JP3184901B2 - Method and apparatus for extracting three-dimensional stereotactic information - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to stereophonic sound recording using acoustic conditions of a human head, and more particularly, to a method for extracting stereoscopic information, an apparatus and an application thereof.

[Prior Art] Conventionally, various proposals have been made for a technique of providing three or more microphones in the vicinity of a simulated human head since 1975. That is, a plurality of microphone units are provided for one ear. However, the purpose is roughly divided into directivity control and a recording method for reproducing three or more channels. For example, a technology that enables 4-channel recording by recording in separate microphone units in the front-back direction. And install a plurality of microphones that differ in the direction of light collection,
An MS microphone system that controls directivity by combining multiple signals obtained from these microphones. Another example is a recording method in which a large number of microphones are installed along the periphery of the head.

[Problems to be Solved by the Invention] However, the binaural information obtained by the above-described conventional technique does not have a substantial difference from a normal voice recording by a simulated human head that accurately imitates. In other words, it is only possible to record information limited to the direction of green collection in the audio signal of the normal binaural audio recording. There are also doubts about its principle.

In the prior art, it is considered based on the theory that the direction of binaural information can be limited by capturing the direction of sound incident on the ear canal. However, for example, the sound from the rear sound source goes around the pinna and enters the external auditory canal from the lateral direction, and the amount of incidence is large. Further, a part of the sound from the rear sound source enters the external auditory canal from the front with a wraparound of more than half a turn around the head. In this way, since the ear canal enters the ear canal in a very complicated incident direction, such as interference of the pinna, reflection on the head, and sneaking noise, the entrance direction of the ear canal entrance must not correspond one-to-one to the sound source in the intended localization direction. I haven't. It is also clear that there is no unnecessary sound among such wraparound sounds reaching the various ears, and therefore, lack of sound in any direction results in loss of fidelity. For this reason, it can be pointed out in principle that the conventional method aiming at recording in the above-described direction is an insufficient recording method with low fidelity from the fundamental principle.

On the other hand, as a use of the source recorded in the above-described conventional technique, there is a problem that the use is limited to a headphone reproduction for four channels which satisfies a recording state and a reproduction state under reversible conditions or a more specific multi-channel reproduction. . For example, it is possible to use the four-channel recorded sound limited in the ear canal incident direction for the four channels of a surround or a speaker, but an accurate reversible theory does not hold, and it is difficult to faithfully reproduce the sound.

As described above, in the prior art, only localization information equivalent to that of general binaural recording is obtained.Therefore, it is still considered that the mistaken front-to-rear mixing, insufficient localization, or the degree of fidelity is still a problem. That is the current situation.

An object of the present invention is to remove the above-mentioned problems, extract unique information that occurs exclusively in the human body and can only be obtained by binaural recording, and skillfully uses the extracted information.
The aim is to get a very clear localization. Also,
At the time of extraction, it is also possible to remove the head noise and enhance the head-borne sound by extracting the inherent noise generated in the head and subtracting or adding it, to obtain a clear localization that was not possible in the past The binaural program source which can be obtained is obtained.

[Means for Solving the Problems] In the present invention, in order to solve the above problems, a first signal obtained from a first microphone exclusively for recording a sound affected by the outer shape of a head provided in an outer ear, A second signal obtained from a second microphone that exclusively captures a sound that is not significantly affected by the outer shape of the head placed farther from the head than the first microphone is calculated by a circuit that extracts a difference component. And a method for extracting three-dimensional stereotactic information.

According to another aspect of the present invention, there is provided a reproduction sound field using a speaker that can be repeatedly reproduced under the same conditions. The reproduction microphone has an imitation head, and the first microphone provided in the outer ear of the imitation head affects the outer shape of the head. The recorded sound is recorded (headed condition), and among the above recording conditions, the condition without the imitation human head is recorded. The head is unconditionally recorded, and the recorded (headed) sound and the recorded (headless) sound are synchronized, and a calculation is performed by a circuit that extracts a difference component between the two recorded sounds. The method for extracting three-dimensional stereotactic information described above is provided.

According to another aspect of the present invention, there is provided a first device which is provided at an arbitrary position in an outer ear and exclusively collects a sound affected by a head outer shape.
A sound pickup means, a second sound pickup means for exclusively recording a sound relatively unaffected by the outer diameter of the head, which is installed further away from the head than the first sound pickup means, and a first sound pickup means There is provided a method for extracting three-dimensional localization information, comprising: a sound obtained from the sound generating means; and a synthesizing means for synthesizing the sound obtained from the second sound collecting means into one microphone in opposite phases.

