JPH08102999A - Stereophonic sound reproducing device - Google Patents

Abstract

PURPOSE: To reproduce a stereoscopic sound at a level satisfied even in a general space having reverberation by interposing a transfer function correction filter and an LPF in parallel between a reproduction sound source and plural speakers. CONSTITUTION: Acoustic information outputted from a reproduction sound source 16 is fed to plural speakers 4L, 4R via a transfer function correction filter 17 and fed to the speakers 4L, 4R via LPFs 18L, 18R. Since the transfer function of the filter 17 is set based on a high frequency component of an impulse response of the acoustic reproduction system, only the high frequency component of the acoustic information passes through the filter 17 and then is fed to the speakers. On the other hand, the acoustic information fed to the LPFs 18L, 18R is fed to the speakers after its high frequency component is eliminated. Thus, number of taps of the filter 17 is decreased without sacrificing the reproduction of the low frequency component and a satisfactory stereophonic sound is reproduced with inexpensive configuration.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention records a sound in an original sound field using, for example, a dummy head microphone, and uses this as a sound source for reproduction, and reproduces this sound source for reproduction by a plurality of speakers arranged in a reproduction acoustic space. By doing so, the present invention relates to a stereophonic sound reproducing device that allows a person in a reproducing sound space to hear a stereoscopic sound having a realistic sensation as if he or she were in the original sound field. A transfer function correction filter that is interposed between a speaker and a sound reproduction system and corrects a transfer function of the sound reproduction system can be realized by a small-capacity digital filter without lowering reproduction characteristics. Incidentally, the transfer function correction filter, by filtering the acoustic information supplied from the reproduction sound source of the stereophonic reproduction apparatus to the plurality of speakers,
It is a filter for appropriately correcting the transfer function of the acoustic transmission system (sound reproduction system) from the speaker to the human ear position in the reproduction acoustic space to remove crosstalk.

[0002]

2. Description of the Related Art As a conventional technique of this kind, for example, JP-A-5-115098 and JP-A-5-227600.
There are some disclosed in Japanese Patent No. These conventional techniques are applied to a free sound field such as a completely anechoic room, and the speaker is placed in the free sound field so that the positional relationship is similar to the positional relationship between the speaker and the human at the time of reproduction. And a dummy head microphone are installed, the transfer characteristic between the speaker and the dummy head microphone (the transfer characteristic of the sound reproduction system) is measured, and the filter coefficient of the transfer function correction filter is calculated based on the measurement result to transfer the transfer function. The correction filter is designed, and when the transfer function correction filter is designed, it is interposed between the reproduction sound source and the speaker. On the other hand, the reproduction sound source records acoustic information by a binaural recording method by recording using a dummy head microphone in an original sound field such as a concert hall.

Then, by supplying acoustic information from the reproducing sound source to each speaker via the transfer function correction filter, the sound pressure at the binaural position of the dummy head microphone is set to the binaural position of the human being in the reproducing acoustic space. I was able to reproduce a stereoscopic sound with a realistic sensation. Here, the design of the transfer function correction filter in the above conventional technology is described in Tokyo Institute of Technology Research Institute Report Vol.2, No.2, 198.
3 “Orthosterephonic for the purpose of standard sound collection and reproduction
Acoustic equivalent circuit (p.13)
~ 18) has been applied. This is because the measured transfer characteristics (impulse response) between the speaker and the dummy head microphone are Fourier-transformed to construct an acoustic equivalent circuit in the frequency domain, and then the transfer function constituting the equivalent circuit is inverse-Fourier-transformed to obtain FIR. It was designed to design filters (finite impulse response filters).

[0004]

Here, the above-mentioned conventional technique is premised on application in a free sound field, which is a very limited space such as a completely anechoic room or the outdoors.
It was not practically applicable to a general indoor space with reverberation. In other words, in such a general indoor space, even if the impulse response between the speaker and the dummy head microphone is measured as described above, the impulse response does not reflect the reflection on the floor or wall surface of the space. Since the wave component is included, if the transfer function correction filter is designed according to the impulse response including such a reflected wave component, the reflected wave component has a bad effect in the stage of creating the filter coefficient, As a result, the localization and sound quality of the reproduced sound is significantly impaired.

Therefore, if it is attempted to sufficiently reproduce in the low frequency band, the impulse response becomes long due to the influence of the reverberation component, and the number of taps is large (the number of filter coefficients is large).
The FIR filter is required, the amount of calculation is increased accordingly, and there arises a problem that a high-speed calculation is required and thus an expensive microprocessor is required.

Further, even if the reproduced sound field has a small effect of reverberation, in order to obtain a sufficient reproduction characteristic up to a low frequency band, a transfer function correction filter should be provided by sufficiently considering the low frequency component of the impulse response. If it is designed, an FIR filter with a large number of taps is required, the amount of calculation is increased, and an expensive microprocessor is required.

