (13) is H _Ar . Similarly, the first letter of the suffix of H represents an equalizer connected to the front stage of the speaker, and the second letter of the second letter is r, l for the right listener (13) and the left listener (14 ) Are represented by R and L. Now, when the input signal S is applied to the input terminal (9), the sound pressures P _r , P _l , P _R , and P _L of the ears of the listeners (13) and (14) are as follows. On the other hand, a specific speaker (11) that localizes the input signal S as a sound image
When applied to the input terminal (12) of each listener (13) (1
The sound pressure P _r ′, P _l ′, P _R ′, P _L ′ of both ears in 4) is expressed by the subscript φ at the position of the specific speaker (11). Here, with respect to the binaural sound pressures of the listeners (13) and (14), the above equations are equal to each other, that is, P _r ′ = P _r , P _l ′ =
If P _l , P _R ′ = P _r , P _L ′ = P _L And solve for A, B, C, D It can be specified uniquely.

However, as described in the explanation of the conventional example, the above formula is satisfied.
Since the circuit that realizes the equalizer of A, B, C, D becomes very complicated, we consider to realize the equalizer of A, B, C, D as a discrete convolver in the time domain. That is, when the transfer function is expressed as an impulse response composed of n discrete sample sequences and the input signal S is a unit impulse, However, the following matrix is created from the impulse response time series data corresponding to each transfer function.

Moreover, the impulse response of the convolver, that is, the tap coefficient is determined as follows.

Furthermore, the impulse response in both ears corrected by the convolver is determined as follows.

Furthermore, the above equation is equivalent to This is expressed by the following formula.

P = h ・ g ... Similarly, from the above equation, if the input signal is a unit impulse, However, P _X ′, h _Y ′ is a matrix obtained by adding m−1 0s to the impulse response sample sequence.

Further, the above equation is expressed as P ′ = h _φ far.

Here, the evaluation function f is determined from P and P ′ as follows, and g that minimizes f is obtained.

^{f = (P-P ')} t (P-P') = (hg-P ') t (hg-P') = g t h t hg- (g t h t P '+ P' t hg) + P ' ^t
P '... Therefore, to obtain the extreme value of f, partial differentiation with g That is, f becomes the minimum when h ^t hg = h ^t P '... Therefore, g = (h ^t h) ^-1 h ^t P '... and g can be obtained by measuring h and P'.

That is, the procedure for obtaining the impulse response of the convolver, that is, the tap coefficient g is as follows.

First, a sequence of impulse response samples from the speakers (8a) to (8d) in FIG. 3 to the ears of the listeners (13) and (14) is measured.

An h matrix is created from the above equation according to the equation and the equation. Ie Next, measure the impulse response sample string from the specific speaker (11) in FIG. A P'matrix is created from the above equation.

Furthermore it is possible to obtain the g being tap coefficients of the convolver be calculated according to the above formula, for example, the coefficients of a clogging a ₁ ~a _m of g coefficient units of the circuit shown in FIG. 4 (15 ₁₎ to ( 5 _m ), the equalizer (10a) of A in FIG. 3 is realized. FIG. 4 is a circuit block diagram of the convolver, (16) is an input terminal,
(17) an output terminal, (18 ₁₎ ~ (18 _m) of the delay element, (2
8) an adder, said delay elements _{(18 1) ~ (18 m} ) tapped BBD or CCD, which is integrated on the same chip as the
Etc. are used.

Similarly, b _{1 to} b _m , c ₁ to _cm , d _{1 of} g obtained
The coefficients of the to d _m Figure 4 of the coefficient unit (15 ₁₎ to the realized by (15 _m) Figure 3 of the B, C, D of the equalizer (10b) (10c) (1
0d) is realized and each listener (13) when a signal is input to the input terminal (9) and when a signal is input to the input terminal (12)
The sound pressures of both ears of (14) are the same, and even when a signal is input to the input terminal (9), each listener (13) (14) perceives a sound image at the position of the specific speaker (11).

Next, as a second embodiment, an example in which the specific speaker (11) in FIG. 3 is set at the same position relative to the listeners (13) and (14) will be described with reference to FIG. In this example, the listener (13) is at the position of the specific speaker (11a),
The listener (14) perceives a sound image at the position of the specific speaker (11b). That is, by applying the same binaural sound pressure to the listeners (13) (14), the listeners (13) (14) can perceive exactly the same sound field. The method of setting the equalizers (10a) to (10d) is the same as that of the first embodiment, and in the above equation, P _r ′ = P _R ′, P _l ′ = P _L ′ or h _φr = h.
_φR , h _φl = _hφL .