[Operation] In the present invention, the first microphone is installed near the entrance of the ear canal suitable for binaural recording, and the interference, reflection and diffraction sounds generated exclusively on the head are difficult to reach, for example, about 1 cm or more from the entrance of the ear canal. A second microphone is installed at the outer position, and the difference component between the two microphones is calculated (eg, subtracted).
By extracting the sound information, it is possible to extract only the sound information of interference, reflection, and diffraction generated exclusively at the head. Then, by amplifying the extracted sound and adding the amplified sound to a signal obtained from either the first microphone or the second microphone, it is possible to obtain a binaural sound in which localization is remarkably emphasized and clearly heard. On the other hand, a third microphone for the middle ear is provided between the first microphone and the eardrum position. This microphone is a microphone that collects various sounds transmitted through the head, and by extracting a difference component between the first microphone and the third microphone, it is possible to extract only the vibration transmitted through the head. By returning the extracted sound to the signal from the first microphone, noise generated in the head is removed, or important propagation sound peculiar to the head other than the noise contained in the head sound is extracted. By amplifying the head-borne sound and returning it to the first microphone, a new trial binaural program source is obtained that actively contains the localization information that has been difficult to include conventionally. Furthermore, it is also possible to extract only these peculiar sounds transmitted inside the human head and use them as special sounds. In order to pick up this intra-head propagation sound, it is particularly effective for external auditory canal insertion type binaural recording using a living human head.
Therefore, a specific shape of a new ear canal insertion type microphone can be obtained.

[Example] Hereinafter, the present invention will be described with reference to the accompanying drawings. In the drawings, the same numbers are used for members having the same contents as much as possible.

1 is a partial sectional left side view showing an ath embodiment of the present invention, FIG. 2 is a partial sectional front view showing an ath embodiment of the present invention, FIG. 3 is a cross sectional view showing an ath embodiment, FIG. 4 is a partial sectional plan view showing the a-th embodiment, FIG. 5 is a perspective view of the a-th embodiment excluding the imitation human head, and FIG. 6 is an explanatory diagram of the recording area.

The state around the microphone according to the present invention will be described as an a-th embodiment with reference to FIGS. 1 to 6. After that, the a
The circuit for processing the signal obtained from the microphone of the embodiment
An embodiment will be described with reference to FIG. Further, a c-th embodiment and a d-th embodiment, which are modifications of the a-th embodiment, will be described with reference to FIGS. 8 to 15.

Embodiment a: Embodiment a is an embodiment of a microphone system for recording audio according to the present invention. Now, the a-th embodiment will be described in detail while listing the individual conditions of the necessary members and the relationship between the members.

Imitation human head ... It is desirable that the imitation human head a4 imitate the external shape and the inside as much as possible with a living human head. For example, an outer ear whose hardness and shape are accurately imitated, a skull highly related to the transmission of sound in the head, a hollow part of the nose and throat connected to a part of the auditory pathway, and the like are used as necessary. Like the conventional imitation human head, the present invention also imitates the auditory transmission path particularly accurately. The outer ear and ear canal are especially important. Also, a
In the embodiment, the sound transmission state near the eardrum and the bovine is further imitated by selecting the transmission speed of the material.

First microphone: The first microphone a1 is a microphone provided exclusively for recording a sound affected by the outer shape of the head of the imitation human head a4. The intended sound is mainly the sound collected by the outer ear. Furthermore, among the sounds collected by the outer ear, not only the direct sound from the sound source is collected, but also various sounds swirling around the outer periphery of the head that are reflected, interfered, or wrapped around the entire head from the sound source Have a role to collect. Furthermore, the focus of these collected sounds becomes the entrance of the ear canal. Therefore, in the embodiment a, the first microphone a1 is installed at the entrance of the ear canal. However, as long as the first microphone a1 is installed inside the outer ear, it is possible to record the sound affected by the target outer shape of the head although there is a difference in sound collection efficiency.

Second microphone: The second microphone is a microphone provided exclusively for recording a sound that is not affected by the external shape of the head of the imitation human head a4. This effect means the effect that occurs in the head environment described for the first microphone a1. In order to achieve this goal, the following relationship must be satisfied at a minimum.

Regarding the content of the sound affected by the head outline of the imitation human head a4 and the sound that reached the microphone without being affected by the head outline, the audio signal recorded by the second microphone was the first. Content rate is lower than audio signal recorded by microphone a1.

Further, there are the following two means for taking the above relation into account.

# Means: When the direction of the sound source including reverberation to be recorded is unidirectional or limited to a single direction, the distance between the first microphone and the sound source is set to ◇, and the distance between the second microphone and the sound source is set to When □ is used, the difference between ◇ and □ can be made large. This is because the distance difference can be corrected by delaying any one of the microphones under the condition that the sound arrival direction is a single direction, and the distance difference can be substantially eliminated. However, the following conditions must be satisfied.

... The distance difference between ◇ and □ is not preferable if the distance between the microphones is too large because the acoustic loss of the recorded result of the first microphone a1 and the second microphone a5 is large due to propagation loss in the air. For example, if you keep it within 30cm, there is no problem.

... The second microphone is positioned away from the first microphone a1 with respect to the imitation human head a4.

... Both the first microphone a1 and the second microphone use microphone units having the same directivity, and the microphones are installed in the same orientation with respect to the head. These conditions are required when calculating the signal obtained from the first microphone a1 and the signal obtained from the second microphone a1 so that the directivity condition must be the same. This is because, even if signals having different directivities are calculated, only the combined directional characteristics of the two microphones are changed as in the conventional MS type. Also, when using both omnidirectional microphones,
Since the orientation of the microphone units is not significantly affected, the orientation of these microphone units does not need to be strictly matched.
Furthermore, the matching of the directivity is necessary only in the frequency band to be calculated, and does not need to be particularly matched in other frequency bands. For example, if the frequency band for the calculation purpose is 100 Hz to 11 kHz, the frequency from 11 kHz to 100 Hz
The configuration of the present invention can be adopted in any state of directivity in the frequency band lower than.

