JP2001346298A - Binaural reproducing device and sound source evaluation aid method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a binaural reproducing device that is readily used, realizes sound image localization for many unspecified people in a front direction out side the head and can attain sound reproduction with fidelity to an original sound from the standpoint of sound quality. SOLUTION: A DAT 6 stores an acoustic signal picked up by a dummy head microphone 2 in an original sound field 1 and outputs the signal again. A sound localization filtering section 3 conducts convolution arithmetic operation or the like by using a filter coefficient for elimination of a crosstalk component to allow a loudspeaker 4 to conduct binaural reproduction thereby realizing sound image localization for many unspecified people in the front direction out side the head. Furthermore, the loudspeaker 4 is placed close to a listener 5. Thus, a reflected sound from a floor can be more decreased than a direct sound and the reproduction with fidelity to the original sound can be realized.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a binaural stereophonic technique for reproducing a sound field with a sense of realism by reproducing a sound recorded by a dummy head microphone and listening with both ears.

[0002]

2. Description of the Related Art Conventionally, in the field of acoustic technology, a binaural system has been known in which a sound image is localized with high fidelity to the original sound by using recording with a dummy head microphone and reproduction with headphones. First, the principle of the binaural method will be briefly described. FIG. 13 is an explanatory diagram of a situation where a human perceives the position of a sound source with both ears in an actual sound field space. FIG. 14 shows a case where sound is recorded by a dummy head microphone installed in an original sound field and reproduced by headphones. FIG. 2 is an explanatory diagram of an example of a binaural recording / reproducing method for listening with both ears. In the figure, 1 is an original sound field, 2 is a dummy head microphone, 5 is a listener, 11
Is headphones.

In order to perceive the position of a sound source, humans recognize interaural characteristics such as a difference in arrival time and level of sound between the left and right ears, monoaural characteristics such as a change in frequency characteristics, and the above information when the head is moved. It is said to use various information such as changes in the visual information. Above all, the interaural characteristics play a major role, and studies focusing on the binaural characteristics have been conducted for a long time.

[0004] Sound waves emitted from a sound source propagate in space,
When reaching the eardrum through the circumference of the human head, the pinna, and the ear canal, phenomena such as reflection, diffraction, and attenuation occur. That is,
In FIG. 13, the sound emitted from the sound source and the sound SL reaching the left ear and the sound SR reaching the right ear of the human greatly differ in amplitude, phase, and the like. Therefore, the above sound S
In order to record L and SR, it is necessary to arrange a small microphone at the entrance of the ear canal of both human ears and the front of the eardrum to record sound.

Generally, due to man-hours and safety issues, the shapes of the heads and pinna of several tens to several hundreds are measured, and dummy heads simulating these are prepared, and the entrances of the ear canals of the left and right ears are prepared. Alternatively, a microphone is placed at a position corresponding to the front of the eardrum to record sound. By using such a dummy head microphone 2, it is possible to record a sound very similar to a sound that a human actually listens to in the original sound field 1. By reproducing this sound with the headphones 11 and making the listener 5 listen to the sound with both ears, it is possible to reproduce a sound field with a sense of reality that is faithful to the original sound field 1.

However, binaural / monoaural characteristics such as arrival time difference, level difference, and frequency characteristic change between the left and right ears used by a human to perceive the position of a sound source vary depending on the delicate shape of the head and pinna. It is known to In the binaural recording / reproducing method using the recording by the dummy head microphone 2 and the reproduction by the headphone 11 as described above, an unspecified number of persons having different binaural / monaural characteristics caused by a difference in the shape of the head or auricle are used. On the other hand, it was not able to fully demonstrate its performance. In particular, it is known that, for a reproduced sound image, a sense of localization in the front direction of the listener is insufficient, and an in-head sound image localization phenomenon in which the sound image is localized in the head occurs.

To solve such a problem, see, for example,
Japanese Patent Publication No. -115099 discloses a method in which extraneous sound image localization information for each individual listener is measured in advance and stored in a storage card. A calculation method is disclosed. According to this method, since the sound transmission characteristics corresponding to the shape of the head and pinna of each individual can be corrected, each individual can obtain a good feeling of localization outside the head.

However, in this method, in order to accommodate an unspecified number of listeners, a small microphone is inserted into the external auditory canal entrance of both ears, preferably the front of the eardrum, for the unspecified number of people to be heard. In addition, it is necessary to measure the head transfer characteristics, and there is a problem that the feasibility is poor due to man-hours and safety problems.

[0009] For example, JP-A-6-178396 and JP-A-3-214896 detect a change in the direction of the head of the listener and adjust the left and right output signals of the headphones accordingly. Thus, there is disclosed a sound field localization control device or an acoustic signal reproducing device in which a sound image is always localized in a fixed direction even when the head direction changes. In a general binaural system, when the head is rotated, the position of the sound image is also rotated. However, in these systems, the position of the sound image is fixed even if the head is moved. This allows the listener to feel as if the sound image is outside his / her head.

The usefulness of these methods can be understood from the fact that when a human is difficult to recognize the position of the sound source in daily life, his head is swung right and left to make the recognition easier.
However, in these systems, the head must be constantly moved in order to maintain the sound image localization outside the head,
There was a problem that the usability was extremely poor and the utility was poor.

