JP7511199B2 - 3D Sound System - Google Patents

Description

The present invention relates to a stereophonic system, a recording device, and a reproduction device.

Conventionally, the Ambisonics method has been known as a method for recording and reproducing three-dimensional surround sound. Non-Patent Document 1 explains that Ambisonics is a method that has evolved into a hierarchical structure from monophonic to two-channel stereo, horizontal plane surround, and three-dimensional sound based on the directional expression of sound incidence, and that spherical harmonic functions have been introduced to express and reproduce three-dimensional sound fields. In other words, Ambisonics is a method in which sound signals recorded by multiple microphones arranged in an audio recording space are hierarchically expanded using spherical harmonic functions in relation to the direction of incidence of the wavefront, and the resulting spherical harmonic functions are used to reproduce surround sound using multiple speakers arranged in the audio reproduction space.

Patent Document 1 discloses, as an example of the Ambisonics method, a sound field reproduction system equipped with a sound field reproduction device, a scale model, a sound source, and a speaker, and discloses A format and B format as sound collection methods in a real sound field, with A format using four cardioid microphones arranged at the vertices of a regular tetrahedron, or a method using a multi-channel microphone installed on the surface of a sphere, and B format using a bidirectional microphone arranged near the center of the microphone unit, one each for the x-, y-, and z-axis directions, and one omnidirectional microphone.

As a method other than the Ambisonics method, Non-Patent Document 2 describes a six-channel recording/playback system developed as a tool for subjective sound evaluation experiments. Non-Patent Document 2 describes that this system is designed to reproduce three-dimensional sound fields in laboratory experiments, and that six unidirectional microphones arranged at 90-degree intervals are used to record sound in an actual sound field, that the playback system involves installing six speakers in an anechoic chamber and playing back recorded signals from each direction, and that this system can be expanded into a general audio recording/playback system.

JP 2019-161455 A

Akio Ando: Highly realistic sound technology and its theory, IEICE Fundamentals Review, Vol.3, No.4, pp.33-46 (2010) S. Yokoyama et al.: 6-channel recording/reproduction system for 3-dimensional auralization of sound fields, Acoustical Science and Technology 23, 2, pp.97-103 (2002)

In the Ambisonics method described in Non-Patent Document 1, in order to reproduce stereophonic sound based on spherical harmonic functions, it is necessary to increase the number of microphones in the sound recording space and arrange them closely together when reproducing the incident direction of the wavefront with high accuracy, and there is a problem that the stereophonic sound system becomes complicated and large. Therefore, an increase in the number of microphones and speakers leads to an increase in the calculation load of the spherical harmonic function, so there is a problem in achieving both recording and reproduction of highly realistic stereophonic sound and a simple stereophonic sound system with low calculation load. Furthermore, there are results of psychoacoustic research that show that the perception of stereophonic sound can be explained by parameters such as interaural time difference, interaural level difference, and interaural envelope time difference in the high frequency range. Therefore, there is a possibility that the high-accuracy wavefront reproduction that requires the conventional system to be complicated and large-sized is not necessarily the only means from the viewpoint of recording and reproduction of highly realistic stereophonic sound.

On the other hand, in the system described in Non-Patent Document 2, a total of three pairs of microphones or speakers are used at vertices facing each other through the center of a regular octahedron to form the orthogonal wavefront incidence directions in the x, y, and z axes in a Cartesian coordinate system, and the vector sum of these is used to express the incidence direction of any wavefront, so that the recorded sound signal can be played back as is, making it possible to easily reproduce the sound recording space in three dimensions with little computational load.

However, in the system described in Non-Patent Document 2, because the wavefront incidence direction is expressed by a total of three pairs of microphones or speakers that cross the space at right angles, there is a problem that when expressing a sound field radiated from two or more highly correlated sound sources placed at different positions, a sound field different from the actual sound recording space may be reproduced.

The present invention focuses on the above problem and aims to provide a stereophonic system that can easily record and reproduce highly realistic stereophonic sound with a low computational load.

The present invention includes multiple means for solving the above problems, one example of which is as follows:

The stereophonic system of the present invention comprises a recording section in which 12 microphones are arranged three-dimensionally in an acoustic recording space, and a reproduction section in which 12 speakers are arranged three-dimensionally in an acoustic reproduction space. The reproduction section reproduces the acoustic recording space based on the acoustic signals recorded by the recording section, and the acoustic recording polyhedron formed by connecting the arrangement points of the 12 microphones and the acoustic reproduction polyhedron formed by connecting the arrangement points of the 12 speakers are roughly the same shape and are cuboctahedrons.

According to the present invention, since the sound recording polyhedron and the sound reproduction polyhedron are roughly the same type, the geometric time-space information of the sound recording polyhedron contained in the recorded sound signal can be reproduced in the sound reproduction space by playing back the sound signal as is or with minor corrections.

In addition, according to the present invention, the direction of incidence of a wavefront is determined by a total of six pairs of microphones or speakers at vertices facing each other through the center of a cubic octahedron, and the direction of incidence of any wavefront is expressed by the vector sum, making it possible to express wavefronts coming from two sound sources that can be recognized by binaural hearing. Therefore, it is possible to easily record and reproduce highly realistic three-dimensional sound with a small computational load.

Problems, configurations, and advantages other than those described above will become clear from the description of the embodiments below.