The second microphone described in the above # means is the second microphone a5 in correspondence with the drawing.

Σ Means: When the sound source to be recorded is in multiple directions regardless of the presence or absence of reverberation, it is necessary to satisfy the condition.

... The second microphone is positioned away from the first microphone a1 with respect to the imitation human head a4.

... It is desirable that the distance between the second microphone and the first microphone 1 is set to be equal to or less than half the wavelength at the upper limit of the frequency range to be extracted.

The above condition is a condition in which the # means and the Σ means are related to each other. When the type of the recorded sound source is not determined in advance, the first microphone a1 and the second microphone a1 are set so as to satisfy both the conditions.
It can be seen that positioning the microphone a2 is preferable.

The above condition is a constraint that occurs because the above delay processing cannot be performed. The reason why delay processing cannot be performed is that if there is only one arrival method of the sound to be recorded, it is possible to correct the phase shift accurately by delay processing. This is because it is impossible to match the phase shift due to the delay processing. For this purpose, each microphone unit must be positioned within a half wavelength of the upper limit frequency to be extracted as described above.

The second microphone described in the above means (2) is the second microphone a2 in each figure. Further, the positional relationship between the first microphone a1 and the second microphone a2 will be described in more detail. If the position of the first microphone a1 is set at a distance of Acm from the head surface excluding the outer ear, the position of the second microphone a2 should be set at a position (A + △) cm with a greater distance than Acm. is necessary. As a result, at least the second microphone a2 can record sound less affected by the imitation human head a4 than the first microphone a1.

By the way, the frequency band affected by the head, which is characterized by binaural sound, has a major effect in the band below 13 kHz. These features are well-known features described in a hearing handbook or the like. The important head effect is the upper limit around 8k, and beyond that is the effect of mere occlusion of the outer ear. This occlusion effect is information that can be recorded relatively accurately even with conventional binaural recording. In addition, the effect of this shielding is not an effect accompanied by wave characteristics due to acoustic diffraction or interference,
It means a shielding effect that is affected only by straightness such as light. On the other hand, there is also a large head effect around 3 kHz. For example, when using a medium having a narrow frequency band such as a telephone line, the frequency upper limit of at least 3 kHz is set as the upper limit of the frequency to be extracted. In this case, the above △ is 5.5c
Setting within m is sufficient. By the way, if 8 kHz is the upper limit of the frequency to be extracted, △ is within 2 cm, and similarly, 13 kHz
Then △ needs to be within 1.3cm. The upper limit of the frequency to be extracted changes as appropriate for each type of media.
Assuming that the upper limit of the average audible frequency of an adult is 16 kHz,
Is within 1 cm, and the microphones are located very close to each other. If the first microphone a1 is placed just around the entrance of the ear canal, the location of the second microphone a2 1 cm away from it is
They come in without going out of the outer ear. As a result, the difference between the signals recorded by these two microphones cannot be sufficiently increased, and the extraction of a difference component described later cannot be sufficiently achieved.

In addition, if only the positional relationship with respect to the outer ear is described as described above, a part with one hand drop occurs. In other words, there is a problem in which direction the distance of △ may be separated. For example, it is assumed that a distance of 設 け is provided to a head (a skull or the like) without separating a distance from the head only by separating from the outer ear. Specifically, when the second microphone a2 is placed near the earlobe of the outer ear, the distance と from the first microphone a1 at the entrance of the ear canal is about 2 cm. In this case, the sound that is collected only depending on the outer ear, for example, the direct sound that is collected by a dent such as a fossa, a scuffer, or a concha that is a part of the outer ear and reaches the entrance of the ear canal reaches only the first microphone a1. Does not reach the second microphone a2. Therefore, this difference is obtained. However, since the distance from the head of the first microphone a1 and the second microphone a2 is almost the same for the sound affected by the entire head, there is almost no difference. Therefore, a signal obtained at the time of extracting a difference component described later has only an influence on the outer ear. Of course, this is sufficiently effective for the purpose of extracting only the sound collection effect in the outer ear. The purpose is diverse and ambiguous. This is because, in the conventional recording method of the binaural relation, there are various grades and ideas, such as no head or with a head and a body, and only an outer ear, and a consistent concept of a dummy head cannot be achieved. This point is not ambiguous in the present invention. On the other hand, it is easy to determine the most preferable state of the dummy head. That is, a human body wearing clothes, having a general hair, and having a general head shape and a body. The first microphone a1 and the second microphone for recording the most desirable multiple sound sources for the ideal imitation human head (human body).
Effective locations for a2 are as follows.