Further, for example, Japanese Patent Application Laid-Open No. 7-15780 discloses a headphone in which a speaker section is disposed in front of a region connecting left and right ear holes of a listener. According to this headphone, the sound is heard from the front of the axis connecting the ear canal because the speaker section is disposed in the front.

However, in this headphone, since the speaker section is arranged at a position far from the ear canal, distortion of the sound transfer characteristic in the space from the speaker section to the entrance of the ear canal cannot be ignored, and the sound collected by the dummy head microphone can be reproduced by the headphone. As a result, the basic concept of binaural reproduction, ie, reproduction near the ear canal entrance, was broken. That is, when the sound collected by the dummy head is reproduced by the above-mentioned headphones, the reproduced sound image is localized as a sound image having no depth on a line connecting the left and right speaker units, as in the case of general stereo reproduction, Alternatively, the sound image cannot be localized in the near side. In addition, since the distortion of the sound transmission characteristic in the space from the speaker section to the entrance of the ear canal cannot be ignored, there is a problem that the sound of the original sound field cannot be faithfully reproduced also in terms of sound quality.

An audio system using a loudspeaker instead of headphones as a reproducing device is described in Bauer (196).
1) and Schroeder (1966). In the case of stereo reproduction using a pair of loudspeakers, generally, in order to obtain a reproduction sound image space as large as possible, the left and right loudspeakers are placed as far as possible from the listener, and are opened about 120 degrees when viewed from the listener. Place them apart. In a sound system using a loudspeaker, since the sound emitted from the loudspeaker propagates through space and is transmitted to both human ears, the sound recorded by the left and right microphones of the dummy head can be heard by the listener's left and right ears. It is difficult to realize the binaural system in which the audio is separated and listened. However, in this acoustic system, the sound transmission characteristics from the loudspeaker to the listener's both ears are measured in advance, and by calculating the inverse characteristic, the sound emitted from the left front (right front) loudspeaker is generated. A filter is configured to eliminate crosstalk that is also transmitted to the right ear (left ear) of the listener, thereby realizing binaural reproduction using a loudspeaker. Further, in this method, since the loudspeaker is arranged in front of the listener, the sound image is not localized inside the head, and a good sound image localization feeling can be obtained. As a system for realizing this system, for example, Japanese Patent Application Laid-Open No. 2-86399 discloses a system provided with a personal filter in the median direction.

However, all of these acoustic systems are based on the assumption that all walls are subjected to sound absorption treatment for preventing sound reflection and are used in a completely anechoic room having a floating floor structure with a wire mesh or the like. When used in a room where a normal floor surface exists, the following problem occurs. That is, in addition to the direct sound in which the sound radiated from the loudspeaker directly propagates through the space and reaches the listener's ear, there is a reflected sound that is reflected on the floor or the like and then reaches the listener's ear. for that reason,
In particular, when a pure tone sound composed of a single or narrow band frequency is reproduced, a phenomenon occurs in which the directly propagated sound and the floor reflected sound interfere in space and the sound is strengthened or weakened,
There is a problem that it is difficult to reproduce a sound field faithfully in terms of sound quality. A typical example of a pure tone sound is a sound generated from a rotating machine such as a motor. In general, it is known that a rotary machine generates a pure tone sound in an extremely narrow band at a frequency that is a positive multiple of the rotation speed. Binaural recording / playback devices are often used for subjective evaluation of the operation sound of such machines in addition to listening to music, but in such evaluations, it is strongly that the sound quality is faithful to the original sound. Therefore, the above-described acoustic system that cannot faithfully reproduce a pure tone sound cannot be used.

Another problem with these acoustic systems is that the psychological and physical burden on the listener is great. That is, in this acoustic system,
A sound transmission characteristic determined from the positional relationship between the loudspeaker and the listener is measured in advance, and a filter for eliminating crosstalk is configured. Therefore, in order for the crosstalk canceling filter to function effectively, the listener must always keep the position and orientation of both ears, that is, the head. Since such a restriction places a great psychological and physical burden on the listener, there is a problem that it cannot be widely used for general purposes.

[0016]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and can be used easily.
It is another object of the present invention to provide a binaural reproduction device capable of realizing localization of a sound image in an out-of-head forward direction for an unspecified number of people and reproducing a sound field faithful to the original sound in terms of sound quality. It is another object of the present invention to provide a sound evaluation support method that makes it easy for a listener to perform a sensory evaluation using such a binaural reproduction device.

[0017]

SUMMARY OF THE INVENTION The present invention relates to a binaural reproduction apparatus for reproducing a binaurally recorded sound signal, using a pair of loudspeakers, and measuring sound transmission characteristics from loudspeakers measured in advance to both ears of a listener. And a sound image localization filtering unit that eliminates crosstalk in which sound radiated from the front left (right front) loudspeaker is also transmitted to the right ear (left ear) of the listener. It is characterized by being arranged close to a listener. According to such a configuration, a good sound image localization feeling ahead of the head can be realized for an unspecified number of listeners as described below, and the sound field reproduction faithful to the original sound in terms of sound quality can be achieved. Will be possible.