Further, even if the transfer function correction filter is designed with a sufficient length on the time axis, if the number of taps is large, the accuracy of the numerical value on the rear side of the impulse response is deteriorated due to a slight phase shift. Will end up. In other words, the initially determined filter coefficient will no longer be effective, which will adversely affect the reproduction of the sound image, and it will be necessary to redesign the filter in parallel with the control in order not to reduce the accuracy. .

The present invention has been made by paying attention to the unsolved problems of the above-mentioned prior art and the technical requirements as described above, and the stereoscopic image is at a satisfactory level even in a general space having reverberation. It is an object of the present invention to provide a stereophonic sound reproducing device that can reproduce sound, can reproduce low frequency components satisfactorily, and is inexpensive.

[0009]

In order to achieve the above object, a stereophonic sound reproducing apparatus according to a first aspect of the present invention comprises a reproducing sound source for outputting acoustic information for reproducing stereoscopic sound and the acoustic information. It is provided with a plurality of input speakers, and is set between the reproduction sound source and the plurality of speakers based on a component having a frequency higher than a predetermined frequency of an impulse response between the plurality of speakers and a predetermined sound receiving position. In addition, a transfer function correction filter that corrects the transfer function of the sound reproduction system and a low-pass filter are interposed in parallel.

According to a second aspect of the present invention, in the stereophonic sound reproducing device according to the first aspect of the present invention, the predetermined frequency and the cutoff frequency of the low pass filter are matched or substantially matched. . The invention according to claim 3 is the stereophonic sound reproducing device according to claim 1 or 2, wherein delay means is provided in series with the transfer function correction filter or the low-pass filter, and the transfer function correction is performed. The transmission time of the acoustic information between the reproduction sound source and the plurality of speakers via the filter and the transmission time of the acoustic information between the reproduction sound source and the plurality of speakers via the low-pass filter match. Then, the delay time of the delay means was set.

Further, the invention according to claim 4 is the stereophonic sound reproducing device according to any one of claims 1 to 3,
An equalizer was provided in series with the low pass filter.
Furthermore, the invention according to claim 5 is the above-mentioned claims 1 to 4.
In the stereophonic sound reproducing device according to the invention described above, the attenuation characteristic when extracting a component having a frequency higher than a predetermined frequency of the impulse response and the attenuation characteristic of the low-pass filter are set to gentle characteristics.

[0012]

According to the first aspect of the invention, the acoustic information output from the reproduction sound source is supplied to the plurality of speakers through the transfer function correction filter, and also transmitted through the low pass filter. Supplied to the speaker. Since the transfer function correction filter is set based on the high frequency component of the impulse response of the sound reproduction system, only the high frequency component of the acoustic information passes through the transfer function correction filter and is supplied to the plurality of speakers. Further, when the transfer function correction filter is set based on the high frequency component, the filter length is shorter than that when the transfer function correction filter is set to include the low frequency component as well, so the amount of calculation in the transfer function correction filter is small.

On the other hand, the acoustic information supplied to the low-pass filter is supplied to the plurality of speakers after the high frequency components are removed. That is, the low frequency component of the acoustic information is reproduced by the same reproduction system as that of the conventional stereo reproduction. Here, the stereophonic sound reproducing device is mainly intended to reproduce the sound image of the original sound field recorded in binaural. However, when the human characteristics are examined, the localization of the sound image is mainly determined in the audible frequency range (usually, 20Hz-20kHz
It has been found to be in a high frequency region (for example, 8 kHz or higher) of the (degree).

Therefore, as in the invention according to claim 1, high-frequency components important for localization of a sound image are corrected by a transfer function correction filter and supplied to the speaker.
When the low-frequency component that makes a small contribution to the localization of the sound image is reproduced by the low-pass filter and supplied to the speaker, the number of taps of the transfer function correction filter becomes small without sacrificing the reproduction of the low-frequency component. In other words, the stereophonic sound is reproduced with an inexpensive structure at a satisfactory level even in a general space having reverberation.

According to the second aspect of the invention, since all the frequency components of the acoustic information are supplied to the plurality of speakers through either the transfer function correction filter or the low pass filter, the acoustic information The reproduction of all frequency components is guaranteed. Here, there is no particular problem if the filter processing times of the transfer function correction filter and the low-pass filter match, but if they do not match, the high frequency component and the low frequency component of the acoustic information deviate on the time axis. Will be replayed. Therefore, if the delay means is provided in series with one of the filters and the delay time of the delay means is appropriately set as in the invention according to claim 3, the acoustic information passing through the transfer function correction filter and the low-pass filter are provided. Since the time difference from the acoustic information passing through is corrected and supplied to the speaker at the same time, the reproduced sounds of both reproducing systems reach the predetermined sound receiving position at the same time.