1 is a schematic diagram showing a sound recording polyhedron and a sound reproduction polyhedron of a stereophonic system according to a first embodiment, and the relationship between them; FIG. 1 is a perspective view illustrating an example of a recording device according to a first embodiment. 1 is a schematic configuration diagram showing an acoustic reproduction polyhedron according to Example 1 that is formed in an acoustic reproduction space having a rectangular parallelepiped shape. FIG. FIG. 2 is a schematic diagram showing a stereophonic model according to the first embodiment, which corresponds to a cuboctahedron; FIG. 2 is a schematic diagram showing the rotational symmetry of the stereophonic sound system according to the first embodiment. 1 is a schematic diagram showing a covariance matrix relating to rotational symmetry of a 24×24 matrix of audio signals in an audio recording polyhedron and an audio reproduction polyhedron of the stereophonic system according to Example 1. FIG. FIG. 11 is a schematic diagram showing a sound recording polyhedron and a sound reproduction polyhedron of a stereophonic system according to a second embodiment, and the relationship between them. FIG. 11 is a schematic diagram showing an expression of a large-scale sound reproduction space by connecting sound reproduction polyhedrons in a stereophonic system according to a second embodiment. FIG. 11 is a schematic diagram showing a sound recording polyhedron and a sound reproduction polyhedron of a stereophonic system according to a third embodiment, and the relationship between them. FIG. 13 is a schematic diagram showing a sound recording polyhedron and a sound reproduction polyhedron of a stereophonic system according to a fourth embodiment, and the relationship between them. 13 is a schematic diagram showing the configuration of a sound recording polyhedron and a dual sound recording polyhedron, as well as a sound reproduction polyhedron and a dual sound reproduction polyhedron of a stereophonic system according to Example 5. FIG. 13 is a schematic diagram showing the configuration of a sound recording polyhedron and a dual sound recording polyhedron, as well as a sound reproduction polyhedron and a dual sound reproduction polyhedron, of a stereophonic system according to a sixth embodiment. FIG. 13 is a perspective view showing a recording device as a specific example of the recording section of FIG. 12. FIG. 13 is a perspective view showing a reproduction device as a specific example of the reproduction unit in FIG. 12 . 13 is a perspective view showing a sound reproduction polyhedron formed in a rectangular parallelepiped sound reproduction space of a stereophonic sound system according to a sixth embodiment. FIG. FIG. 13 is a schematic diagram showing a physical model in a two-dimensional cross section of an acoustic space filled with rhombic dodecahedrons in a stereophonic system according to a sixth embodiment. 13 is an eigenmode image showing the analysis result of a two-dimensional acoustic space using a physical model of the acoustic space by the stereophonic system according to Example 6. FIG. 13 is a schematic diagram showing a sound recording polyhedron and a sound reproduction polyhedron of a stereophonic system according to a seventh embodiment, and the relationship between them. FIG. 13 is a perspective view showing a specific example of a speaker enclosure that is a part of the reproduction device of the stereophonic sound system according to the seventh embodiment. FIG. 1 is a diagram showing the configuration of a conventional stereophonic system using the Ambisonics method. FIG. 1 is a schematic diagram showing the relationship between the sound recording polyhedron and the sound reproduction polyhedron of a conventional stereophonic system that employs regular octahedrons as the sound recording polyhedron and the sound reproduction polyhedron.

The present invention relates to a stereophonic system that records stereophonic sound and reproduces the stereophonic sound based on the recorded sound signals, as well as a recording device and a reproduction device used therein.

In this embodiment, an embodiment of a stereophonic sound system according to the present invention will be described with reference to the drawings. Note that the same elements in each drawing will be denoted by the same reference numerals, and duplicated explanations will be omitted.

First, the first problem that the present invention aims to solve will be explained in detail using Figure 20.

Figure 20 shows a conventional 3D sound system using the Ambisonics method described in Patent Document 1.

As shown in this diagram, a conventional 3D audio system first uses an Ambisonics encoder 4 to hierarchically expand an audio signal 3 recorded using multiple microphones 2 installed in an audio recording space 1 using spherical harmonic functions in relation to the direction of incidence of the wavefront, obtaining an Ambisonics audio signal 5. Next, an Ambisonics decoder 7 generates a reproduced audio signal 8 so that the audio signal 3 recorded in the audio recording space 1 is reproduced in an audio reproduction space 6, and the reproduced audio signal is reproduced by multiple speakers 9 installed in the audio reproduction space 6.

As described above, since the Ambisonics method expresses the incident direction of a wavefront using spherical harmonic functions, when reproducing the incident direction of a wavefront with high accuracy, it is necessary to increase the number of microphones 2 in the sound recording space 1 and arrange them densely to record the sound signal 3, and also increase the number of speakers 9 in the sound reproduction space 6 and arrange them densely. This leads to an increase in the complexity and size of the stereophonic sound system related to recording and reproduction, and an increase in the calculation load of the spherical harmonic functions in the Ambisonics encoder 4 and the Ambisonics decoder 7. For this reason, there has been a challenge in achieving both the recording and reproduction of highly realistic stereophonic sound and a simple stereophonic sound system with a low calculation load.

Next, the second problem that the present invention aims to solve will be explained in detail using Figure 21.

Figure 21 shows a conventional 3D sound system based on the method described in Non-Patent Document 2.