As described above, if there is a subtle difference in the head transfer characteristics of each individual, the sense of sound image localization in the frontal direction of the listener is particularly insufficient, and the sound image localization phenomenon in the head in binaural reproduction using ordinary headphones. Occurs. However, in binaural reproduction using a loudspeaker arranged in front of the listener, no intra-head sound image localization occurs. This means that when a situation where the sound image localization feeling in the front direction is insufficient due to a delicate difference in head transfer characteristics occurs, the reproduced sound image position is the true sound source direction, that is, the vicinity of the pinna in the headphones, the front arrangement This suggests that the loudspeaker will be pulled forward.

In the binaural reproduction apparatus of the present invention, the binaural reproduction using the loudspeakers is realized by providing the pair of loudspeakers and the sound image localization filtering unit as described above, and the subtleties of the dummy head microphone and the head-related transfer function are obtained. Thus, even when the reproduction is performed for an unspecified number of different listeners, it is possible to impart a sound image localization feeling in the forward direction.

Further, in the binaural reproduction device of the present invention, since the loudspeaker is arranged close to the listener, it is possible to reproduce faithfully the original sound in terms of sound quality. That is, since the loudspeaker is arranged close to the listener, the propagation distance of the directly propagating sound is sufficiently smaller than the propagation distance of the reflected sound. Due to the distance attenuation of the sound pressure, in which the sound pressure of the propagated sound decreases in inverse proportion to the square of the propagation distance, the ratio of the reflected sound to the directly propagated sound can be made negligibly small. As a result, it is possible to suppress the interference phenomenon in a space that occurs when a pure tone sound is reproduced, which has been a problem in the conventional reproduction device using a loudspeaker, and that the reproduction faithful to the original sound can be achieved in terms of sound quality. Will be possible. Generally, the discrimination threshold of human hearing is said to be 1 dB. In order to prevent the listener from perceiving the change in sound pressure level due to the interference phenomenon, the distance from the listener to the loudspeaker must be 1 m or less. A loudspeaker may be arranged.

Further, the pair of loudspeakers may be arranged close to each other. By disposing the pair of loudspeakers close to each other, it becomes possible to maintain good binaural reproduction even when the listener's head moves during reproduction. As described above, the interaural characteristics such as the arrival time difference and the level difference between the left and right ears play a large role in human sound source position perception. Therefore, in order to realize good binaural reproduction, it is important to realize and maintain interaural characteristics similar to those of the original sound field. However, in a conventional binaural reproduction device using loudspeakers, both left and right loudspeakers are spaced apart from each other by about 120 degrees in order to obtain a sound image space as large as possible. In such an arrangement, for example, when the listener moves his / her head to the right, the distance between the right front loudspeaker and the right ear decreases, and the distance between the left front loudspeaker and the left ear increases. Therefore, a change in the opposite tendency occurs between the left and right sides, which greatly affects the interaural characteristics. For this reason, if the listener moves his or her head during playback, there is a problem that binaural playback cannot be maintained.

On the other hand, the present invention arranges a pair of loudspeakers in front of the listener in a state of being close to each other, so that even if the listener moves his or her head left and right, the tendency is reversed between the two ears. Such a large change does not occur. This makes it possible to maintain good binaural reproduction even if the listener moves his or her head during reproduction.

However, when the loudspeakers are arranged close to each other as described above, the crosstalk in which the sound radiated from the left front (right front) loudspeaker propagates to the right ear (left ear) of the listener increases. . This makes it difficult to calculate the filter coefficients of the sound image localization filtering unit for eliminating crosstalk, and may impair good binaural reproduction depending on the environment of the room where the reproducing apparatus is used. In order to suppress such a problem, for example, it is desirable to install a sound insulation plate between the left and right loudspeakers to partition the front space of the listener left and right. In order to realize sound field reproduction that is faithful to the original sound in terms of sound quality, it is desirable that a part of the sound insulating plate or the entire sound insulating plate be made of a sound absorbing material.

Further, even if the space between the two ears of the listener from the loudspeaker is left-right asymmetric, a sense of localization in the forward direction can be obtained. It is desirable that the positional relationship of the space between the loudspeaker and the listener's both ears determined from the spread angle, the orientation, the distance from the loudspeaker to the listener's ear, and the like be symmetrical. Also, when the left and right loudspeakers are housed in one housing, the positional relationship and shape of the space of the housing, and when a sound insulating plate is provided, the positional relationship and shape of the space of the sound insulating plate are also roughly described. It should be symmetric.

The binaural reproduction apparatus of the present invention as described above can be used for a sound evaluation supporting method for supporting sound evaluation, for example, by comparing and evaluating the noise of a plurality of mechanical devices. . That is, the sounds emitted from the plurality of measurement targets are respectively recorded, and only the comparison target portion is taken out from each of the recorded acoustic signals and recorded, and the binaural reproduction device of the present invention provides:
The sound signals of the comparison target portion are arranged in time series and reproduced. Accordingly, the listener can continuously listen to the plurality of comparison target portions and perform the comparative evaluation, and can assist in realizing the comparative evaluation without impairing the sense of the listener.

[0026]

FIG. 1 is a schematic diagram showing an example of a system including a first embodiment of a binaural reproduction device according to the present invention. In the figure, 1 is an original sound field, 2 is a dummy head microphone, 3 is a sound image localization filtering unit, 4 is a loudspeaker, 5 is a listener, and 6 is a DAT. In a system including this binaural reproduction device, a dummy head microphone 2 for recording sound in an original sound field 1 and a DAT (Digital) for recording or re-outputting the same.
l Audio Tape Recorder 6, a pair of loudspeakers 4 for reproducing sound, a sound image localization filtering unit 3 for eliminating crosstalk from acoustic transfer characteristics from the previously measured loudspeakers 4 to both ears of the listener 5, and the like. It is configured.