Further, according to the invention of claim 4, the low frequency component of the acoustic information is supplied to the plurality of speakers after the frequency characteristic is changed by the equalizer. Therefore, if the characteristic of the equalizer is set to be the inverse characteristic of the low-pass characteristic due to the transfer characteristic between the reproduction sound source and the predetermined sound receiving position through the low-pass filter, for example,
The frequency characteristic of the signal reproduced by the low-pass filter becomes flat or substantially flat.

In the invention according to claim 5, since the attenuation characteristic is made gentle rather than steep, for example, even when the sound source of a single frequency sweeps and the frequency of the original sound changes, the cutoff frequency is exceeded. Sometimes the sound image is prevented from changing discontinuously and natural reproduction is performed.

[0018]

Embodiments of the present invention will be described below with reference to the drawings. 1 to 6 are diagrams showing a first embodiment of the present invention, FIG. 1 is a block diagram showing a configuration of a transfer function correction filter design device 1, and FIG. 5 shows a configuration of a stereophonic sound reproducing device 15. It is a block diagram.

First, the structure of the transfer function correction filter design device 1 will be described. The transfer function correction filter design device 1 includes an M sequence signal generation device 2 for generating and outputting a white noise signal x and a reproduced sound field 3. A pair of loudspeakers 4L and 4R, which are appropriately spaced apart from each other on the wall surface, and microphones 5L and 5R, which are placed so as to face the loudspeakers 4L and 4R from the front and which come into contact with human ears.
And a dummy head 6 simulating a human head in which is embedded.

Amplifiers 7L and 7R are arranged on the input sides of the speakers 4L and 4R, and the input sides of these amplifiers 7L and 7R are connected via the changeover switch 8 to the M-sequence signal generation device 2
Is connected to the output side of. That is, by appropriately switching the changeover switch 8, the white noise signal x is supplied to either one of the speakers 4L and 4R.

The speakers 4L and 4R and the amplifier 7
L and 7R are the same as those used in the stereophonic sound reproducing device 15 to be described later, and the positions of the speakers 4L and 4R are also the same as those in the stereophonic sound reproducing device 15. Then, the dummy head 6 is arranged at the same position as the position of the human head when reproducing the stereophonic sound in the reproduction sound field 3 later.

On the other hand, the output sides of the microphones 5L and 5R are connected to a transfer characteristic measuring device 10 via a changeover switch 9. That is, by appropriately switching the changeover switch 9, the sound pressure measured by either one of the microphones 5L and 5R is supplied to the transfer characteristic measuring device 10 as the sound pressure measurement signal y. Therefore, in the transfer characteristic measuring device 10, by properly switching the changeover switches 8 and 9, the white noise signal x emitted from the left speaker 4L and received by the left microphone 5L, and emitted from the left speaker 4L. White noise signal x received by the right microphone 5R, white noise signal x emitted by the right speaker 4R and received by the left microphone 5L, and white noise signal x emitted by the right speaker 4R and received by the right microphone 5R. The white noise signal x is selectively input as the sound pressure measurement signal y.

The white noise signal x is also supplied from the M-sequence signal generator 2 to the transfer characteristic measuring device 10. Then, the transfer characteristic measuring device 10 uses the white noise signal x and the sound pressure measurement signal y that is selectively input as described above based on the speakers 4L and 4R.
And four types of impulse responses between each of the microphones 5L and 5R are measured.

Here, the transfer characteristic measuring device 10 of the present embodiment.
Are configured as shown in FIG. That is, the transfer characteristic measuring apparatus 10 is actually configured to include an interface circuit such as a microcomputer and an A / D converter, and functionally has I filter coefficients W _i (i =
0,1,2, ..., and the adaptive digital filter W of each filter coefficient W _i is a variable consists I-1), each filter coefficient W _i of the adaptive digital filter W according LMS algorithm is one of the adaptive algorithm It has a filter coefficient updating unit 10A for updating and an adding unit 10B. Then, the result of filtering the white noise signal x with the adaptive digital filter W and the sound pressure measurement signal y are input to the addition unit 10B.
The addition result of the addition unit 10B is input to the filter coefficient update unit 10A as an error signal e.

The white noise signal x is also input to the filter coefficient updating unit 10A, and the filter coefficient updating unit 10A uses the white noise signal x and the error signal e based on the adaptive digital filter W according to the LMS algorithm. The filter coefficient W _i is updated. That is, the LMS algorithm executed by the filter coefficient updating unit 10A has the evaluation function E [e
² (n)] is an algorithm for sequentially updating the filter coefficient W _i of the adaptive digital filter W so as to minimize [ ² (n)], and its concrete updating formula is as shown in the following formula (1).