The stereophonic system shown in this diagram is composed of a recording unit 10 that records by arranging six microphones 2 three-dimensionally in an acoustic recording space 1, and a reproduction unit 12 that reproduces the acoustic recording space 1 by arranging six speakers 9 three-dimensionally in an acoustic reproduction space 6 based on the acoustic signal 3 recorded by the recording unit 10. An acoustic recording polyhedron 11 is formed by connecting the arrangement points of the six microphones 2 that make up the recording unit 10. An acoustic reproduction polyhedron 13 is formed by connecting the arrangement points of the six speakers 9 that make up the reproduction unit 12. The acoustic recording polyhedron 11 and the acoustic reproduction polyhedron 13 are both roughly isomorphic to a regular octahedron 52.

According to this method, a pair of microphones 2 or speakers 9 in an opposing relationship 53 through the center of a regular octahedron form the orthogonal wavefront incidence directions in the X-, Y-, and Z-axis directions in the Cartesian coordinate system, and the vector sum of these forms the incidence direction of any wavefront. This makes it possible to reproduce the sound recording space 1 three-dimensionally with ease and low computational load by simply playing back the recorded sound signals as is. On the other hand, when expressing a sound field radiated from two or more highly correlated sound sources placed at different positions, there is a problem in that a sound field different from the actual sound recording space 1 may be reproduced.

In consideration of these problems, the stereophonic system disclosed in this specification, as shown in FIG. 1, which will be described in detail later as an embodiment, has been devised based on the basic idea of setting the number of microphones 2 constituting the recording section 10 in the sound recording space 1 to 12 or more, and making the sound recording polyhedron 11, which is a geometric arrangement, and the sound reproduction polyhedron 13, which is a geometric arrangement of speakers 9 constituting the reproduction section 12 in the sound reproduction space 6, approximately the same type 15. Here, the sound recording polyhedron 11 and the sound reproduction polyhedron 13 are approximately the same type 15 as polyhedrons composed of 12 or more vertices, for example, cuboctahedrons composed of 12 vertices, so as to be able to express wavefronts arriving from two sound sources that can be recognized by binaural hearing.

The following examples are explained with reference to the drawings.

In this specification, "isomorphism" refers to a geometric relationship that includes "congruence" and "similarity" in the mathematical sense. "Approximately the same type" means that two corresponding vertices of two figures are roughly the same type, and even if the angle between the coordinate vectors of the two vertices is not 0 (zero) in polar coordinates (spherical coordinates) with the center of gravity of the figure as the origin, the two figures are considered to be "approximately the same type." Here, it is desirable for the angle between the coordinate vectors of the two vertices to be within 15 degrees. Therefore, in this specification, when the angle between the coordinate vectors of the two vertices is within 15 degrees, the concept of the two figures being "approximately the same type" is included.

The term "almost congruent" used in the examples below also means roughly "congruent" like the above "almost identical type". For example, two corresponding vertices of two figures are "almost congruent" even if the angle between the coordinate vectors of the two vertices is not 0 (zero) in polar coordinates (spherical coordinates) with the center of gravity of the figures as the origin. Here, it is desirable for the angle between the coordinate vectors of the two vertices to be within 15 degrees. Therefore, in this specification, when the angle between the coordinate vectors of the two vertices is within 15 degrees, it is considered to be included in the concept that the two figures are "almost congruent".

Figure 1 shows the stereophonic system of Example 1.

As shown in this diagram, the stereophonic system is composed of a recording unit 10 that records by arranging 12 microphones 2 in a three-dimensional manner in a sound recording space 1, and a reproduction unit 12 that reproduces the sound recording space 1 by arranging 12 speakers 9 in a three-dimensional manner in a sound reproduction space 6 based on the sound signal 3 recorded by the recording unit 10. A sound recording polyhedron 11 is formed by connecting the arrangement points of the 12 microphones 2 in the recording unit 10. Furthermore, a sound reproduction polyhedron 13 is formed by connecting the arrangement points of the 12 speakers 9 in the reproduction unit 12. The sound recording polyhedron 11 and the sound reproduction polyhedron 13 are both roughly isomorphic to a cuboctahedron 15.

In this embodiment, the recorded sound signal 3 contains spatial and temporal geometric information of the sound recording polyhedron 11 related to the sense of realism. Therefore, when the distortion between the sound recording polyhedron 11 and the sound reproduction polyhedron 13 is small, the recorded sound signal 3 can be reproduced as is. When the distortion between the sound recording polyhedron 11 and the sound reproduction polyhedron 13 is large or when the difference in directivity between the microphone 2 and the speaker 9 cannot be ignored, the stereophonic sound of the sound recording space 1 can be reproduced in the sound reproduction space 6 with a simple and low calculation load by performing the reproduction after minor corrections related to the distortion. In addition, the incident direction of the wavefront is formed by a total of six pairs of microphones 2 or speakers 9 whose vertices face each other through the center of the cubic octahedron, and the incident direction of any wavefront is expressed by the vector sum. Therefore, it is possible to express the wavefront coming from two sound sources that can be recognized by binaural hearing.

Figure 2 shows a recording device as a specific example of the recording unit 10 in Figure 1.

In FIG. 2, twelve acutely oriented microphones 2 are mounted on a microphone holder 17, with each microphone 2 positioned at a vertex of a cuboctahedron. The sound receiving portion of the microphone 2 faces outward. The microphone holder 17 is supported by a microphone stand 16. This allows the twelve microphones 2 to be radially positioned so as to be roughly the same shape as the cuboctahedron, and allows the wavefront incident on the center of the polyhedron to be recorded three-dimensionally.