The first microphone a1 is provided at the entrance of the ear canal, and then the installation position of the second microphone a2 is on the line connecting the entrances of both ear canals of both ears.
Installed at a distance of cm. This △ is an important numerical value that limits the upper limit frequency when calculating a difference component described later, and it is desirable that the value be as small as possible. On the other hand, as shown in FIG. Is related to the relative positional relationship. That is, the influence of the shielding of the outer ear changes extremely between the second microphone a2s in which the second microphone a2 is positioned outside the outer ear and the second microphone a2n in which the second microphone a2 is positioned inside the outer ear.
That is, in the second microphone a2s, the difference in the shielding effect is sufficiently extracted when extracting a difference component to be described later. On the other hand, in the second microphone a2n, the difference in the shielding effect is hardly extracted. As described above, the generally preferable numerical value of は is the maximum value of △ when the state of the second microphone a2s slightly protruding from the outer ear, and the upper limit of the frequency required for extracting a difference component described later may be low. If so, use this maximum. On the other hand, if the upper limit of this frequency is 12 kHz, it is about 1.5 cm. This value should slightly change depending on the size of the head, but is a value that generally satisfies the various conditions described above.

In addition, as long as the sound source is a sound reproduced by a speaker, reproduction can be performed many times under the same conditions. Therefore, recording with the first microphone a1 in the presence of the imitation human head 4a, and further recording the first microphone a1 The signal from the first microphone a1 is recorded in a state where the imitation human head 4a is removed, and the separately recorded two signals are subjected to the synchro processing later, whereby the extraction of a difference component described later can be achieved.

Third microphone: The third microphone a3 is a microphone used exclusively for recording the sound propagating inside the imitation human head a4. This microphone is provided between the first microphone a1 and the position of the eardrum in the embodiment a. That is, it is provided in the middle of the ear canal. However, since it is most effective to capture the sound vibration that propagates inside the head and directly arrives at the cow,
The most preferable position of the third microphone a3 is the over cow position (the sixth position).
See the figure). However, when the third microphone a3 is installed, it is expected that the third microphone a3 will be most often used by inserting it into the head of the body. Therefore, in the a embodiment, the third microphone a3 is provided in the middle of the ear canal (see FIG. 6). . Since the sound propagating inside the head excites the wall of the ear canal to further excite the air in the ear canal, there is also a propagation path that excites the eardrum, so it is sufficient to install this third microphone a3 in this ear canal. There is. The third microphone a3 is preferably a microphone having the same characteristics as the first microphone a1. However, from the nature that the sound that is the purpose of sound collection is the sound that has propagated through the head,
In addition to the method in which the microphone is provided in the same direction as the microphone a1, a method in which the microphone is pressed against the wall of the external auditory canal so as to more efficiently collect the propagation sound inside the head can be adopted. The required frequency range may be about the same as the frequency characteristic of the middle sound range in which the propagated sound is transmitted. For example, empirically, it is considered sufficient to secure the range of 400 Hz to 3 kHz. Thereby, the third microphone a3 and the first microphone
The condition that the distance of a1 is within 5.5 cm is obtained. In addition, the distance between the microphones is not strictly set, because the phase of the collected sound due to the difference in microphone position is due to coming through the air medium and the head material medium. This is because it is complicated and cannot be specified. Further, when the ear canal is completely closed by the first microphone a1 located at the entrance of the ear canal, there may be a case where the sound from the air medium does not mix into the third microphone a3. In this case, although the problem of recording with a microphone located several cm outside of the target over cow position toward the outer ear remains, the sound of the third microphone a3 remains as it is, and propagates directly inside the head and directly It can be used as a sound vibration arriving at the cow. Then, in this case, a sound very similar to a sound signal that would be directly vibrated to the excessive position at the excessive position is obtained. The audio signal obtained here is a purely head-propagated sound. Among them, a signal unique to the head-propagation and a signal that is not different from the sound recorded by a microphone outside the head, for example, the second microphone a2 are also included. ing. In other words, speech that arrives after complicated interference due to the skull and otolaryngeal cavity effects is a signal unique to the head, while the sound that propagates through the head is relatively simple. The sound that arrives without distortion of the sound waveform is a signal similar to the second microphone a2, and there is a sound that cannot be distinguished from a signal unique to the propagation sound.

Hemisphere member... A hemisphere member a6 is a sound absorbing material that rotates around the imitation human head a4, and is a member used to limit the recording sound source direction to a single direction. And it has the following features.

It is mainly made of a sound absorbing material, and if necessary, a plate material such as wood can be provided in a sandwich shape in the sound absorbing material. This plate also has the role of a bootstrap that increases the low-frequency absorption.

The rotation of the stereo tone arm is performed in substantially the same manner as the rotation of the stereo tone arm about the approximate center of the imitation human head a4. That is, the vertical and horizontal rotation is performed with the lateral direction regulated. In order to perform this rotational movement, a gimbal type rotation support part a7 is provided. A servo motor for controlling vertical and horizontal rotation and a detection device a8 also serving as a detector for accurately knowing the rotation angle are provided at each rotation portion of the rotation support portion a7.

The center of the rotational movement is most preferably the center position connecting both ears of the imitation human head a4.

The overall shape is a hemisphere with a thickness of 5 cm to 10 cm, and a notch below the support column a 9 when turning
a10, a handle a11 for manual operation on the rear side, and a second microphone a5 on the left and right inner sides, separated from the inner wall of the sound absorbing material by several cm.
Further, an auxiliary cylindrical sound absorbing material a12 can be attached to the front opening side with a double-sided tape or the like.