The dummy head microphone 2 has an original sound field 1
Record the sound at. As the dummy head microphone 2, a microphone is mounted at the entrance of the binaural ear canal of a dummy head that simulates the shape and acoustic characteristics of the human head and auricle, and a microphone is mounted at the eardrum position also simulating the shape of the ear canal. HATS (Head and T) that imitates not only the head but also the shape of the upper body such as the shoulders and torso
Orso Simulator) or the like can be used. Here, as an example, the head and upper body are simulated from data of 28 Japanese adults, and small microphones are arranged at the entrances of the external auditory canals of the left and right auricles, respectively.
ATS (Head and Torso Simulat)
or) was used.

DAT 6 is a dummy head microphone 2
The recorded sound can be recorded and output again. DA
As a specific example of T6, 48 kHz sampling, 16
A stereo 2-channel DAT for general music with bits can be used. As a device for recording and re-outputting the sound recorded by the dummy head microphone 2, DAT6
In addition, tape recorders and CD-Rs (CD-ROM R
recorder), a memory of a personal computer or a workstation, an external storage device, or the like.
Of course, the sound recorded by the dummy head microphone may be directly reproduced by the loudspeaker 4 without using such a recording and re-output device.

The sound image localization filtering unit 3 performs a sound image localization filtering process on the basis of previously measured acoustic transfer characteristics of the reproduction space including spatial acoustic transfer characteristics from the loudspeaker to specific portions of the left and right ears and electroacoustic conversion characteristics of the loudspeakers. And a filter coefficient for eliminating a crosstalk component is calculated in advance, and a sound signal to be reproduced is filtered. Note that the measurement of the spatial acoustic transfer characteristic can use an existing transfer characteristic measuring method such as a sine wave sweep method or a cross spectrum method using broadband noise.

The sound image localization filtering unit 3 can use a general analog or digital arithmetic element. When a digital operation element is used, a convolution operation unit that performs a convolution operation on a digital acoustic signal and a filter coefficient, and a digital-analog (D / A) conversion that converts the convolution operation result into an analog signal and outputs the result to the loudspeaker 4 as an acoustic signal It can be constituted by a vessel or the like. If the input audio signal is an analog signal, for example, when the audio signal is directly input from the dummy head microphone 2 or when the analog signal is output from the DAT 6, the input analog signal is output. An analog-to-digital (A / D) converter for converting an acoustic signal into a digital signal is provided.

As these computing units, at least the convolution computing unit is a DSP (Digital Signal Pr).
processor) can be used. The digital filter for convolving the input audio signal with a filter coefficient is configured by cascading a plurality of one-sample delay elements to the input audio signal, and multiplying the input audio signal and each delay signal by a filter coefficient. Is summed up. Specifically, a digital filter having 2048 taps can be used.

The filter coefficients used in the convolution operation can be measured and calculated in advance as follows. FIG. 2 is an explanatory diagram illustrating an example of a measurement state of a sound transfer characteristic from the loudspeaker 4 to both ears of the listener 5.
The loudspeaker 4 is installed in the same manner as when performing binaural reproduction as described later. In addition, the dummy head microphone 2 can be used instead of the listener 5 in order to record acoustic signals at the positions of both ears of the listener 5.

First, a broadband noise signal is input to the right front loudspeaker 4 to generate wideband noise, and the original broadband noise signal and the signal recorded by the dummy head microphone 2 reach the left and right ears by the cross spectrum method. The sound transfer characteristics HRS and HRO are measured. HLS and HLO are measured in the same manner for the left front loudspeaker 4. The HRS, HRO, HLS, and HLO include an electroacoustic conversion characteristic of the loudspeaker 4 and a spatial acoustic transfer characteristic from the loudspeaker 4 to the dummy head microphone 2 via the space. This HRS, HRO, HLS, H
From LO, CL = −HLO / HLS (1) CR = −HRO / HRS (2) TL = 1 / ((1-CL · CR) · HLS) (3) TR = 1 / ((1− CL · CR) · HRS) (4) The coefficients of the digital filter may be calculated so as to realize the following characteristics CL, CR, TL, and TR.

According to the sound image localization filtering unit 3 configured as described above, when the sound field signal recorded by using the dummy head microphone 2 is reproduced from the loudspeaker 4 to both ears of the listener 5, The sound radiated from the front (left front) loudspeaker cancels the crosstalk component reaching the left ear (right ear), and the sound generated due to the non-flat electroacoustic conversion characteristics and spatial sound transmission characteristics of the loudspeaker 4. It is possible to remove distortion with respect to the amplitude and phase of the signal.

Returning to FIG. 1, the pair of loudspeakers 4
The sound after the filtering process for preventing crosstalk and the like is reproduced by the sound image localization filtering unit 3. As the loudspeaker, various types of speakers such as an electrodynamic type, an electromagnetic type, an electrostatic type, a piezoelectric type, and an electrostrictive type can be used. The left and right loudspeakers are not limited to one each,
Each may use a plurality of loudspeakers. For example, a high frequency reproduction tweeter and a low frequency reproduction woofer 2
It can be composed of loudspeakers. The left and right loudspeakers 4 are preferably arranged so as to be substantially symmetric.

