JPH0795696A - Image normal positioning device - Google Patents

Abstract

PURPOSE:To simulate a real sound field and to clarify normal position feelings before and behind by allocating plural reflected sounds generated by delaying an input signal and the input signal as a direct sound to each circuit which sets the normal position of an image arbitrarily. CONSTITUTION:A control part 1 including a CPU controls the whole system by a program and data set by a user in a storage part 4, and furthermore, input information such as a sound source, a listener position, and an indoor reflection coefficient, etc., from a spatial information input part 3. A reflection part forming part 6 inputs the input signal as the direct sound from the sound source 5 and the signals of plural reflected sounds formed by delaying the input signal to an image normal positioning device 7. Plural normal position circuits in the device 7 divide those signals into respective frequency band by an LPF, a BPF, and an HPF, and the image can be normal-positioned by a delay circuit and a FIR filter, etc. In this way, it is possible to simulate an image normal position, distance difference between the right and left directions of a listener, and a transfer function around the ear and to obtain the sound field with clear normal position feelings before and behind.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sound image localization device which is built in or connected to an electronic musical instrument, a game device, an audio device, etc.

[0002]

2. Description of the Related Art Localization of sound images of electronically produced sounds and reproduced sounds has been performed conventionally.
There is a binaural method as one of the methods. With this binaural method, microphones are attached to the right and left ear canals of a life-size artificial head (called a dummy head), the target song or music is collected by this microphone, and the outputs are output to the right and left headphones, respectively. It is a method to listen in. in this case,
The collected sound may be recorded once on a magnetic tape or the like, and the recorded sound may be reproduced and listened to through headphones. It should be noted that the dummy head is made to have an average size and shape of a human head in order to approach an actual human, and is made of a material such as silicone rubber to approximate the softness of the skin of the face and the ear shell. There is. With the recent improvement in digital signal processing technology, digital signal processing technology has been adopted not only in the binaural system but also in sound image localization apparatus.

An example of a sound image localization device using such a digital signal processing technique will be described with reference to FIG.
Reference numeral 100 shown in this figure denotes a sound source such as an electronically generated sound or a sound reproduced by an audio device or the like.
Is a sound image localization circuit that localizes a sound image at a predetermined position,
FIR coefficient table 102 and FIR filters 103, 1
04 and 04. Reference numeral 102 is an FIR coefficient table that stores the coefficients of the FIR filters 103 and 104 for localizing the sound image at a predetermined position. Also, 105
Is a reverb circuit that generates reverb (reverberation sound), and the output of the reverb circuit 105 is the sound image localization circuit 10
The outputs of the two systems of 1 are added in the adders 106 and 107, amplified by the amplifiers 108 and 109, and the left channel output L is output from the output terminal 110 to the right channel output R.
Is output from the output terminal 111.

The coefficient stored in the FIR coefficient table 102 is the data of the transfer function obtained by measuring the impulse response from the localization position where the impulse sound source is installed in advance to the listening position, and the coefficient is digitally calculated. FIR (Finite Impulse Response) which is a filter
) It is a coefficient that can set the filter to simulate the measured transfer function. In the sound image localization apparatus shown in this figure, in order to localize a sound image at a desired position, first, the coordinates of the sound image localization position are determined by the localization coordinate input terminal 11
Enter 2. Then, based on the input coordinates, FI
From the R coefficient table 102, the coefficient for localizing the sound image at the coordinate is read out, and this coefficient is read by the FIR filter 103.
And the FIR filter 104, respectively. The signals from the sound source 100 are subjected to localization processing by the FIR filters 103 and 104 in which this coefficient is set, and these F
After converting the output signals from the IR filters 103 and 104 into analog signals, it is possible to listen as a sound image localized at a desired position by listening with headphones.

When a song or music is heard indoors, the sound reflected on the walls, ceiling, floor, etc. is heard as a reverberation sound. Therefore,
A reverb circuit 105 is provided as a circuit for simulating the reverberation sound, and the reverberation sound generated by the reverb circuit 105 is applied to the adders 106 and 107 to add the reverberation sound.