W _i (n + 1) = W _i (n) −αx (n) e (n−i) (1) However, the terms with (n) and (n + 1) are at discrete times n and n + 1, respectively. Is a value, and α is a convergence coefficient _{relating to} the convergence stability and convergence speed of the filter coefficient W _i . In other words, the sound pressure measurement signal y is a signal that the white noise signal x is emitted from the speaker 4L or 4R and measured by the microphone 5L or 5R.
Theoretically, when the error signal e becomes zero, the adaptive digital filter W identifies the impulse response between the speakers 4L, 4R and the microphones 5L, 5R. The impulse response between the speakers 4L and 4R and the microphones 5L and 5R represented by the adaptive digital filter W has a waveform on the time axis as illustrated in FIG. 3, for example.

Here, the impulse response can be divided into a direct sound component, an initial reflected sound component, and a reverberation sound component when viewed on the time axis. Of these, the direct sound component is the component of the sound that directly reaches the microphones 5L, 5R from the speakers 4L, 4R, and the initial reflected sound component is the speaker 4
L, 4R and microphones 5L, 5R are the components of the sound that reaches the microphones 5L, 5R after being reflected one or more times on the floor or wall surface. The reverberation sound component is the microphone after repeating reflections many times. It is the component of the sound that reaches 5L and 5R.

Then, the transfer characteristic measuring device 10 appropriately changes the changeover switches 8 and 9 so that the four combinations of impulse responses h ₀₀ , h ₀₁ , h ₁₀ and h of the above-mentioned combinations are obtained.
₁₁ be adapted to measure, they measured impulse response h ₀₀ to h ₁₁ is adapted to be supplied to the filter coefficient calculation unit 11. Filter coefficient calculation device 1
1 is also actually composed of a microcomputer and necessary interface circuits, and the transfer characteristic measuring device 10
Based on the impulse responses h _{00 to} h ₁₁ supplied from
The processing for calculating the filter coefficient of the transfer function correction filter of the stereophonic sound reproducing device is executed.

FIG. 4 is a flow chart showing an outline of processing executed in the filter coefficient calculation device 11. That is, reading the impulse response h ₀₀ to h ₁₁ supplied from the transfer characteristic measurement device 10 in step 101, then the process proceeds to step 102, the impulse response h ₀₀ to h ₁₁
Are subjected to Fourier transform to obtain impulse responses H ₀₀ , H ₀₁ , H ₁₀ and H ₁₁ in the frequency domain.

Next, in step 103, the filter coefficients A ₀ and A ₁ of the transfer function correction filter in the frequency domain are calculated according to the following equations (2) and (3). A ₀ = -H ₀₁ / H ₀₀ (2) A ₁ = -H ₁₀ / H ₁₁ (3) Then, the process proceeds to step 104, and the following formula (4),
Other filter coefficients B ₀ and B ₁ in the frequency domain are calculated according to the equation (5).

B ₀ = 1 / {(1-A ₀ A ₁ ) H ₀₀ } (4) B ₁ = 1 / {(1-A ₀ A ₁ ) H ₁₁ } (5) In this way The reason why the filter coefficients A ₀ , A ₁ , B ₀ , and B ₁ are calculated in the frequency domain is that unlike the case of performing them in the time domain, simple arithmetic operations are sufficient and the calculation load is reduced.

Next, in step 105, the filter coefficients A ₀ , A ₁ , B ₀ and B ₁ are high pass filtered to remove components below a predetermined frequency.
Specifically, the human audible frequency range (20 Hz to 20 kHz)
In z), the filtering process is performed using a high-pass filter in which a component in a frequency band (for example, 8 kHz or more) important for localization of a sound image remains.

Then, in step 106, the filter coefficients A ₀ , A ₁ , B ₀ , B ₁ from which the low frequency components have been removed in step 105 are inverse Fourier transformed to filter coefficients a ₀ , a in the time domain. ₁ , b ₀ , b ₁ are calculated, and in step 107, the filter coefficients a ₀ , a ₁ ,
The b ₀ and b ₁ are output to, for example, a predetermined external storage device, and the transfer function correction filter is configured with the filter coefficients a ₀ , a ₁ , b ₀ and b ₁ and incorporated in the stereophonic sound reproducing device.

In this embodiment, since the low frequency component is removed in step 105, the transfer function correction filter also has a function as a high pass filter.
FIG. 5 is a block diagram showing the configuration of the stereophonic sound reproducing device 15. The stereophonic sound reproducing device 15 uses the dummy head 6 as shown in FIG. 1 for recording in an original sound field such as a concert hall. Reproduction sound source 16, speakers 4L and 4R arranged in reproduction sound field 3, and speaker 4
Amplifiers 7L and 7R disposed on the input sides of L and 4R, and an adder 20 disposed on the input sides of the amplifiers 7L and 7R
L, 20R, reproduction sound source 16 and adders 20L, 20
A transfer function correction filter 17 interposed between R and a low-pass filter 18 interposed between the reproduction sound source 16 and the adders 20L and 20R in parallel with the transfer function correction filter 17.
L, 18R and the low-pass filters 18L, 18R
And delay circuits 19L and 19R as a delay means provided in series on the output side of the.