FIG. 3 is a plan view showing an installation position of a pair of loudspeakers in the first embodiment of the binaural reproduction device of the present invention, and FIG. 4 is a side view of the same. FIG.
Although the spread angle α of the pair of loudspeakers 4 as viewed from the listener 5 shown in FIG. 4 is arbitrary, the listener 5 is set so that the distance b between both ears of the listener 5 and the surface position of the loudspeaker 4 becomes smaller. A pair of loudspeakers 4 are installed close to each other.
The distance b is preferably set to be about 1 m or less as described later.

A pair of loudspeakers 4 viewed from the listener 5
Is arbitrary as described above, for example, 12
It can be arranged symmetrically in the front direction of the listener 5 so that it becomes 0 degree. However, in order to maintain good binaural reproduction even when the listener 5 moves his or her head, it is desirable to arrange the left and right loudspeakers 4 close to each other to reduce the spread angle. When a pair of loudspeakers 4 are arranged close to each other, they may be housed in one housing to make one loudspeaker unit apparently. Even in this case, the loudspeaker 4 and the housing may be arranged so that their spatial arrangement and shape are substantially symmetrical on the left and right.

As shown in FIG. 4, the position of the loudspeaker 4 in the height direction may be set so that the height of both ears of the listener 5 and the center of the loudspeaker 4 are the same. Specifically, when the height of the listener 5 is 1.5 m, the center of the tweeter of the loudspeaker 4 can be set to have the same height.

Hereinafter, the effectiveness of the binaural reproduction by the arrangement of the loudspeakers 4 will be described. FIG. 5 and FIG. 6 are explanatory views of an embodiment of the binaural reproducing apparatus of the present invention and an example of an experimental result on a reproduced sound image space in various conventional reproducing apparatuses. 5 and 6, a sound source area of 1-2 m in the depth direction and 2 m in the width direction is assumed in front of the listener, and a sound source such as a machine (printer) is installed at an arbitrary position in the sound source area. In addition to recording the radiated sound of the sound source, the reproduced sound is reproduced by several reproduction methods, and the result of an experiment on how a region (reproduced sound image region) in which the listener perceives the sound image of the sound source is formed is shown.

The example shown in FIG. 5A shows a case of a general stereo system in which recording is performed using a general stereo microphone and reproduction is performed using headphones. In this case, since sound is reproduced from the left and right speakers for the left and right ears of the listener 5, a reproduced sound image area is formed in the head connecting the left and right ears of the listener 5. Also, the example shown in FIG. 5B shows a case of a general stereo system in which recording is performed using a general stereo microphone and reproduction is performed using a loudspeaker. In this case, a reproduced sound image area is formed in a space connecting the loudspeakers. As described above, in the reproduction by the conventional stereo method, the position of the reproduced sound image is localized at the position of the headphone or the loudspeaker, and the listener 5 recognizes that the sound source exists at a position completely different from the actual sound source position. .

The example shown in FIG. 5 (C) shows a case in which reproduction is performed by binaural method using headphones. In this case, the reproduced sound image area is considerably enlarged as compared with the case of reproducing by the stereo method shown in FIG. However, as described above, the sense of localization in the front direction is insufficient, and an in-head sound image localization phenomenon occurs. Therefore, the listener 5 has recognized that the sound source exists at a position different from the actual sound image position.

The example shown in FIG. 6A shows a case in which reproduction is performed in a binaural system using a conventional loudspeaker arrangement. Conventionally, loudspeakers are arranged as far as possible from the listener, and the loudspeakers are arranged so that the angles of the left and right speakers are as wide as possible. FIG.
(B) and FIG. 6 (C) show the configuration of the binaural reproduction device according to the first embodiment of the present invention,
FIG. 6 (B) shows an example in which the loudspeakers are arranged close to the listener 5 and the left and right loudspeakers are separated from each other, and FIG. An example is shown in which left and right loudspeakers are installed close to each other. 6 (A) to 6 (C)
In each of the examples shown in FIGS. 5A to 5C, a reproduced sound image is formed in front of the listener as compared with the examples shown in FIGS.
The actual sound source position almost coincided with the sound source position recognized by the listener.

However, in the case of performing binaural reproduction with the conventional loudspeaker arrangement shown in FIG.
As described above, the influence of the reflected wave due to the floor surface or the like is large, and the original sound may not be faithfully reproduced due to interference with the direct wave. FIG. 7 is a diagram illustrating an example of a difference in reproduced sound quality due to a difference in distance from the loudspeaker to the listener.
Here, a machine serving as a sound source was installed at a position 2 m in front of the listener, loudspeakers were arranged at a spread angle of 120 degrees, and reproduction was performed by the binaural method. FIG.
(A) shows the distance from the listener to the loudspeaker is 2 m.
FIG. 7B shows a case where the distance from the listener to the loudspeaker is set to 70 cm, and the reproduced sound and the original sound are recorded and measured, respectively, and the sound pressure level of the sound is measured. The difference is shown as a frequency characteristic diagram.