FIG. 6 shows how the sound from the impulse sound source 122 is transmitted to the listener 121 in the room 120. S _L and S _R are direct sounds from the impulse sound source 122, and the direct sound S _L is a distance d from the impulse sound source 122.
The direct sound S _R propagates a distance (do + d1) from the impulse sound source 122 and reaches the right ear. Furthermore, the sound from the impulse sound source 122 reflected by the wall of the room 120 is heard by the listener 121 as indicated by the broken line.
Reaching the left and right ears of. This reflected sound shows only the sound reflected only once by the wall of the room 120. Such a reflected sound is said to be an initial reflected sound, and shows a characteristic characteristic depending on the type of the room 20 and is twice. It can be distinguished from the reflected sound reflected by the third wall.

The mode of this early reflected sound is shown in the graph of FIG. The horizontal axis of the graph shown in this figure is time t, and the vertical axis is level. In the graph shown as L on the upper side of this figure, the listener 121 from the impulse sound source 122 shown in FIG.
The direct sound S _L reaching the left ear of the listener and the early reflection sound are shown. In the graph shown as R on the lower side, the direct sound S reaching the right ear of the listener 121 from the impulse sound source 122 shown in FIG. _R and early reflections are shown.

Referring to this figure, the direct sound S _L is the listener 12 after t _do after the impulse sound is emitted from the sound source 122.
1 reaches the left ear and the direct sound S _R is received by the listener 12 t _{(do + d1)} after the impulse sound is emitted from the sound source 122.
It reaches the right ear of 1. The initial reflected sound has the level and delay time characteristics shown in the figure. in this way,
It can be seen that there is a time shift and level difference between the direct sound S _L and the direct sound S _R, and also a time shift and level difference between the initial reflected sound.

[0009]

By the way, when the sound source is in front of the listener in the anechoic room (S _{L in the} case of FIG. 6).
= S _R ), it is impossible to determine whether the sound of the sound source comes from the front or the back. On the contrary, when the sound source is at the back, there is a phenomenon that it cannot be determined whether the sound source is at the front or at the back. This is because this experiment is performed in an anechoic room and the reflected sound is not generated. However, such a phenomenon is rarely recognized in the actual living space. The reason is that not only the direct sound but also the reflected sound is generated from the sound source in the real life space, and the direction of the sound source can be determined by determining the direction of the reflected sound. Therefore, it is necessary to localize the sound image not only for the direct sound but also for the reflected sound.

However, in the conventional sound image localization apparatus, as shown in FIG. 5, the initial reflected sound and the rear reverberant sound produced by the reverb 105 are not localized, and the problem is that the front and rear localization is unclear. There was a point. Therefore,
It is an object of the present invention to provide a sound image localization device that simulates a more realistic sound field, and in particular, makes the front and rear localization feeling clear. Further, another object of the present invention is to provide a sound image localization device in which the circuit scale is not so large as compared with the conventional one.

[0011]

In order to solve the above-mentioned problems, the present invention provides a reflected sound creating means for creating a plurality of reflected sounds by delaying an input signal, and localizing a sound image at a predetermined position. Multiple sound image localization means capable of
Each of the plurality of reflected sounds created by the reflected sound creating means and an assigning means for arbitrarily assigning the input signal as a direct sound to the plurality of sound image localization means are provided.

[0012]

According to the present invention, since the direct sound and the plurality of reflected sounds are respectively localized by the sound image localization device, the sense of localization before and after can be clarified and depends on the shape of the room and the like. The direction of reflected sound can be simulated. In addition, FI is obtained by removing at least low-frequency components.
Since the sound image is localized by the R filter, F
The structure of the IR filter can be simplified. Furthermore, since the transfer function simulated by the FIR filter is a transfer function only around the ear of the human body, the order of the FIR filter can be made low.

[0013]

1 shows an electronic musical instrument equipped with a sound image localization apparatus of the present invention. In FIG. 1, reference numeral 1 denotes a control unit that includes a CPU and controls each block. The main program of the CPU, data set by the user, and the like are stored in the storage unit 4. From the spatial information input unit 3, the sound source position, the listener position, the reflection coefficient of the wall of the room, the shape of the room, etc. are input,
From these information, the coefficients of the reflected sound forming unit 6 and the sound image localization device 7, the read addresses are obtained by calculating or referring to a table, and the reflected sound forming unit 6 and the sound image localization device 7 are obtained.
To It is set by the control unit 1.