Embodiment b A circuit for processing a signal obtained from each microphone of the embodiment a will be described with reference to FIG. This embodiment is appropriately referred to as a b-th embodiment.

In the figure, b1, b2, b3 are microphone amplifiers, b4, b5, b6, b7 are filters, b8, b9, b10, b11 are delay circuits, b12, b13 are circuits for extracting difference components, b14, b15, b16 is a mixer, b17, b
18 is an amplifier.

By the processing of the circuit of the b-th embodiment to be described below,
You can get the novel information of the Lord. First, a circuit B1 for extracting a difference component between the first microphone a1 and the second microphone a5 will be described.

The signal captured on the right side of the first microphone a1 is the microphone amplifier b2
Amplified by This amplified signal is defined as β ′. β '
Passes through a filter b5 passing between 100 Hz and 11 kHz, and a signal β ″ delayed by K + X seconds by a delay circuit b10 is input to a circuit b12 for extracting a difference component.

On the other hand, the signal captured on the right side of the second microphone a5 is amplified by the microphone amplifier b1. This amplified signal is referred to as α ′. α ′ passes through a filter b4 passing between 100 Hz and 11 kHz, and α ″ delayed by K seconds by a delay circuit b9 is input to a circuit b12 for extracting a difference component.
Is also connected to another circuit that simply delays by the delay circuit b8 without passing through the filter, and the resulting signal is the output signal out5. The role of the filters b4 and b5 that pass between 100Hz and 11kHz is to narrow down the frequency range in which the difference component is to be calculated, and to predetermine the frequency range that is meaningless even if there are differences in microphone position and directional characteristics. Has the role of cutting. Therefore, it is necessary to change this filter to an arbitrary characteristic depending on the position and characteristics of the microphone.

The output from the circuit b12 for extracting the difference component is amplified by the amplifier b14 at an arbitrary magnification. The result is the output signal out1. The output signal out1 and the output signal out5 are mixed by the mixer b14, and the result becomes the output signal out2.

Now, K is an algebra that is the time when the sound moves from the first microphone a1 on one side to the first microphone a1 on the other side. For example, when the second microphone a5 is installed at a right angle to the sound source and the first microphone a1 is freely rotated, the first microphone a1
The maximum value of the phase difference generated by the positional difference between the right side (left side) of a1 and the right side (left side) of the second microphone a5 corresponds to K. For example, assuming that the linear distance between the binaural microphones of the first microphone a1 is 20 cm, and this distance difference is assumed that the sound moves linearly, about 0.6 msec
Becomes Actually, it often reaches around the head, so it is longer than the above 0.6 msec. On the other hand, in the state of FIG. 8, the distance difference between the right microphone of the first microphone a1 and the right microphone of the second microphone a5 is 6 cm. A few meters to two microphones
The time difference that the sound emitted from the distant sound source arrives is about 0.18
msec. Therefore, X = 0.18 msec. Also this X
Can be calculated by the relative rotation between the first microphone a1 and the second microphone a5. In this embodiment, X is determined by sequentially calculating in accordance with the detection information obtained by the detection device a8 provided on the above-mentioned hemispherical member a6 or the detection device a8 shown in FIG.

The description of the b-th embodiment will be supplemented based on the specific numerical values shown in FIG.

The right side of the first microphone a1 is located 6 cm behind the right side of the second microphone a5 with respect to the sound source. Therefore, the time difference between the signals β and α obtained from the respective microphones is such that β is X seconds later than α with respect to α. To simply eliminate this difference, set the delay time of the delay circuit b9 to 0 seconds,
If the delay time of the delay circuit b10 is X seconds, the above time difference is eliminated. However, since α is slower than β on the left side, the delay circuit on the β side (not shown on the left side) must advance X seconds. However, information can be delayed but not advanced. Therefore, by adding a biasing delay time K to all of the delay circuits of the left and right α and β, the left and right α
And the time difference of all signals in β can be zero.

When the first microphone and the second microphone are at 90 °, the time difference between the two microphones is the largest. Here, if K = 0.6mse, delay circuit
b9 is a fixed delay time of 0.6 ms. Then, the delay circuit b10 adds -0.18 msec in addition to 0.6 msec, resulting in a delay time of 0.42 msec. The α ″ and β ″ obtained as described above have no time lag, and the calculation in the circuit b12 for extracting the difference component becomes smooth.

Next, a supplementary explanation of the circuit b12 for extracting the difference component will be given.
The simplest example of this circuit is a circuit that performs a simple subtraction while keeping an analog signal. This can be easily achieved by using a known circuit, an operational operational amplifier. It is also an effective means that, if desired, an input signal is operated for each of a plurality of spectrums by a plurality of operational amplification operational amplifiers having different amplification factors and their progress is combined. By these calculations, the first signal β obtained from the first microphone a1 exclusively for recording the sound affected by the external shape of the head provided near the entrance of the ear canal, and the first signal β obtained from the first microphone a1 and separated from the head. The second signal α obtained from the second microphone a5 that exclusively records the sound that is not greatly affected by the external shape of the head
And the difference is obtained as a result. Among the signal components of β, which are normal binaural signals, components that are not affected by the head are included. A circuit that removes this signal and extracts the difference component only for the audio signal that is strongly affected by the head.
Retrieved at b12. That is, information specific to the binaural, which is exclusively affected by the head, is extracted.