That is, the reproduction sound source 16 uses the so-called binaural recording method to record the acoustic data D for the left ear (recorded by the microphone 5L on the left ear side of the dummy head 6).
_L and the sound data D _R for the right ear (recorded by the microphone 5R on the right ear side of the dummy head 6) are recorded in a distinguishable manner, and these sound data D _L , D _R are transferred function correction filter 17 Is input to the adders 20L and 20R provided on the input side of the amplifiers 7L and 7R.

The acoustic data D _L and D _R are also input to the low pass filters 18L and 18R, and the outputs of the low pass filters 18L and 18R are passed through delay circuits 19L and 19R. Adders 20L, 2
It is designed to be input to 0R. And adder 2
0L and 20R are acoustic data supplied through the transfer function correction filter 17 and low-pass filters 18L and 1L.
8R and the audio data supplied via the delay circuits 19L and 19R are superimposed and supplied to the amplifiers 7L and 7R, and the outputs of the amplifiers 7L and 7R are supplied to the speakers 4L and 4R. .

When expressed in the frequency domain, the transfer function correction filter 17 (actually, the filter coefficient a ₀ in the time domain,
It is composed of a ₁ , b ₀ , and b ₁ , but is expressed in the frequency domain for the sake of easy understanding here), and the multiplication unit 17a that multiplies the acoustic data D _L for the left ear by the filter coefficient A _0.
A multiplication unit 17b that multiplies the sound data D _R for the right ear by a filter coefficient A ₁ , an addition unit 17c that adds the output of the multiplication unit 17b and the sound data D _L for the left ear, and a multiplication unit 17a.
Adder 1 for adding the output of _R and the sound data D _R for the right ear
7d, a multiplier 17e for multiplying the output of the adder 17c by a filter coefficient B ₀ and supplying the output to the adder 20L, and an adder 17
The multiplication unit 17f that multiplies the output of d by the filter coefficient B ₁ and supplies it to the adder 20R.

On the other hand, the low-pass filters 18L and 18R
The cutoff frequency is set to the same or substantially the same frequency as the cutoff frequency of the transfer function correction filter 17 that also functions as a high-pass filter. The delay times of the delay circuits 19L and 19R are the same as the processing time in the transfer function correction filter 17 and the low-pass filter 18L,
It is set so as to compensate for the difference from the processing time at 18R.
That is, the processing time in the transfer function correction filter 17, the series circuit of the low-pass filter 18L and the delay circuit 19L, and the low-pass filter 18R and the delay circuit 19R.
The delay times of the delay circuits 19L and 19R are set so that the processing time in the serial circuit of 1 becomes the same.

Next, the operation of this embodiment will be described. That is,
Left and right acoustic data D _L output from the reproduction sound source 16,
D _R is input to the transfer function correction filter 17 and the low-pass filters 18L and 18R. Since the transfer function correction filter 17 also functions as a high-pass filter as described above, the acoustic data D _L , D _R Of the high frequency component of the adder 20L, 20L through the transfer function correction filter 17.
The low frequency components of the acoustic data D _L and D _R supplied to _R are low pass filters 18L and 18R and a delay circuit 19L,
It is supplied to the adders 20L and 20R through 19R.

In this case, when the transfer function correction filter 17 and the low-pass filters 18L and 18R are compared with each other, the processing time in the transfer function correction filter 17 is generally longer because a function for removing crosstalk is required. However, the difference in processing time between them is the delay circuit 19L,
Supplemented by 19R. That is, the high-frequency component passing through the transfer function correction filter 17 and the low-pass filter 1
The low frequency components passing through 8L and 18R are time-matched and supplied to the adders 20L and 20R, and are simultaneously reproduced by the speakers 4L and 4R. Sounds corresponding to the components are not reproduced by the speakers 4L and 4R with a time lag, and the acoustic data D _L and D _R can be accurately reproduced.

If the transfer function correction filter 17 is constructed as shown in FIG. 5 (which is an expression in the frequency domain), the sound corresponding to the high frequency component of the left ear acoustic data D _L is heard by the listener. The sound that is reproduced in the left ear of M and corresponds to the high frequency component of the sound data D _R for the right ear is reproduced in the right ear of the listener M. For example, the transfer function correction filter 17
The transfer characteristic between the left ear acoustic data D _L of the reproduction sound source 16 and the left ear of the listener M via B is as follows: B ₀ H ₀₀ + A ₀ B ₁ H ₀₀ = {1 / (1-A ₀ A ₁ )} (1-H ₀₁ H ₁₀ / H ₀₀ H ₁₁ ) = {1 / (1-H ₀₁ H ₁₀ / H ₀₀ H ₁₁ )} (1-H ₀₁ H ₁₀ / H ₀₀ H ₁₁ ) = It is 1 that the sound corresponding to the high frequency component of the left ear acoustic data D _L is completely reproduced in the left ear of the listener M.
Similarly, the transfer characteristic between the acoustic data D _R for the right ear and the left ear of the listener M via the transfer function correction filter 17 is A ₁ B ₀ H ₀₀ + B ₁ H ₁₀ = (− H _{10 / H 11) H 00 / {} (1-a 0 a 1) H 00} + H 10 / {(1-a 0 a 1) H 11} = 0 , and the high frequency component of the acoustic data D _R for the right ear It can be seen that the corresponding sound is not reproduced at all in the left ear of the listener M (crosstalk is removed). The results of the above transfer characteristics are the same for the right ear of the listener M.