When a key depression event of the keyboard section 2 is detected by the control section 1, the control section 1 instructs the sound source 5 to generate sound in response to the key depression event, and the sound source data from the sound source 5 is reflected. It is sent to the forming unit 6. A large number of reflected sounds are created by the reflected sound forming unit 6 and input to the sound image localization device 7 together with the direct sound. Based on the information input by the spatial information input unit 3, each of the direct sound and the multiple reflected sounds is generated. It is processed to localize at the set position. In this way
The signal subjected to the sound image localization processing is converted into an analog signal and directly produced by the headphones 11, or the crosstalk cancellation circuit 8 cancels the crosstalk and the L channel speaker 9 and the R channel speaker 10 perform the crosstalk cancellation. Is pronounced.

The crosstalk canceling circuit 8 cancels the signal component that crosstalks from one ear to the other ear, so that when the cancelled signal is sounded from the speakers 9 and 10, it is at a predetermined localization position. This is a circuit for localizing the sound image. Such a circuit is a conventionally known circuit as described in JP-A-4-150400. Next, the reflected sound forming unit 6
The details of the above are shown in FIG.

In this figure, the sound source data input from the sound source 5 is applied to the delay line 21. The delay line 21 is composed of, for example, a RAM,
A direct sound is formed by being output from the delay line 21 to the output line 21-1 with a delay of a time td corresponding to the distance from the localization position to the listener, and a first initial reflected sound is output with a delay of time tr. -2, and a number of initial reflected sounds are output line 2 after a predetermined time delay.
It is output to 1-3 to 21-n. These direct sound and a large number of early reflection sounds are processed by the coefficient unit 22-1 to the coefficient unit 22-n.
To a predetermined level and are supplied to the sound image localization device 7, respectively.

The delayed output output from the rear end of the delay line 21 is supplied to a circuit that forms a rear reverberation sound following the initial reflected sound. This rear reverberation is generated by the room walls, ceiling,
It is a reverberant sound that is generated by being reflected twice or more by a floor or the like. Since it is reflected many times, it is generated later than the initial reflected sound. In order to generate such a rear reverberation sound, the output of the delay line 21 is applied to all-pass filters APF1, APF2, APF3 connected in cascade. The characteristic of this all-pass filter is a filter in which the frequency attenuation characteristic is flat but the phase is shifted according to the frequency, as is conventionally known, and when an impulse signal is supplied, the signal is delayed slightly. Are output in time series. Therefore, it is possible to form a reflected sound having a slight time delay. By using this, the all-pass filter APF with different delay time
1, APF2, APF3 form a large number of slightly delayed reflected sounds immediately after the signal.

Further, the reflected sound thus formed is added to adders 25-1 to 25-m and delay lines 26-1 to 26-26.
-M and low pass filter (LPF) 27-1 to 27-
m and the third coefficient unit 28-1 to 28-m. This comb filter has delay line 2
The output delayed every 6-1 to 26-m is output, and the LPF 27-
The high frequency component is attenuated by 1-27-m, and
The level is attenuated by the coefficient units 28-1 to 28-m. This has the characteristic that the level is attenuated each time the sound is reflected and the high-frequency component is attenuated, and this characteristic is for simulating this characteristic. In addition, the second coefficient unit 24 that is placed in front of these comb filters
1 to 24-m are all-pass filters APF1 to APF3
By adjusting the level of the reflected sound that is output more, the level is adjusted so that it becomes like a rear reverberation sound.

Although the rear reverberation is formed in this way, the delay lines 26 of the plurality of comb filters are formed.
The reflected sounds of coarse delay are formed by -1 to 26-m, and all-pass filters APF1 to AP are filled so as to fill the gap between them.
A reflected sound with a fine delay is formed by F3. Then, the delay lines 26-1 to 26-1 forming the comb filter
The delay times of 26-m are made different, and the output is obtained even from the middle of the delay lines 26-1 to 26-m. Further, the all-pass filter APF1
By making the delay times of APF3 different from each other, a larger number of reflected sounds are formed.

Next, the details of the sound image localization device 7 for locating the direct sound, the early reflection sound and the rear reverberation sound will be described with reference to FIG. As shown in FIG. 3, the sound image localization device 7 includes a distribution unit 40 and localization circuits 41-1 to 41-k. The distribution unit 40 distributes the input direct sound 31, the initial reflected sound 32, and the rear reverberation sound 33 so as to be localized at desired localization positions. This distribution is controlled by the control signal from the control unit 1 shown in FIG. 1, and the direct sound 31 and the initial reflected sound are generated depending on the sound source position, the listener position, the shape of the room, etc. input by the user in the spatial information input unit 3. 32, a plurality of sounds forming the rear reverberation sound 33 are supplied to each of the localization circuits 41-1 to 41-k for each sound. From the distribution unit 40, localization circuits 41-1 to 41-k
The output of only a few k of the output sound is output, and the direct sound 31,
The early reflection sound and the rear reverberation sound 33 can be assigned respectively. The k localization circuits 41
The localization positions of -1 to 41-k are set to the positions shown in FIG. 4, for example, and the sound image can be localized at any of the localization positions.