It can be understood that a binaural signal in which only three-dimensional information generated exclusively in the head is emphasized by the circuit B1 described above.

When the signal of the second microphone a2 is processed by the circuit B1 instead of the second microphone a5, all delay circuits are set to zero, or only the delay circuit b9 connected to the second microphone a2 is used. Delay. It is desirable that the delay time is set to a value that does not exceed the above-mentioned △ cm, that is, 3.3 / △ (sec), the time required for the sound to move. That is, when the sound source is located right beside the head, the time difference between the second microphone and the first microphone is minimized. The time difference is larger by 3.3 / △ (sec) than when the delay circuit is set to zero. Therefore, by taking the middle of these contradictory conditions and setting the delay time to 3.3 / 2 △ (sec), the time difference caused by the distance difference に 対 し て for sound sources in almost all directions is halved, and the difference component can be calculated. The upper limit of the frequency is doubled. Alternatively, take advantage of this and double △ to
The microphone a2 can be further separated from the head.

Next, a circuit B2 for extracting a difference component between the first microphone a1 and the third microphone a3 will be described.

The signal captured on the right side of the third microphone a3 is the microphone amplifier b3
Amplified by This amplified signal is referred to as γ ′. The signal γ 'passes through a filter b7 passing between 400 Hz and 3 kHz and enters one input of a circuit b13 for extracting a difference component. A filter b for passing the signal β 'described in the circuit B1 between 400 Hz and 3 kHz is applied to the other input of the circuit b13 for extracting the difference component.
You are typing through 6. The circuit b13 for extracting the difference component has the same circuit configuration as the circuit b12 for extracting the difference component in the circuit B1. With this circuit, the first signal β obtained from the first microphone (a1) exclusively recording the sound affected by the external shape of the head provided near the entrance of the ear canal, and the head installed inside the head from the first microphone a1 As a result, a difference from the third signal γ obtained from the third microphone a3 that exclusively records the sound propagating inside is obtained. γ is transmitted along the inside of the head including various interference sounds affected by the outer shape of the head similarly to β which is a normal binaural signal, and is collected by the circuit b13 for extracting this difference component Only unique information such as reflections and interferences caused by the human body tissue inside the head that occurs when traveling through the inside of the head is extracted. That is, information that is exclusively affected inside the head is extracted. Circuit b1 for extracting this difference component
The signal output in 3 is amplified by the amplifier b18 at an arbitrary magnification, and the obtained signal is the signal out3.

The signal out3 and the signal β 'are respectively input to the mixer b15, and the outputs thereof are delayed by a fixed K second by the delay circuit b11, and further input to the mixer b16. The signal out2 described in the circuit B1 is further input to the mixer b16. Then, the mixed signal becomes a signal out4.

The filter b6 passes between 400 Hz and 3 kHz. Since the phase difference between left and right is small at 400 Hz or less in the reduction, the value is not so much extracted as a three-dimensional sound signal. In addition, 3k in high frequency
The upper limit of Hz is used to eliminate the frequency difference between the two components when extracting the difference component by restricting the high frequency region that is difficult to propagate by head propagation.

FIG. 9, FIG. 10, FIG. 11, FIG. 12, FIG.
FIG. 3b shows a modification of the a-th embodiment, which is referred to as a c-th embodiment for explanation.

The c-th embodiment is based on the premise that the a-th embodiment mainly uses the imitation human head.
An example of a so-called sonde method that can be used even when worn on a human head will be described.

FIG. 9 is a side view showing the state of use of the c-th embodiment, FIG. 10 is an explanatory view showing a state in which the c-th embodiment is inserted in a human head, and a cross-sectional view thereof, and FIG. It is the front view which showed the external shape of the Example. FIGS. 12, 13a and 13b are explanatory views showing a modified example of the c-th embodiment.

The members provided in the c-th embodiment are substantially the same as those in the a-th embodiment, and the differences will be described below.

The purpose is to enable the special binaural recording of the present invention by inserting it into the human head.

The outer wall c1 is made of a porous material that is considered to be flexible and has little effect on the body even when inserted into the ear canal. Here, polyethylene in a porous state is used. Further, a convex part c3 is provided on the middle part of the outer wall c1. This is a catch for positioning the outer wall c1 at the entrance of the ear canal.