Next, a third embodiment will be described. In this example, the first
In the above embodiment, a matrix in which the above equation or the impulse sequence of the equation is shifted by one sample is created, and g is obtained by using it as a new P ′ or h _φ . That is, for example, if two samples are shifted in the above equation, Then, by substituting this into h _φ , another solution of g can be found. The operations for shifting the sample are sequentially performed. G for each of them, g of each
Then, the evaluation function f of the above formula is calculated, and g when f is minimized is taken as the optimum solution.

Next, a fourth embodiment will be described with reference to FIG. Sixth
In the figure, (19a) and (19b) are input terminals, and (20a) to (20
h) is an equalizer by convolver, (21a) ~ (21d)
Is an adder, (22a) to (22d) are speakers, (23a) is a first specific speaker, (23b) is a second specific speaker, and (2
4a) is an input terminal of the first specific speaker (23a), (2a)
4b) is an input terminal of the second specific speaker (23b). The binaural sound pressure of each listener (13) (14) due to the signal input to the input terminal (19a) is equalized (20a) to (20d).
Causes the same sound pressure as when a signal is input to the input terminal (24a) of the first specific speaker (23a), and each listener (13) (14) by the signal input to the input terminal (19b) in the same manner. )
The binaural sound pressure is set to the same sound pressure as when a signal is input to the input terminal (24b) of the second specific speaker (23b) by the equalizers (20e) to (20h). In this way, the same operation as in the first embodiment can be performed for two channels or more input channels.

Next, a fifth embodiment will be described with reference to FIG. 7th
In the figure, (25 ₁ ) to (25 _k ) are k listeners and (2
6 ₁ ) to (26 _2k ) are 2k speakers, (27 ₁ ) to (27 _2k )
Are 2k equalizers. This example, the first embodiment k people of the listener thinking extended to the _{(25 1) ~ (25 k} )
2k convolver equalizers (27 ₁ ) to (27 _2k ) are provided to control the. When setting the equalizers (27 ₁ ) to (27 _2k ), the difference from the first embodiment is that the 2-row 2-column matrix of the above equation and the like becomes only 2k rows and 2k columns, and the coefficient matrix of the convolver that is an unknown number. g ₁
If all of ~ g _2k are obtained and set in the coefficient unit of the convolver configuring the equalizers (27 ₁ ) to (27 _2k ), the input terminals ( _{k 1} ) to (25 ₁ ) to (25 _k ) can be _obtained. k people every listener by applying an input signal to the 9) (25 ₁₎ to
It is possible to make (25 _k ) perceive a sound image at the position of the specific speaker (11).

As described above, according to the sound reproducing device of the present invention, a plurality of listeners can each perceive a sound image at an arbitrary position. Further, a convolver having a simple structure can be used as the equalizer. In this case, the precision is defined by the number of taps of the convolver and the number of impulse response measurement samples, and high precision can be easily obtained.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram of a conventional device, FIG. 2 is a configuration diagram of an equalizer in the conventional device, FIGS. 3 and 4 show an embodiment of the present invention, FIG. 3 is an overall configuration diagram, and FIG. FIG. 7 is a block diagram of a convolver, and FIGS. 5 to 7 are overall block diagrams showing different embodiments. (8a) ~ (8d) ( 22a) ~ (22d) (26 1) ~ (26 2k) ...
… Speakers (10a) to (10d) (20a) to (20h) (2
7 ₁ ) to (27 _2k ) ... Equalizer, (11) (11a) (11b)
(23a) (23b) …… Specific speaker, (13) (14) (2
5 ₁ ) ~ (25 _k ) ... listener

Claims (2)

[Claims] 1. A convolutional integration circuit, comprising 2n convolutional integration circuits each of which is formed by connecting a multiplier to the output of each of m cascaded delay elements and adding the outputs of the respective multipliers. Input in parallel, the output of each convolutional integration circuit is input to each of 2n speakers, and the matrix h having the impulse response from each speaker to both ears of the listener as a column and its transposed matrix h ^t and the desired sound image localization 'g = by and ^{(h t h) -1 h t} P' impulse response matrix P to a column of by a particular speaker installed in the position to both ears of the listener the values of elements of the matrix g represented by A sound reproducing device set as a coefficient value of each multiplier of the convolutional integration circuit. 2. The position of the specific speaker is set to a relatively equal position for each listener.
The sound reproducing device according to the item.