That is, by using the transfer function correction filter 17, the high-frequency component of the three-dimensional sound in the original sound field recorded in the reproduction sound source 16 is faithful to both ears of the listener M in the reproduction sound field 3. Is reproduced. on the other hand,
The low frequency components of the acoustic data D _R and D _L are supplied to the speakers 4L and 4R as they are through the delay circuits 19L and 19R without passing through a transfer function correction filter for removing crosstalk and the like, The sound corresponding to the low frequency component of D _R is emitted from the left speaker 4L to the reproduced sound field 3, and the sound corresponding to the low frequency component of the acoustic data D _L is emitted from the right speaker 4R to the reproduced sound field 3.

Therefore, the sound corresponding to the low frequency component of the sound data D _L emitted from the left speaker 4L reaches not only the left ear of the listener M but also the right ear, and similarly the sound from the right speaker 4R. The sound corresponding to the low frequency component of the emitted sound data D _R reaches not only the right ear but also the left ear of the listener M. That is, with respect to the low frequency components of the sound data D _L and D _R , basically, reproduction similar to normal stereo reproduction is performed, which seems to adversely affect reproduction of stereoscopic sound. Since the low-frequency component has a small contribution to the localization of the sound image, even if only the low-frequency component is reproduced without passing through the transfer function correction filter, the influence of the mismatch of the sound image with the original sound field can be small.

In other words, low-frequency components can be reproduced well without passing through the transfer function correction filter 17 having a large calculation load. With the configuration of this embodiment, the low frequency component is
The sound in the low frequency band is sufficiently reproduced because the sound is reproduced in the same manner as the conventional stereo reproduction, not completely ignored. Therefore, such a stereophonic sound reproducing device 1
Considering the application of No. 5 as a vehicle interior noise evaluation apparatus, for example, the vehicle interior noise often contains low-frequency components with a high level, and the evaluation apparatus should reproduce good low-frequency components. However, with the configuration of the present embodiment, the requirement can be sufficiently satisfied.

Moreover, since only the high frequency components of the impulse responses h _{00 to} h ₁₁ are taken into consideration when the transfer function correction filter 17 is created, the number of taps of the transfer function correction filter 17 can be reduced. There is an advantage. That is, since the time required for sufficient attenuation is shorter as the frequency is higher, the factor that lengthens the filter length is the low frequency component of the impulse response. That is, F
In consideration of all frequency components of the impulse response, the transfer function correction filter 17, which is an IR filter, has a waveform as shown in FIG. 6A and has a large number of taps. If only the components are taken into consideration, the waveform becomes as shown in FIG. 6B, and the number of taps becomes small.

The smaller the number of taps is, the more the amount of calculation is reduced. Therefore, even an inexpensive microprocessor can be sufficiently realized. For example, the lowest frequency component included in the transfer function correction filter 17 is
When the frequency is changed to about 8 kHz in this embodiment, the time until the sound of 80 Hz and the sound of 8 kHz are attenuated to the same level is 1: 100, and the filter length is also shortened,
The amount of calculation is reduced accordingly. Further, if the filter length becomes short, the accuracy of the numerical value on the rear side of the impulse response does not become a problem due to the phase shift, so that there is an advantage that the filter coefficient initially obtained can be used without being reset.

As described above, with the configuration of this embodiment,
The number of taps of the transfer function correction filter 17 can be reduced to reduce the amount of calculation, while the acoustic data D _L ,
Since the low-frequency component contained in D _R can also be reproduced, it is possible to reproduce a stereoscopic sound at a satisfactory level even in a general reproduction sound field 3 having reverberation with a low-cost configuration, and to reduce the low-frequency component. It does not sacrifice frequency components.

Since the cutoff frequency of the high-pass filter and the cutoff frequencies of the low-pass filters 18L and 18R when designing the transfer function correction filter 17 are made to match or substantially match, the acoustic data D _L ，
Since the reproduction of all the frequency components of D _R is guaranteed, the sound of the original sound field can be reproduced in the reproduced sound field 3 without omission. Therefore, in reality, rather than matching the cutoff frequencies of both filters, the low-pass filter 18
It is desirable that the cutoff frequencies of L and 18R be set slightly higher than the cutoff frequency of the high pass filter so that the pass band characteristics of both filters overlap slightly.