The distance from the listener to the loudspeaker is 2
In the case of m, as shown in FIG. 7A, the sound pressure level between the reproduced sound and the original sound has a large difference of about 6 dB in a plurality of frequency bands, and the original sound cannot be faithfully reproduced.
It was confirmed that the region where the large difference occurred coincided with the region where the original sound contained pure tone components. On the other hand, when the listener and the loudspeaker are arranged close to each other and the distance is set to 70 cm as in the binaural reproduction device of the present invention, as shown in FIG. Is reproduced with a difference from the original sound of 1 dB or less, and a sound field reproduction faithful to the original sound can be realized also in terms of sound quality.

As described above, if the loudspeaker 4 is installed at a distance from the listener 5 as shown in FIG. 6A, the original sound may not be faithfully reproduced, and especially the pure sound cannot be faithfully reproduced. However, FIGS. 6B and 6C
By arranging the loudspeaker 4 close to the listener 5 as in the first embodiment of the binaural reproduction device of the present invention shown in FIG. 1, it is possible to faithfully reproduce the original sound including the pure sound. became.

This corresponds to both the ears of the listener 5 and the loudspeaker 4.
By reducing the distance b to the surface position of the loudspeaker, compared to the sound directly propagating from the loudspeaker 4 to the listener 5,
This is because the sound reflected from another path, for example, the floor surface, becomes sufficiently small. The sound pressure decreases in inverse proportion to the square of the propagation distance. By installing the loudspeaker 4 close to the listener 5, the distance from the loudspeaker 4 to the ear of the listener 5 via the floor surface is very small compared to the distance between the loudspeaker 4 and the ear of the listener 5. And the amount of attenuation also increases. As a result, the ratio of the reflected sound can be made negligibly small, and the interference phenomenon due to the reflected sound and the like can be suppressed. For this reason, it is possible to reproduce the original sound faithfully without installing it in an anechoic room or the like.

FIG. 8 is a graph showing an example of an experimental result obtained by simulating the influence of interference between a directly transmitted sound and a floor reflected sound. An experiment was conducted to determine how much the interference phenomenon between the directly transmitted sound and the floor reflected sound affects the sound pressure level of the reproduced sound. FIG. 8 shows the result. As is clear from FIG. 8, the influence of the floor reflected sound increases exponentially as the distance from the listener to the loudspeaker increases. The discrimination threshold of human hearing is 1d
In order to prevent the listener from recognizing a change in the sound pressure level due to the interference phenomenon, the loudspeaker 4 is set so that the distance from the listener 5 to the loudspeaker 4 is 1 m or less. It can be seen that it is sufficient to arrange. In the experimental example in FIG. 7B described above, the distance b between the both ears of the listener 5 and the surface position of the loudspeaker 4 is set to about 70 cm. It is included in a range of 1 m or less, and there is little influence of interference or the like due to a floor reflected sound, and reproduction faithful to the original sound can be performed.

A pair of loudspeakers 4 viewed from the listener 5
The spread angle α is arbitrary as described above. For example, the spread angle α can be opened as shown in FIG. 6B and arranged symmetrically in the front direction of the listener 5. However, in order to maintain good binaural reproduction even when the listener 5 moves his / her head, the left and right loudspeakers 4 are arranged close to each other as shown in FIG. It is desirable to make it smaller.

As described above, the interaural characteristics such as the arrival time difference and the level difference between the left and right ears play a large role in human perception of the sound source position. Therefore, in order to realize good binaural reproduction, it is important to realize and maintain interaural characteristics similar to those of the original sound field. By arranging the pair of loudspeakers 4 in front of the listener 5 in a state of being close to each other, even if the listener 5 moves his or her head to the left and right, a large change that reverses the tendency between both ears does not occur. . That is, if the listener 5 moves his / her head to the left, the distance to both the left and right loudspeakers 4 changes, and the tendency of the sound reaching the left and right ears does not change much. for that reason,
Good binaural reproduction can be maintained even if the listener moves his or her head during reproduction.

FIG. 9 is a graph showing the results of an experiment in which the effect of the listener moving his or her head was measured in the first embodiment of the binaural reproduction device of the present invention. Here, the sound recorded by installing a machine serving as a sound source at a position 2 m in front of the listener 5 was reproduced using the binaural method by the loudspeaker 4. At this time, (a) a case where the distance from the listener 5 to the loudspeaker 4 is 2 m and the spread angle of the loudspeaker is 100 degrees, and
(B) The distance from the listener 5 to the loudspeaker 4 is 70
cm and a loudspeaker with a spread angle of 20 degrees were reproduced with a dummy head microphone. In order to investigate what effect will be exerted on the sound to be reproduced when the listener 5 moves his or her head during reproduction, the sound pressure level is measured by moving the dummy head microphone,
FIG. 9 shows the variation of the sound pressure level. In FIG. 9, the case of the above condition (a) is shown by a hatched graph, and the case of the condition (b) is shown by a white graph. The moving area of the head assumed in the experiment is ± 10 cm in the front-rear direction, ± 10 cm in the left-right direction, and ± 25 degrees in the rotation direction.