Figure 3 shows an acoustic reproduction polyhedron formed in a rectangular parallelepiped acoustic reproduction space.

As shown in this figure, a normal room (acoustic reproduction space) such as a living room has a rectangular parallelepiped shape, so the recorded acoustic signal can be reproduced by an acoustic reproduction polyhedron 13 that corresponds to the rectangular parallelepiped shape. Even if the shape corresponds to a rectangular parallelepiped shape, it can be said to be roughly isomorphic to a cuboctahedron. The cuboctahedron shown in Figure 1 is a solid body in which the eight vertices of a cube are cut off to the midpoints of the sides, so it has the characteristic that it is easy to configure the acoustic reproduction polyhedron 13 into a rectangular parallelepiped room.

Figure 4 shows a cuboctahedron and the corresponding spatial audio model.

In this figure, the acoustic element core 18 is a spherical acoustic element that serves as the core, and 12 spherical acoustic elements 19 are arranged in contact around the core to form a close-packed structure.

When energy is injected into a certain point in a three-dimensional acoustic space, a propagation phenomenon of acoustic energy occurs due to the expansion and compression of the medium, centered around that point. As the minimum components for expressing such an acoustic phenomenon, we consider a spherical acoustic element core 18 as the energy injection point, and a similarly spherical acoustic element 19 as the target of propagation of the acoustic energy. The three-dimensional acoustic space is then considered to be a homogeneous field in which the acoustic element cores 18 and acoustic elements 19 are densely packed.

Then, the arrangement of the acoustic element core 18 and the acoustic elements 19 will be as shown in the figure. In this case, the geometric structure formed by connecting the centers of gravity of the surrounding acoustic elements 19 will be a cuboctahedron.

In this embodiment, the stereophonic system is constructed based on a cubic octahedron. Therefore, the stereophonic system of this embodiment corresponds to a stereophonic model based on geometric thinking.

Figure 5 is a schematic diagram showing the rotational symmetry of the stereophonic system of this embodiment.

In this diagram, a cuboctahedron 20 has rotation axes 21, 22, and 23 on the equilateral triangular faces, square faces, and vertices. The rotation axes 21, 22, and 23 are representative examples of axes. Considering the positive and negative directions of the axes, there are four rotation axes 21 that pass through the centers of gravity of the equilateral triangular faces, three rotation axes 22 that pass through the centers of gravity of the square faces, and six rotation axes 23 that pass through the vertices.

These rotation axes 21, 22, and 23 represent the rotational symmetry axes in the sound recording polyhedron and sound reproduction polyhedron of the stereophonic system. In the stereophonic system of this embodiment, calculations can be performed that utilize the rotational symmetry of a cuboctahedron.

The cuboctahedron has 8 rotations every 120 degrees through the center of gravity of the equilateral triangular faces ( _T1 , _T12 , ^T2 , _T22 , T3, T32, _T4 , _T42 ), 9 rotations every 90 degrees through ^the center of gravity of the square faces ( ^S1 , _S12 ^, _S2 , _S22 , ^S3 , _{S32 , S4} , ^S42 ), 6 rotations every ₁₈₀ degrees through the vertices ( _V1 , _V2 , _V3 , ^V4 _, ^V5 , _V6 ), _and the identity transformation I, _{for a} total _of 8 + 9 + ₆ + 1 = 24 rotational symmetry transformations.

As an example of a calculation that uses this rotational symmetry transformation, let us first consider 3D panning. Here, we refer to this method, which uses a group of rotational transformations, as the GBAP method (Group Based Amplitude Panning).

The acoustic signal 3 recorded by the cuboctahedral recording unit 10 shown in FIG. 2 is expressed by the following matrix (Equation (1)).

Here, each row corresponds to the acoustic signals 3 recorded by the 12 microphones 2, and the length n of the columns corresponds to the number of time samples, so the matrix X is a 12×n matrix. Now, if one of the rotations _T1 passing through the center of gravity of the face of an equilateral triangle is performed on the matrix X in 120 degree increments, the matrix will have a form in which the elements of the rows are replaced according to the rotation, for example, as expressed by the following formula (2).

Therefore, in the GBAP method, if the virtual stereo sound angle φ [rad] and the maximum stereo sound angle φ ₀ = π/3 [rad] are defined based on the intermediate angle of 60 degrees (= π/3 [rad]) of a 120 degree rotation, the gain w _I of matrix X and the gain w _T1 of T ₁ (X) have the relationship expressed by the following equation (3), for example, according to the sine law.

Or, similarly, according to the tangent law, the relationship is expressed by the following equation (4).

Any of these can be used to achieve diverse 3D panning using 24 types of rotationally symmetric transformations as a natural extension of amplitude panning.

Next, as another example of a calculation using a rotationally symmetric transformation, we will consider the covariance matrix related to the rotational symmetry of the recorded acoustic signal 3.

Normally, the covariance matrix is calculated by calculating the covariance between the acoustic signals 3 recorded by each microphone 2, and the result is used, for example, to analyze the acoustic recording space or correct the acoustic reproduction space. However, its meaning is statistical and difficult to interpret physically.