All of the localization circuits 41-1 to 41-k have the same configuration, and only the details of the localization circuit 41-1 among them are shown in FIG. In this localization circuit 41-1, different localization processing is performed in the delay circuit 42 and other circuit parts. This is shown in FIG.
6, when a sound image is localized at the position of the impulse sound source 122 shown in FIG. 6, the common distance between the direct sound S _L and S _R from the localization position 121 to the listener 121 is do. Therefore, it is understood that it is necessary to further propagate the listener 121's right ear by the distance d1. Therefore, a delay time corresponding to the distance d1 corresponding to the orientation of the listener with respect to the localization position 122 is obtained by the delay circuit 42, and the transfer function around the ear of the listener 121 that localizes the sound image at the localization position 121 is simulated. The processing to be performed is performed in another circuit portion.

That is, the direct sound 31, the initial reflected sound 32, and the rear reverberant sound 33 supplied from the distributor 40 are delayed by the delay circuit 42 and correspond to the distance difference d1 between the left side and the right side of the listener with respect to the localization position. L with time difference to
The outputs of the channel and the R channel are obtained from the lines 42-L and 42-R. And this line 4
The outputs from 2-L and line 42-R are low-pass filters (LPF) 43-L and 43-R, band-pass filters (BPF) 45-L and 45-R, and high-pass filter 4.
Input to 8-L and 48-R. This LPF43-
The signals of the frequency components of 20 to 1000 Hz extracted by the L and 43-R are adjusted by the fourth coefficient units 44-L and 44-R in the level attenuation amount based on the distance difference d1 and the adder 51- It is output to L and 51-R. Like this
Only the level difference is processed for the low-frequency components extracted by the PFs 43-L and 43-R. The reason for this is that since the wavelength of the low-frequency signal is long, the wraparound becomes large and the effect of sound image localization processing is weakened, and in order to localize a signal with a long wavelength, the filter for localizing the sound image becomes extremely large, and the cost This is because it becomes a factor that raises.

Next, the notch filters 46-L and 46-R remove some of the frequency components of the signals having frequency components of 1000 to 8000 Hz extracted by the BPFs 45-L and 45-R, and the distance difference d1. Is adjusted by the fifth coefficient units 47-L and 47-R and output to the adders 51-L and 51-R. The notch filters 45-L and 45-R are filters for localizing the sound image up and down, and the notch filters 45-L and 45-R.
The fact that the localization of the sound image changes vertically depending on the frequency component eliminated by R is used. As described above, the components extracted by the BPFs 45-L and 45-R are only subjected to processing for sound image localization in the vertical direction.

Next, the frequency components above 8000 Hz extracted by the HPF48-L and 48-R are FIR (Fini).
te Impulse Responce) filters 49-L and 49-R are applied to simulate the transfer function around the listener's ear, and the level attenuation amount based on the difference in distance is the sixth
The adder 51 is adjusted by the coefficient units 50-L and 50-R.
It is output to -L and 51-R. The FIR filters 49-L and 49-R are supplied from the control unit 1 with a coefficient capable of simulating the transfer function of the fixed position measured at the listener's position with the sound source placed at one of the plural fixed positions shown in FIG. Has been set. This transfer function is a transfer function measured in an anechoic chamber, that is, a transfer function only around the ear lobes, shoulders, and face of the human body and does not include reverb, and therefore FIR filters 49-L and 49-R are used. The order of can be quite low.

In this way, the L channel signal and the R channel signal output from the adders 51-L and 51-R which have been subjected to the localization processing by the localization circuit 41-1 are other localization circuits 41.
The outputs of −2 to 41-k and the adders 52-L and 52-R are added to obtain an L channel signal and an R channel signal. The present invention thus provides a direct sound, an early reflection sound,
Since the rear reverberations are localized respectively, it is possible to clarify the localization feeling before and after. Further, in locating these sounds, only the high frequency signals extracted by the HPF are used as the signals for locating the sound image by the FIR filter, so that the configuration of the FIR filter can be simplified. Furthermore, since the transfer function simulated by this FIR filter is the transfer function only around the ear of the listener measured in the anechoic chamber, the order of the FIR filter can be made low.