As is apparent from FIG. 9, in the conventional arrangement condition (a) of the loudspeaker 4 indicated by hatching, when the listener 5 moves his / her head, a sound pressure level fluctuation exceeding 2 dB occurs. Good binaural reproduction cannot be maintained. On the other hand, under the condition (b) in which the loudspeaker 4 is arranged close to the listener 5 and a pair of loudspeakers are arranged close to each other as in the binaural reproduction device of the present invention, the front and rear of the head In all cases of movement, left-right movement, and rotation, the amount of fluctuation in sound pressure level is suppressed to 1 dB or less, and good binaural reproduction can be maintained. In particular, for such movement of the head, it is effective to arrange the pair of loudspeakers 4 close to each other.

The arrangement condition (a) in FIG. 9 is the case shown in FIG. 6A, and the arrangement condition (b) is the case shown in FIG. 6C. That is, if the left and right loudspeakers 4 are arranged as wide as possible as far as possible and are arranged as far as possible from the listener 5, the binaural reproduction cannot be maintained due to the movement of the head of the listener 5. But,
By placing the left and right loudspeakers 4 close to each other and close to the listener 5, good binaural reproduction can be maintained even when the head of the listener 5 moves.

FIG. 10 is a schematic diagram showing an example of a system including a second embodiment of the binaural reproduction device of the present invention. In the figure, the same parts as those in FIG. 7 is a sound insulating plate. In the second embodiment, the loudspeakers 4 are arranged adjacent to each other with the sound insulating plate 7 interposed therebetween. When the loudspeakers 4 are arranged close to each other, the loudspeakers 4 on the left front (right front)
The crosstalk in which the sound radiated from the sound propagates to the right ear (left ear) of the listener 5 increases. This crosstalk is eliminated by the sound image localization filtering unit 3. However, if the crosstalk becomes large, it becomes difficult to calculate a filter coefficient for eliminating the crosstalk, and good binaural reproduction may be performed depending on the environment of the room where the reproducing apparatus is used. There is a risk of damage. In order to suppress this problem, a sound insulating plate 7 that partitions the front space of the listener 5 left and right between the left and right loudspeakers 4.
Is installed.

Although any material can be used for the sound insulating plate 7, it is necessary to suppress the reflection of sound waves on the surface of the sound insulating plate 7 in order to realize a sound field reproduction faithful to the original sound in terms of sound quality. . Therefore, it is desirable that the surface of the sound insulating plate 7 or the entire sound insulating plate is made of a sound absorbing material made of glass wool, porous plastic, foam material or the like.

FIG. 11 is a plan view showing an installation position of a pair of loudspeakers and a sound insulating plate in a second embodiment of the binaural reproduction device of the present invention, and FIG. 12 is a side view of the same. Also in the second embodiment, the distance b from the surface position of the loudspeaker 4 to the both ears of the listener 5 is within 1 m in the same manner as in the first embodiment with respect to the arrangement position of the loudspeaker 4. It is desirable that the distance be, for example, about 70 cm. The left and right loudspeakers 4 are arranged close to each other, and the distance b from the surface position of the loudspeaker 4 to both ears of the listener 5 is set to 70.
cm and the opening angle α is about 20 degrees. The position of the loudspeaker 4 in the height direction may be set so that the height of both ears of the listener 5 and the center of the tweeter of the loudspeaker 4 are the same. For example, the height C is about 1.5 m. can do.

The sound insulating plate 7 is disposed so as to protrude from the surface of the loudspeaker 4 toward the listener 5 in order to separate sounds radiated from the left and right loudspeakers 4. That is, as shown in FIG. 11, the distance a from the end of the sound insulating plate 7 to both ears of the listener 5 is smaller than the distance b from the surface position of the loudspeaker 4 to both ears of the listener 5. A sound insulation plate 7 is provided. In addition, as shown in FIG. 12, it is preferable to extend vertically. In this example, the sound insulating plate 7 is provided from the floor surface, and the height d of the sound insulating plate 7 is set higher than the height C. Thereby, the listener 5
Can be satisfactorily partitioned left and right, and crosstalk can be reduced. Even if they cannot be completely separated, there is no problem because the sound image localization filtering unit 3 performs signal processing so as to cancel crosstalk.

As a specific example of the sound insulating plate 7, a thickness 10c
m, using a glass wool sound absorbing material having a length (d) of 200 cm and a width of 120 cm, in the height direction up to 30 cm above the upper surface of the loudspeaker 4, and in the direction of the listener 5 from the binaural position of the listener. a can be set to be 30 cm.

Although the space between the ears of the listener from the loudspeaker is left-right asymmetric, a sense of localization in the forward direction can be obtained. However, in order to match the sense of localization in the front direction with the left and right ears, the loudspeaker 4 is required. The spatial relationship between the loudspeaker and the listener's both ears, which is determined from the spread angle and orientation of the speaker, the distance from the loudspeaker to the listener's ear, the spatial position of the sound insulating plate 7, and the like, is symmetrical. Is desirable.

According to the second embodiment of such a binaural reproduction device, similar to the first embodiment,
The original sound can be faithfully reproduced. If the listener 5 does not move beyond the sound insulating plate 7, binaural reproduction can be maintained.