In contrast, the 3D audio system of this embodiment utilizes rotational symmetry, making it possible to introduce a geometric interpretation into the covariance matrix values. That is, for the matrix X of 24 types of audio signals 3 that have been subjected to 24 types of rotational symmetry transformation, if each is converted into a one-dimensional vector and then a covariance matrix calculation is performed, a covariance matrix related to the rotational symmetry of a 24 x 24 matrix is obtained.

Figure 6 shows a simplified covariance matrix for the rotational symmetry of a 24x24 matrix.

As shown in this figure, each element of the matrix shows the dispersion relations for 24 types of rotationally symmetric transformations: one type 24 of identity transformation, eight types 25 of rotations in increments of 120 degrees through the center of gravity of an equilateral triangular face, nine types 26 of rotations in increments of 90 degrees through the center of gravity of a square face, and six types 27 of rotations in increments of 180 degrees through a vertex. From this, we can obtain a geometric interpretation of the 24 x 24 types of rotational symmetry.

As a result, even if the acoustic reproduction polyhedron 13 configured in the acoustic reproduction space 6 is distorted relative to a cubic octahedron as shown in FIG. 3, or if the difference in directivity between the microphone 2 and the speaker 9 cannot be ignored, it is possible to reproduce a highly realistic sound field by performing a geometric correction so that the covariance matrix related to the rotational symmetry obtained in the acoustic reproduction space 6 coincides with the covariance matrix related to the rotational symmetry obtained in the acoustic recording space 1.

Although the rotational symmetry transformation has been described here as a symmetry operation, the 3D audio system of this embodiment can also use a reflection symmetry transformation using a mirror plane 352, which is a plane that includes a rotation axis 21 passing through the center of gravity of an equilateral triangle and a rotation axis 22 passing through the center of gravity of a square face, as shown in FIG. 5. This makes it possible to express the reflection of sound, and by calculating the covariance matrix related to the reflection symmetry, a geometric interpretation of spatial sound reflection can be obtained.

Figure 7 shows the stereophonic system of Example 2.

As shown in this diagram, the stereophonic system is composed of a recording unit 10 that records by arranging 14 microphones 2 in a three-dimensional manner in a sound recording space 1, and a reproduction unit 12 that reproduces the sound recording space 1 by arranging 14 speakers 9 in a three-dimensional manner in a sound reproduction space 6 based on the sound signal 3 recorded by the recording unit 10. A sound recording polyhedron 11 is formed by connecting the arrangement points of the 14 microphones 2 in the recording unit 10. Furthermore, a sound reproduction polyhedron 13 is formed by connecting the arrangement points of the 14 speakers 9 in the reproduction unit 12. The sound recording polyhedron 11 and the sound reproduction polyhedron 13 are both roughly isomorphic to a rhombic dodecahedron 28.

In this embodiment, the wavefront incidence direction is determined by a total of seven pairs of microphones 2 or speakers 9, with the vertices facing each other through the center of a rhombic dodecahedron, and the vector sum of these pairs represents the incidence direction of any wavefront. This makes it possible to provide wavefront expression power that adds one more pair to the minimum six pairs that represent wavefronts arriving from two sound sources that can be recognized by binaural hearing.

The rhombic dodecahedron is a dual polyhedron with the centers of gravity of the equilateral triangle and square faces of the cubic octahedron as new vertices, so in terms of rotational symmetry, it has 24 types of rotational symmetry transformations, just like the cubic polyhedron.

Another unique feature of the rhombic dodecahedron is that it can fill three-dimensional space by itself.

Figure 8 shows how a rhombic dodecahedron fills a three-dimensional space.

As shown in this figure, when an acoustic reproduction polyhedron 13 roughly equivalent to a rhombic dodecahedron is assumed in the acoustic reproduction space 6, the acoustic reproduction space 6 can be filled without gaps by connecting 12 acoustic reproduction polyhedrons 113 that have been translated in parallel to the surroundings. This allows smooth connections between polyhedrons when reproducing the sound field when a listener moves physically or virtually from the acoustic reproduction polyhedron 13 to the acoustic reproduction polyhedrons 113 connected to the surroundings, making it possible to reproduce a large-scale acoustic reproduction space 6.

Figure 9 shows the stereophonic system of Example 3.

As shown in this diagram, the stereophonic system is composed of a recording unit 10 that records by arranging 20 microphones 2 in a sound recording space 1 in three dimensions, and a reproduction unit 12 that reproduces the sound recording space 1 by arranging 20 speakers 9 in a sound reproduction space 6 in three dimensions based on the sound signal 3 recorded by the recording unit 10. An sound recording polyhedron 11 is formed by connecting the arrangement points of the 20 microphones 2 in the recording unit 10. Furthermore, an sound reproduction polyhedron 13 is formed by connecting the arrangement points of the 20 speakers 9 in the reproduction unit 12. The sound recording polyhedron 11 and the sound reproduction polyhedron 13 are both roughly isomorphic to a regular dodecahedron 29.

In this embodiment, the direction of incidence of the wavefront is determined by a total of 10 pairs of microphones 2 or speakers 9, with the vertices facing each other through the center of a regular dodecahedron, and the direction of incidence of any wavefront is expressed by the vector sum of these pairs. This makes it possible to provide the expressive power of a wavefront by adding an additional four pairs to the minimum configuration of six pairs that expresses a wavefront coming from two sound sources that can be recognized by binaural hearing.