Further, when the transfer function correction filter 17 is designed by obtaining only the high frequency component of the impulse response, FIG.
As is clear from comparison between (a) and (b), there is an advantage that the direct sound and the reflected sound can be easily separated. Therefore, the transfer function correction filter 17 considering only the direct sound
Can also be created, which enables the reproduction of stereophonic sound without the influence of reflection. That is, since the initial reflected sound component and the reverberant sound component are likely to change later on the time axis as described above, the transfer function correction filter 17 including them is also included.
If it is designed, it may not be possible to accurately obtain the filter coefficient, but it is possible to obtain a stable effect by extracting only the direct sound and calculating the filter coefficient.

FIG. 7 is a diagram showing a second embodiment of the present invention, and like the first embodiment shown in FIG.
5 is a block diagram showing the configuration of FIG. The same components as those in the first embodiment are designated by the same reference numerals, and their duplicate description will be omitted. That is, in this embodiment, the delay circuits 19L,
Equalizers 21L and 21R are provided on the output side of 19R, whereby low-pass filters 18L and 18R are provided.
The frequency characteristic of the low frequency component extracted by
L and 21R are appropriately adjusted and then supplied to the adders 20L and 20R.

The gain change amount of each frequency component in the equalizers 21L and 21R is determined by the reproduction sound source 16 and the listener M through the low pass filters 18L and 18R.
It is set so that the inverse characteristic of the transfer characteristic between the ear positions of is realized. That is, the transfer characteristic H ₀₀ , obtained by the processing of FIG.
The frequency characteristics of the equalizers 21L and 21R are determined based on the inverse characteristics of H ₁₁ .

With such a configuration, the equalizers 21L and 21R may change the spectral shape of the low frequency components of the acoustic data D _L and D _R reproduced through the low pass filters 18L and 18R during the reproduction. Since it is canceled, there is an effect that a reproduced sound closer to the sound in the original sound field can be obtained. The transfer characteristics H ₀₀ and H ₁₁ required to determine the frequency characteristics of the equalizers 21L and 21R are as shown in FIG.
If the low-pass filters 18L and 18R and the delay circuits 19L and 19R are arranged on the input side of the 7R, the transfer characteristics H ₀₀ and H ₁₁ can be measured by the low-pass filters 18L and 18R and the delay circuits 19L and 19L. Since the characteristics of 19R can be included, equalizers 21L and 21R
It becomes possible to more accurately design the frequency characteristic of.

9 and 10 are diagrams showing a third embodiment of the present invention. FIG. 9 shows a high-pass filter used when designing the transfer function correction filter 17 in this embodiment (see step 105 in FIG. 4). 10) is a frequency characteristic diagram of FIG. 10 and FIG. 10 is a frequency characteristic diagram of the low-pass filters 18L and 18R in the present embodiment. The configurations of the transfer function correction filter designing device 1 and the stereophonic sound reproducing device 15 are the same as those in the first embodiment, and therefore their illustration and description are omitted.

That is, in this embodiment, the cutoff frequency of the high-pass filter is set to be slightly lower than 8 kHz, and the cutoff frequency of the low-pass filters 18L and 18R is set to 8 kHz.
It is set slightly higher than Hz to slightly overlap the pass characteristics of the high-pass filter and the low-pass filters 18L and 18R, and the attenuation characteristics are not steep but gentle (for example, 12 dB / octave or less).

When the high-pass filter and the low-pass filters 18L and 18R have the characteristics as in this embodiment, the reproduced sound by the transfer function correction filter 17 and the reproduced sound by the system including the low-pass filters 18L and 18R are generated. , The sound image changes discontinuously when exceeding the cutoff frequency, for example, when the sound source of a single frequency sweeps and the frequency of the generated sound rises or falls because they overlap near the cutoff frequency. This is advantageous in that it can be prevented, natural reproduction is possible, and stereophonic sound with no discomfort can be reproduced.

In the above embodiment, the impulse response h ₀₀
.About.h ₁₁ are obtained by using the adaptive digital filter W as described with reference to FIG. 2, but the present invention is not limited to this, and for example, the white noise signal x
Alternatively, the sound pressure measurement signal y and the sound pressure measurement signal y may be Fourier-transformed, the transfer characteristic may be calculated for each frequency component, and the result may be inversely Fourier-transformed to be directly obtained.