On the other hand, in front of the first microphone a1 and the second microphone a2,
A space c5 is provided between c1 and each microphone. This space also serves as a windscreen. In particular, as shown in FIG. 12, when the external appearance of the c-th embodiment is similar to that of a normal inner-ear type headphone, the position where the normal headphone unit enters can be the space c5. At first glance, it looks like ordinary headphones, but it actually has an outer wall structure provided for a windscreen. Further, a second microphone a2 is supported at one end of the outer wall. FIG. 13a shows a c-th embodiment in which the second microphone a2 is provided on the side wall of the outer ear. The second microphone a2 is simple and has a small number of resonance elements of the support body so that unnecessary vibration sound is hard to enter the microphone by a clip for an outer ear conventionally known as an accessory, and is supported by a clip formed to be attached to a side of the outer ear. The feature is that it can be installed at the tip of the outer ear without adjustment. Further, FIG. 13b shows an example in which the second microphone a2 is supported by ear hanging.
The ear hook is made of a flexible material coated with a polyethylene material around a metal core, and the direction and position of the microphone can be freely and stably performed. In addition, it is characterized in that the same position can be obtained when re-mounting. The operation of the above-mentioned c-th embodiment is the same as that of the a-th embodiment.
A microphone to achieve the same purpose as the means.

FIG. 14 and FIG. 15 are a sectional view and a front view showing a d-th embodiment. The d-th embodiment shows an example in which the difference component can be obtained mechanically without using an electric circuit. That is, this is an example in which the complicated means described in the a-th embodiment and the b-th embodiment can be obtained with only one microphone. The fourth microphone d1 is a microphone unit having an 8-shaped directional characteristic, and has a tube d3 on the inner ear side. This tube d3 extends to a position several cm away from the outer ear, and the opening of this tube d3
d5 is open like a morning glory. Further, it is desirable that the tube d3 be in a tapered state that becomes narrower as approaching the microphone. Now, with this simple structure, the sound incident from the opening d5 includes a sound that is not much affected by the head like the second microphone a2. The sound directly entering the fourth microphone d1 is a sound that is greatly affected by the head as well as the input sound of the first microphone. Then, the tube d3 is connected to the back side of the fourth microphone d1, and the vibration of the microphone is excited in the opposite phase. Therefore, with the above structure, only the difference component of the sound directly entering the opening d5 and the fourth microphone d1 can shake the diaphragm of the fourth microphone d1 and extract the signal.

In the d-th embodiment, not only the difference component can be mechanically extracted, but also the position for picking up the sound corresponding to the second microphone a2 is relatively far from the ear canal because the tube d3 is used. Has the advantage that the phase difference does not increase. In FIG. 14, when there is a sound source to be recorded right to the left, when the second microphone a2 is brought to the opening d5 of the tube d3, there is a time difference due to the distance d7 with respect to the fourth microphone d1. In the d-th embodiment, only the distance d8 at which the tube turns to the back side of the fourth microphone d1 appears as a time difference, and there is an advantage that the phase difference is small. However, when there is a sound source to be recorded in front, there is no difference. Therefore, the difference can be eliminated in all directions by dispersing a large number of pipes d3 not only right beside but also dozens of directions. Then, by providing the sound beyond the inside of the head, for example, as indicated by a dotted tube d3, it becomes possible to record the sound ignoring the presence of the head. It is possible to eliminate the influence of the head that did not appear, and it is possible to extract more accurate three-dimensional information.

By the way, finally, using the circuit of the above-mentioned a-th embodiment and the circuit of the b-th embodiment, an arbitrary direction impulse signal sound source is arranged, and various impulse signals obtained by the b-th embodiment are obtained. The output signal is stored in a DSP circuit or the like to configure a transfer characteristic compensation circuit. By preparing this transfer characteristic compensation circuit for each required direction and installing it between the effector processing circuits such as the mixer, the reverberation of currently available DSP for signals that do not have three-dimensional information such as line recording sound It can be used as a new effector that can add direction information as easily as adding.

[Effects of Embodiment] As described above, the first microphone that records the sound that is more affected by the head acoustically, and the second microphone that is less affected by the head acoustically than the first microphone This is a microphone for extracting three-dimensional stereo localization information, characterized by comprising:
Therefore, it is possible to immediately subtract the signals from these two types of microphones, extract the difference signal, and extract only the audio signal affected by the acoustic influence of the head. These 2
This microphone makes it possible to record the signals from the various types of microphones with a multi-channel recorder or the like, and then perform the extraction of the difference component at any time thereafter.

Also, a three-dimensional stereo system comprising: a third microphone for collecting green voice that has received a large amount of sound propagating inside the head; and a first microphone that receives a small amount of voice that propagates inside the head than the third microphone. This is a microphone for extracting localization information. Therefore, by immediately subtracting the signals from these two types of microphones and extracting the difference signal, it is possible to extract only the audio signal propagating inside the head. Also, the signals from these two types of microphones are collected by a multi-channel recorder or the like, and thereafter, the recording that can extract the difference component at any time can be performed by the microphones.

Further, a phase difference between two slightly different audio signals is canceled by a delay circuit, and a difference component for performing processing such as subtraction in a state where the phases are matched is passed through a circuit for extracting a difference component, thereby enhancing the extracted signal. This is a circuit for extracting three-dimensional stereo localization information in which an amplifier is provided and an enhanced extracted signal is added to either of the two signals. Therefore, by processing signals from various microphones described in the a-th embodiment and the c-th embodiment, for example, an audio signal having two major characteristics can be obtained by this circuit. One is a special binaural sound that arbitrarily exaggerates the acoustic effects caused by the outer shape of the head, and the other is a special binaural sound that arbitrarily exaggerates the acoustic effects caused by propagating inside the head. Binaural audio is obtained.