Regarding rotational symmetry, the stereophonic system of the present invention has 20+24+15+1=60 types of rotational symmetry transformations, including 20 rotations every 120 degrees through the vertices, 24 rotations every 72 degrees through the center of gravity of the regular pentagonal faces, 15 rotations every 180 degrees through the midpoints of the sides, and identity transformation I.

Therefore, compared to Example 1, which has 24 types of rotationally symmetric transformations, this example, which has 60 types of rotationally symmetric transformations, enables the expression of more advanced and complex stereophonic sound through three-dimensional panning using the GBAP method and the calculation and analysis of the covariance matrix related to the rotationally symmetric transformations.

Figure 10 shows the stereophonic system of Example 4.

As shown in this diagram, the stereophonic system is composed of a recording unit 10 that records by arranging 12 microphones 2 in a three-dimensional manner in a sound recording space 1, and a reproduction unit 12 that reproduces the sound recording space 1 by arranging 12 speakers 9 in a three-dimensional manner in a sound reproduction space 6 based on the sound signal 3 recorded by the recording unit 10. A sound recording polyhedron 11 is formed by connecting the arrangement points of the 12 microphones 2 in the recording unit 10. Furthermore, a sound reproduction polyhedron 13 is formed by connecting the arrangement points of the 12 speakers 9 in the reproduction unit 12. The sound recording polyhedron 11 and the sound reproduction polyhedron 13 are roughly isomorphic to a regular icosahedron 30.

In this embodiment, the direction of incidence of the wavefront is determined by a total of six pairs of microphones 2 or speakers 9, each pair being located at a vertex facing each other through the center of a regular icosahedron, and the direction of incidence of any wavefront is expressed by the vector sum of these pairs, providing the expressive power to express wavefronts coming from two sound sources that can be recognized by binaural hearing.

The regular icosahedron is a dual polyhedron with the center of gravity of the regular pentagonal face of the regular dodecahedron as a new vertex, and therefore has 60 types of rotational symmetry transformations, just like the regular icosahedron. Therefore, compared to Example 1, which has 24 types of rotational symmetry transformations, it realizes more diverse 3D panning using the GBAP method, and calculation and analysis of covariance matrices related to rotational symmetry transformations.

Figure 11 shows the stereophonic system of Example 5.

As shown in this diagram, the stereophonic system is composed of a recording unit 10 that records by arranging six microphones 2 and eight dual microphones 32 (a total of 14 microphones) in a three-dimensional manner in the sound recording space 1, and a reproduction unit 12 that reproduces the sound recording space 1 by arranging six speakers 9 and eight dual speakers 35 (a total of 14 speakers) in a three-dimensional manner in the sound reproduction space 6.

A sound recording polyhedron 11, which is a regular octahedron, is formed by connecting the placement points of the six microphones 2 of the recording unit 10. A dual sound recording polyhedron 33, which is a cube, is formed by connecting the placement points of the dual microphones 32 that are placed at the center of gravity 31 of the faces of the sound recording polyhedron 11 (including the point on the normal line that passes through the center of gravity 31 of the face).

The acoustic reproduction polyhedron 13, which is a regular octahedron, is constructed by connecting the arrangement points of the six speakers 9 of the reproduction unit 12. The dual acoustic reproduction polyhedron 36, which is a cube, is constructed by connecting the arrangement points of the dual speakers 35 that are arranged at the center of gravity 34 of the face of the acoustic reproduction polyhedron 13 (including the point on the normal line passing through the center of gravity 34 of that face) that are in contact with each other on their edges. The acoustic recording polyhedron 11 and the acoustic reproduction polyhedron 13 are approximately isomorphic 14. The dual acoustic recording polyhedron 33 and the dual acoustic reproduction polyhedron 36 are approximately isomorphic 14.

In other words, it looks like this:

The sound recording polyhedron 11 formed by connecting the arrangement points of the six microphones 2 and the sound reproduction polyhedron 13 formed by connecting the arrangement points of the six speakers 9 are roughly isomorphic to each other and have a regular octahedron shape. The dual sound recording polyhedron 33 formed by connecting the arrangement points of the eight dual microphones 32 and the dual sound reproduction polyhedron 36 formed by connecting the arrangement points of the eight dual speakers 35 are roughly isomorphic to each other and have a cube shape. The cube is dual to a regular octahedron that is roughly isomorphic to the regular octahedron.

Here, a dual polyhedron, which is a polyhedron (dual sound recording polyhedron 33) that is in a dual relationship with another polyhedron (for example, sound recording polyhedron 11), refers to a polyhedron formed by taking the center of gravity of the faces of the other polyhedron as new vertices, connecting the centers of gravity of the faces that meet at an edge with an edge, and using a polygon that connects the centers of gravity of the faces that meet at a vertex as a face.

The sound recording polyhedron 11 and the dual sound recording polyhedron 33 are in contact with each other along their edges. Also, the sound reproduction polyhedron 13 and the dual sound reproduction polyhedron 36 are in contact with each other along their edges.

The 3D sound system shown in this diagram uses a regular octahedron and its dual polyhedron, a cube, to form the sound recording space 1 and sound reproduction space 6. Therefore, a total of seven pairs of microphones or speakers - three pairs of vertices that face each other through the center of the regular octahedron and four pairs of vertices that face each other through the center of the cube - form the direction of incidence of the wavefront, and the vector sum of these pairs represents the direction of incidence of any wavefront. This makes it possible to provide the expressive power of a wavefront that adds one more pair to the minimum six pairs that represent a wavefront coming from two sound sources that can be recognized by binaural hearing.