The reproduction sound source 16 is not limited to the one previously recorded as in the above embodiment, but the dummy head 6 having the microphones 5L and 5R as shown in FIG.
May be arranged in the original sound field, and the measurement sound may be used as it is as a reproduction sound source to reproduce the stereophonic sound in real time. In the above embodiment, the delay circuits 19L and 19R are provided in series with the low-pass filters 18L and 18R in order to match the processing time, but the low-pass filter 18L is shorter than the processing time in the transfer function correction filter 17. , 18R, the delay time 19L, 19R in series with the low-pass filters 18L, 18R is not provided and the processing time of the transfer function correction filter 17 and the low-pass filters 18L, 1R are not provided.
If a delay circuit for compensating for the difference from the 8R processing time is provided in series with the transfer function correction filter 17, the processing times can be matched as in the above embodiment, and reproducibility is not impaired. .

[0058]

As described above, according to the first aspect of the invention, the transfer function correction filter and the low-pass filter are interposed in parallel between the reproduction sound source and the plurality of speakers. While the number of taps of the function correction filter can be reduced to reduce the amount of calculation, the low-frequency component included in the output of the reproduction sound source can also be reproduced, so that it is an inexpensive configuration and generally has reverberation. It is possible to reproduce a three-dimensional sound at a satisfactory level even in various reproduction sound fields, and to obtain an effect that the low frequency component is not sacrificed.

According to the second aspect of the invention, since the reproduction of all the frequency components of the acoustic information output from the reproduction sound source is guaranteed, the sound in the original sound field can be reproduced without omission. effective. In the invention according to claim 3, the reproduced sound by the transfer function correction filter and the reproduced sound by the low-pass filter reach the predetermined sound receiving position at the same time, so that the sound in the original sound field is accurately reproduced. There is an effect that can be done.

Further, according to the invention of claim 4, there is an effect that a sound closer to the original sound field can be reproduced by appropriately setting the frequency characteristic of the equalizer. According to the invention of claim 5, the sound image is prevented from changing discontinuously when the cutoff frequency is exceeded, and natural reproduction is performed, so that stereophonic sound can be reproduced without discomfort.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a transfer function correction filter designing device in a first embodiment of the present invention.

FIG. 2 is a block diagram showing an example of a transfer characteristic measuring device.

FIG. 3 is a waveform diagram showing an example of an impulse response.

FIG. 4 is a flowchart showing an outline of processing in a filter coefficient calculation device.

FIG. 5 is a block diagram showing a configuration of a stereophonic sound reproducing device.

FIG. 6 is a waveform diagram showing the shape of a transfer function correction filter.

FIG. 7 is a block diagram showing a configuration of a stereophonic sound reproducing device in a second embodiment.

FIG. 8 is a block diagram showing a desirable configuration of a transfer function correction filter design device in the second embodiment.

FIG. 9 is a frequency characteristic diagram of a high pass filter in the third embodiment.

FIG. 10 is a frequency characteristic diagram of a low-pass filter according to the third embodiment.

[Explanation of symbols]

 1 Transfer Function Correction Filter Design Device 2 White Noise Generation Device 3 Playback Sound Field 4L, 4R Speaker 5L, 5R Microphone 6 Dummy Head 10 Transfer Characteristic Measurement Device 11 Filter Coefficient Calculation Device 15 Stereophonic Sound Reproduction Device 16 Playback Sound Source 17 Transfer Function Correction Filter 18L, 18R Low-pass filter 19L, 19R Delay circuit (delay means) 20L, 20R Adder 21L, 21R Equalizer

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical indication location G10K 15/12 H03H 21/00 8842-5J H04S 7/00 E (72) Inventor Tomoyuki Matsuzawa Kanagawa 2 Takaracho, Kanagawa-ku, Yokohama-shi Nissan Motor Co., Ltd.

Claims (5)

[Claims] 1. A reproduction sound source for outputting acoustic information for stereophonic reproduction, and a plurality of speakers to which the acoustic information is input, the plurality of speakers being provided between the reproduction sound source and the plurality of speakers. A transfer function correction filter, which is set based on a component having a frequency higher than a predetermined frequency of the impulse response between the speaker and the predetermined sound receiving position and corrects the transfer function of the sound reproduction system, and a low-pass filter are interposed in parallel. A stereophonic sound reproducing device characterized in that 2. The stereophonic sound reproducing apparatus according to claim 1, wherein the predetermined frequency and the cutoff frequency of the low pass filter are matched or substantially matched. 3. A delay means is provided in series with the transfer function correction filter or the low-pass filter, and transmission time of acoustic information between the reproduction sound source and a plurality of speakers via the transfer function correction filter, The stereophonic sound according to claim 1 or 2, wherein the delay time of the delay means is set so that the transmission time of the acoustic information between the reproduction sound source and the plurality of speakers via the low-pass filter coincides with each other. Playback device. 4. The stereophonic sound reproducing device according to claim 1, further comprising an equalizer provided in series with the low-pass filter. 5. The attenuation characteristic when extracting a component of a frequency higher than a predetermined frequency of the impulse response and the attenuation characteristic of the low-pass filter are set to gentle characteristics. The stereophonic sound reproduction device described.