The binaural reproduction device of the present invention as described above can be used as an acoustic evaluation support method for comparing and evaluating the noise of, for example, a plurality of mechanical devices. That is, sounds emitted from a plurality of measurement targets are recorded, and only the comparison target portion is extracted from each recorded acoustic signal and recorded. The recorded audio signals of the comparison target portion are arranged in chronological order and are reproduced by the above-described binaural reproduction device of the present invention. Thereby, the listener can continuously listen to the plurality of comparison target portions and perform the comparative evaluation, and can realize the comparative evaluation without impairing the sense of the listener. At this time, even the pure tone component which has been difficult to faithfully reproduce in the past is faithfully reproduced by the binaural reproduction device of the present invention, so that the listener can perform an accurate evaluation.

[0061]

As is clear from the above description, according to the present invention, the loudspeaker is arranged in front of the listener, and the sound image localization for eliminating the crosstalk caused by such a structure. By providing the filtering unit, it is possible to realize an unspecified number of listeners with a good frontal sound image localization feeling in the forward direction. Further, in the present invention, since the loudspeaker is arranged close to the listener, the phenomenon that the sound directly transmitted from the loudspeaker and the sound reflected by the floor surface interfere with each other in space, and the sound field that is faithful to the original sound even in the sound quality surface Reproduction is possible. Furthermore, by arranging the left and right loudspeakers close to each other, it is possible to maintain good binaural reproduction even when the listener moves his or her head.

Further, according to the present invention, the measurement of the head-related transfer function for an individual, the head must always be moved, and conversely, the head must always be kept at a fixed position, It does not require the restrictions imposed in the past, such as the necessity of using a room, and can achieve a good sense of frontal orientation and faithful reproduction in terms of sound quality, making it easy to implement binaural reproduction. There is an effect that can be.

Furthermore, since the sound source can be faithfully reproduced by using such a binaural reproduction device of the present invention, there is an effect that accurate evaluation can be performed when evaluating the sound source.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram showing an example of a system including a first embodiment of a binaural reproduction device of the present invention.

FIG. 2 is an explanatory diagram showing an example of a measurement state of a sound transfer characteristic from a loudspeaker 4 to both ears of a listener 5;

FIG. 3 is a plan view showing installation positions of a pair of loudspeakers in the first embodiment of the binaural reproduction device of the present invention.

FIG. 4 is a side view showing an installation position of a pair of loudspeakers in the first embodiment of the binaural reproduction device of the present invention.

FIG. 5 is an explanatory diagram of an embodiment of a binaural reproduction device of the present invention and an example of an experimental result on a reproduced sound image space in various conventional reproduction devices.

FIG. 6 is an explanatory diagram of an embodiment of a binaural reproduction device of the present invention and an example of an experimental result on a reproduced sound image space in various conventional reproduction devices.

FIG. 7 is a diagram illustrating an example of a difference in reproduced sound quality due to a difference in distance from a loudspeaker to a listener.

FIG. 8 is a graph showing an example of an experimental result obtained by simulating the influence of interference between a directly transmitted sound and a floor reflected sound.

FIG. 9 is a graph showing an experimental result obtained by measuring an effect when a listener moves his / her head in the first embodiment of the binaural reproduction device of the present invention.

FIG. 10 is a schematic configuration diagram showing an example of a system including a second embodiment of the binaural reproduction device of the present invention.

FIG. 11 is a plan view showing installation positions of a pair of loudspeakers and a sound insulating plate in a second embodiment of the binaural reproduction device of the present invention.

FIG. 12 is a side view showing an installation position of a pair of loudspeakers and a sound insulating plate in a second embodiment of the binaural reproduction device of the present invention.

FIG. 13 is an explanatory diagram of a situation where a human perceives the position of a sound source with both ears in an actual sound field space.

FIG. 14 is an explanatory diagram of an example of a binaural recording / reproducing method in which sound is recorded by a dummy head microphone installed in an original sound field, reproduced by headphones, and listened to by both ears.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Original sound field, 2 ... Dummy head microphone, 3 ... Sound image localization filtering part, 4 ... Loudspeaker, 5 ... Listener, 6 ... DAT, 7 ... Sound insulation board, 11 ... Headphones.

Claims (7)

[Claims] 1. A binaural reproduction device for reproducing a binaurally recorded audio signal, wherein a crosstalk is eliminated from a pair of loudspeakers for reproducing sound and a sound transmission characteristic from a loudspeaker measured in advance to both ears of a listener. A binaural reproduction device, comprising: a sound image localization filtering unit that performs a sound image localization, wherein the pair of loudspeakers are arranged close to a listener. 2. The binaural playback device according to claim 1, wherein the pair of loudspeakers are arranged close to each other. 3. The binaural reproduction device according to claim 1, wherein the layout of the space of the loudspeakers is substantially symmetric in left and right channels. 4. The binaural according to claim 1, wherein a sound insulating plate that partitions a front space of the listener left and right is disposed between the pair of loudspeakers. Playback device. 5. The sound insulating plate according to claim 4, wherein the whole or a part thereof is made of a sound absorbing material.
A binaural playback device according to claim 1. 6. The binaural reproduction device according to claim 1, wherein a distance from a surface of the pair of loudspeakers to both ears of the listener is 1 m or less. 7. A binaural recording of acoustic signals respectively emitted from a plurality of measurement targets, extracting only a comparison target portion from the recorded acoustic signals, and arranging the comparison target portion acoustic signals in a time series. A sound source evaluation support method, wherein the sound source evaluation is reproduced by the binaural reproduction device according to any one of claims 1 to 6.