In addition, by using a dual polyhedron as a reference, it is possible to record and reproduce the duality behind acoustic phenomena, namely the duality of elasticity accompanying the compression and expansion of the medium in acoustic phenomena, and the duality of inertia accompanying the movement of the medium, in a physically interpretable form.

Figure 12 shows the stereophonic system of Example 6.

As shown in this diagram, the stereophonic system is composed of a recording unit 10 that records by arranging 26 microphones 2 in a three-dimensional space 1, and a reproduction unit 12 that reproduces the sound recording space 1 by arranging 26 speakers 9 in a three-dimensional space 6.

A sound recording polyhedron 11, which is a cuboctahedron, is formed by connecting the placement points of the 12 microphones 2 of the recording unit 10. A dual sound recording polyhedron 33, which is a dual polyhedron of the cuboctahedron, is formed by connecting the placement points of the dual microphones 32 that are placed at the center of gravity 31 of the faces of the sound recording polyhedron 11 (including the point on the normal line that passes through the center of gravity 31 of the face) that are in contact with each other on their edges.

The sound reproduction polyhedron 13 is constructed by connecting the arrangement points of the 14 speakers 9 of the reproduction section 12. The dual sound reproduction polyhedron 36 is constructed by connecting the arrangement points of the dual speakers 35 that are arranged at the center of gravity 34 of the face of the sound reproduction polyhedron 13 (including the point on the normal line that passes through the center of gravity 34 of that face) that are in contact with each other on their sides. The sound recording polyhedron 11 and the sound reproduction polyhedron 13 are approximately isomorphic 37. The dual sound recording polyhedron 33 and the dual sound reproduction polyhedron 36 are approximately isomorphic 37.

The rhombic dodecahedron shown in this figure is a spherical rhombic dodecahedron whose vertices are located on a sphere 38 of the same radius from the acoustic center.

Figure 13 shows a recording device as a specific example of the recording unit 10 in Figure 12.

In FIG. 13, twelve acutely directional microphones 2 are installed in a microphone holder 17, with each microphone 2 positioned at a vertex of a cubic octahedron. The sound receiving parts of the microphones 2 are installed facing outward. The microphone holder 17 is supported by a microphone stand 16.

Furthermore, 14 acutely oriented dual microphones 32 are installed in the microphone holder 17, and each dual microphone 32 is arranged at a vertex of a rhombic dodecahedron. The sound receiving parts of the dual microphones 32 are installed facing outward. The microphone holder 17 is supported by the microphone stand 16.

This allows the wavefront incident on the center of the polyhedron to be recorded in a way that can be physically interpreted in terms of the duality behind acoustic phenomena.

In addition, an omnidirectional microphone 39 may be provided at the acoustic center for purposes such as evaluating the absolute sound pressure of the sound field to be recorded. This allows the absolute sound pressure to be captured by the omnidirectional microphone 39, and the physical relationship between the arrival of the wave front and the absolute sound pressure to be evaluated using the acoustic signal 3 recorded by the surrounding acoustic recording polyhedron 11 and the dual acoustic recording polyhedron 33 shown in FIG. 12. This can be used for noise control, design, and other purposes that require evaluation of the absolute sound pressure of the sound field.

Figure 14 shows a reproduction device as a concrete example of the reproduction unit 12 in Figure 12.

In FIG. 14, twelve speakers 9 are arranged on a speaker stand 40 so as to form a shape that is roughly equivalent to a cuboctahedron, and fourteen dual speakers 35 are arranged so as to form a shape that is roughly equivalent to a rhombic dodecahedron.

As a result, by playing back the audio signal 3 recorded by the recording unit 10 in FIG. 13 as is, or by playing back the signal after performing operations such as the GBAP method, the three-dimensional audio of the audio recording space 1 can be easily reproduced in the audio reproduction space 6 with a small computational load.

Figure 15 shows an acoustic reproduction polyhedron formed in a rectangular parallelepiped acoustic reproduction space.

As shown in this diagram, a normal room (acoustic reproduction space) such as a living room has a rectangular parallelepiped shape, so the recorded acoustic signal can be reproduced by the acoustic reproduction polyhedron 13 and dual acoustic reproduction polyhedron 36, which correspond to the rectangular parallelepiped shape. Even if the shape corresponds to a rectangular parallelepiped shape, it can be said to be roughly isomorphic to a cuboctahedron and a rhombic dodecahedron. A cuboctahedron is a solid body obtained by cutting off the eight vertices of a cube to the midpoints of the sides, and a rhombic dodecahedron is roughly isomorphic to a solid body obtained by connecting the eight vertices and the centers of gravity of the six faces of a rectangular parallelepiped, so it is easy to configure the acoustic reproduction polyhedron 13 in a rectangular parallelepiped room.

Regarding the distortion of the acoustic reproduction space 6 caused by constructing the room in a rectangular parallelepiped shape, one method is to calculate the covariance matrix related to the rotational symmetry as shown in Figure 6 for the acoustic recording space 1 and the acoustic reproduction space 6, respectively, and then correct the acoustic signal so that the two match.

The 3D sound system of this embodiment also has the advantage of being compatible with a physical model of 3D sound based on the dual relationship between elasticity and inertia.