It is also possible to apply a signal including a direct sound, an early reflection sound and a rear reverberation sound to an FIR filter to perform sound image localization processing. For example, a sound image localization up to a reverberation sound of about 2 seconds is possible. And the sampling frequency is 4
When set to 8 kHz, the number of stages of the FIR filter is about 9600
This is difficult to achieve because it has 0 stages. Further, although the localization position shown in FIG. 4 is only the position in the horizontal direction, the localization position of the present invention is not limited to this, and a localization circuit for performing the up and down localization is further provided to perform the up and down sound image localization. You may do it.

[0027]

Since the present invention is configured as described above, it is possible to localize the direct sound and a plurality of reflected sounds by the sound image localization device, respectively. Therefore, the sense of localization before and after can be made clear. At the same time, the direction of the reflected sound that depends on the shape of the room can be simulated. Further, since the sound image is localized by the FIR filter by removing at least the low frequency component of the signal, the configuration of the FIR filter can be simplified. Furthermore, since the transfer function simulated by the FIR filter is a transfer function only around the ear of the human body, the order of the FIR filter can be set to a low order.

[Brief description of drawings]

FIG. 1 is a block diagram in which a sound image localization device of the present invention is applied to an electronic musical instrument.

FIG. 2 is a detailed block diagram of a reflected sound forming unit of the present invention.

FIG. 3 is a block diagram of a sound image localization apparatus of the present invention.

FIG. 4 is a diagram showing sound image fixed positions of a direct sound and a reflected sound.

FIG. 5 is a block diagram of a conventional sound image localization device.

FIG. 6 is a diagram showing how sound from a sound source is transmitted to a human body in a room.

FIG. 7 is a graph showing arrival times of a direct sound and an initial reflected sound.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Control part 2 Keyboard part 3 Spatial information input part 4 Storage part 5,100 Sound source 6 Reflection sound formation part 7 Sound image localization device 8 Crosstalk cancellation circuit 9,10 Speaker 11 Headphones 21, 26-1 to 26-m, 42 Delay Line 21-1 to 21-n Delay line output 22-1 to 22-n, 24-1 to 24-m, 28-1 to
28-m, 44-L, 44-R, 47-R, 47-L,
50-L, 50-R Coefficient unit 23-1 to 23-3 All-pass filter 25-1 to 25-m, 51-L, 51-R, 52-L,
52-R, 106, 107 Adder 27-1 to 27-m Low-pass filter 31 Direct sound 32 Initial reflection sound 33 Rear reverberation sound 34, 42-L L channel output 35, 42-RR channel output 40 Distributor 41- 1 to 41-k localization circuit 43-L, 43-R low-pass filter 45-L, 45-R band-pass filter 46-L, 46-R notch filter 48-L, 48-R high-pass filter 49-L, 49- R, 103, 104 FIR filter 101 Sound image localization circuit 102 FIR coefficient table 105 Reverb circuit 108, 109 Amplifier 110, 111 Output terminal 120 Room 121 Listener 122 Impulse sound source

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical indication location H03H 17/06 8842-5J H04S 7/00 F 8421-5H

Claims (2)

[Claims] 1. Reflected sound creating means for creating a plurality of reflected sounds by delaying an input signal, a plurality of sound image localization means capable of localizing a sound image at a predetermined position, and the reflected sound creating means. A sound image localization apparatus comprising: each of a plurality of reflected sounds; and an allocation unit that arbitrarily allocates the input signal as a direct sound to the plurality of sound image localization units. 2. A dividing means for dividing an input signal into a plurality of frequency bands, and a difference between a distance connecting the sound image localization position and the left side of the listener and a distance connecting the sound image localization position and the right side of the listener. The sound source localization means is composed of a first localization means composed of a delay circuit for performing such localization processing and a second localization means composed of an FIR filter simulating a transfer function around the ear of the human body. A sound image is localized by the sound image localization means with respect to a signal in a frequency band excluding at least a low frequency band signal in the divided frequency band, and a level difference is generated with respect to at least a low frequency band signal. A sound image localization device characterized by being imparted.