Now, by combining the above two kinds of special binaural, the information about the outer circumference of the head and the information propagating inside are extracted arbitrarily, and this extracted signal is concentrated by means of amplification, made stronger, and finally The signal from the most normal microphone, for example, the second with the least acoustic influence of the head
By adding this concentrated sound signal to the microphone, an intense program source with a localization that is impossible in the past can be obtained. Until now, it is possible to obtain a signal in which localization is emphasized, which can cover insufficient localization due to the listener and the head shape error at the time of recording, and also has a useful effect of solving the problem of individual differences in binaural. In order to obtain the above effects, a binaural microphone including the first microphone and the second microphone or the first microphone and the third microphone is provided.

[Effect of the Invention] According to the invention defined in claim 1 of the present invention, the sound affected by the external shape of the head provided in the outer ear is removed from the sound that is not greatly affected by the external shape of the head. Sound affected by the outer shape, that is, rich three-dimensional stereo localization information is extracted. The extracted three-dimensional stereo localization information can be used as it is as a sound source whose stereo information is clear, or a new stereo sound source can be provided by adding an appropriate amount to another sound source.

According to the invention defined in claim 4 of the present invention, by removing the sound that is not affected by the head outline at all from the sound affected by the head outline, the sound that is exclusively affected by the head outline, that is, Complete and rich three-dimensional stereotactic information is extracted. This extracted three-dimensional stereo localization information can be used as it is as a sound source with clear stereo information, or
By adding an appropriate amount to another sound source, a new three-dimensional sound source can be provided.

According to the invention defined in claim 5 of the present invention, a sound which is not greatly affected by the outer shape of the head is subtracted as a sound wave from a sound affected by the outer shape of the head provided in the outer ear, and is recorded by a microphone. This eliminates the need for a circuit for extracting the difference component. As a result, cost reduction can be realized without being affected by circuit quality.

[Brief description of the drawings]

FIG. 1 is a partially sectional left side view showing an a-th embodiment of the present invention.
FIG. 2 is a partial sectional front view showing an ath embodiment of the present invention, FIG. 3 is a sectional view showing an ath embodiment, FIG. 4 is a partial sectional plan view showing an ath embodiment, and FIG. FIG. 6 is a perspective view of the a-th embodiment excluding the imitation human head, FIG. 6 is an explanatory view of a recording area, and FIG.
FIG. 13 is a block circuit diagram showing a b-th embodiment, which is an example of processing a signal obtained from each microphone of the a-th embodiment.
Figures are explanatory diagrams showing the positional relationship of microphones, FIG. 9, FIG.
FIGS. 11, 12, 13a and 13b are explanatory views showing a modification of the a embodiment, FIG. 14 is a sectional view showing the d embodiment, and FIG. 15 is a d embodiment. FIG. In the figure, a1 ... first microphone, β ... first signal, a2, a5 ... second
Microphone, α second signal, b12, b13... Circuit for extracting difference components, a3... Third microphone, γ... Third signal, a6.

Claims (5)

(57) [Claims] 1. A first signal (β) obtained from a first microphone (a1) exclusively for recording a sound affected by an external shape of a head provided in an outer ear, and a head from the first microphone (a1). A second signal (α) obtained from a second microphone (a2, a5) that exclusively records a sound that is not significantly affected by the head outer shape that is set apart from a second component (α) and a circuit that extracts a difference component ( A method for extracting three-dimensional stereotactic information characterized by performing the calculation in b12). 2. A circuit for extracting at least a difference component (b12)
3. The three-dimensional microphone according to claim 1, wherein in the frequency band calculated in step (1), microphones having omnidirectional characteristics are used for both the first microphone (a1) and the second microphone (a2, a5). Extraction method of stereotaxic information. 3. A circuit for extracting at least a difference component (b12)
In the frequency band calculated by the above, it is assumed that the directional characteristics of the first microphone (a1) and the second microphone (a2, a5) are both directional and that the microphone directions are set in substantially the same direction. The method for extracting three-dimensional stereotactic information according to claim 1, characterized in that: 4. In a reproduction sound field by a speaker which can be repeatedly reproduced under the same conditions, the head has an imitation human head, and is affected by the outer shape of the head by a first microphone (a1) provided in an outer ear of the imitation human head. By recording the sound (headed condition) and setting the above recording condition to the condition without imitation human head,
The sound which is not affected by the head outline of the imitation human head than the recording is recorded (headless condition), and the recorded (headed condition) sound and the recording (headless condition)
A method for extracting a difference component between the two recorded sounds by performing a synchronization process on the extracted sounds and calculating the difference component between the two recorded sounds by a circuit (b12). 5. A first sound pickup means (d1) which is provided at an arbitrary position in the outer ear and exclusively picks up a sound affected by the outer shape of the head, and a head which is higher than the first sound pickup means (d1). A second sound pickup means (d5) exclusively recording sound relatively unaffected by the outer diameter of the head, which is set apart from the head, a sound obtained from the first sound pickup means (d1), Synthesizing means for synthesizing the sound obtained from the second sound collecting means (d5) into one microphone in the opposite phase, and extracting the three-dimensional stereotactic information.