Figure 16 shows a physical model representation of a two-dimensional cross section of an acoustic space filled with rhombic dodecahedrons.

In this diagram, acoustic compliance 42, represented as a spring, is shown stretching in six of a total of 12 directions (including directions not shown because this is a two-dimensional cross section) from elastic center 41. Acoustic inertance 43, represented as a center of inertia, is located between the springs. In other words, the polyhedron connecting the elastic centers is a cubic octahedron 20, and the polyhedron with the center of inertia as a face is a rhombic dodecahedron 152. For this reason, the cubic octahedron 20 and its dual polyhedron, the rhombic dodecahedron 152, correspond to a physical model of stereophonic sound.

This model is a physical model, so acoustic calculations are possible. For simplicity, we will only refer to two-dimensional waves here.

The coordinate system for the acoustic calculation is based on three axes, x ₁ , x ₂ , and x ₃ , rotated by 120 degrees each, as shown in this figure.

First, consider the law of conservation of mass for a region V enclosed by a closed surface S of a space-filling cell (a hexagon in two dimensions, a rhombic dodecahedron in three dimensions). At a given time, the mass contained in V is expressed by the following equation (5).

where δ is the diameter of the acoustic element and ρ is the density of the air.

Therefore, the mass change per unit volume is expressed by the following equation (6):

This mass change occurs when sound waves flow into region V through surface S. Now, consider the mass flowing out through area element dS of surface S. It is _ρvn dS per unit volume. Here, _vn is the component of the outward normal of particle velocity v with respect to surface S, and in the two-dimensional model of FIG. 16, three axial components are considered in total. The total outflow amount is expressed by the following equation (7).

Here, for the surface elements of regular hexagons, the following relation (8) was used.

Therefore, from the above equations (6) and (7), we obtain the continuity equation (9), which is expressed as follows:

Next, consider the equation of motion. As shown in Fig. 16, the motion of the acoustic inertance 43, which is the center of inertia for each axis _x1 , _x2 , and _x3 , is considered to be one-dimensional. Therefore, the equation of motion is given by the following equation (10), just like in the one-dimensional case.

Furthermore, for numerical analysis, the above continuity equation (9) and equation of motion (10) are discretized. The elastic center 41 is set as the sound pressure reference point, and the acoustic inertance 43, which is the center of inertia, is set as the particle velocity reference point, and the variable names are given as shown in Figure 16. The gradient of the particle velocity in the above continuity equation (9) is discretized by the Euler method as shown in the following equation (11).

By substituting the above equation (11) into the above continuity equation (9) and rearranging, we obtain the following discretized continuity equation (12).

The gradient of the sound pressure in the above equation of motion (10) is discretized as shown in the following equation (13).

By substituting the above equation (13) into the above equation of motion (10) and rearranging it, we obtain the following equation (14) for the _x1 axis direction, for example.

Finally, from the above equations (12) and (14), we obtain the following state space equation (15).

The matrix in the first term of the above equation represents the energy storage characteristics in elasticity and inertia, and the matrix in the first term on the right-hand side represents the connection between the elastic and inertial elements.

As an example of numerical analysis, consider a two-dimensional rectangular acoustic space expressed by the following equation (16).

The boundary was assumed to be rigid (particle velocity is zero), and the above formulated equation (15) was combined to obtain the state space equation of the target. The vertical and horizontal directions were divided into 25 parts, and the calculation was performed using a total of 625 elements.

Figure 17 shows the eigenmodes and eigenfrequencies of the acoustic space obtained by performing eigenvalue analysis on the state space equation obtained from the above equation (15).

In this figure, white indicates nodes of sound pressure, and black indicates antinodes of sound pressure. For example, (1,0) in the figure indicates a natural vibration with one node in the horizontal direction and zero nodes in the vertical direction. In addition, Sim in the figure indicates the natural frequency obtained by this model, and Exact indicates the exact solution of the natural frequency in a rectangular room of l _x ×l _y . This exact solution is expressed by the following equation (17).

However, n _x and n _y are all 0, 1, 2, 3....

These results show that the model analysis results and the exact solution match with a maximum error of 2.3%.

The above demonstrates that acoustic analysis can be performed based on the geometric configuration of the 3D audio system of this embodiment.

Figure 18 shows the stereophonic system of Example 7.

In the stereophonic system shown in this figure, the acoustic recording space 1 and the acoustic reproduction space 6 are the same space, and the acoustic recording polyhedron 11 and the acoustic reproduction polyhedron 13 are approximately congruent 44. Therefore, it is a stereophonic system that combines the recording unit 10 and the reproduction unit 12, and by playing back the acoustic signal 3 recorded by the recording unit 10 after performing operations 45 such as rotation, adding reverberation, and filtering such as reducing sound pressure, in the reproduction unit 12, an interactive stereophonic system is constructed, making it possible to provide a creative sound field.

Figure 19 shows a part of a reproduction device as a concrete example of the recording unit 10 and reproduction unit 12 in Figure 18.

Figure 19 shows a speaker enclosure 46, which is part of the reproduction device, installed at the apex 252 of a rectangular parallelepiped room. The speaker enclosure 46 has a microphone 2 provided near the speaker unit 50 so that the sound recording polyhedron 11 and the sound reproduction polyhedron 13 are approximately congruent.

The speaker enclosure 46 is pyramidal in shape and is in contact with wall A (47), wall B (48) and wall C (49).

This makes it possible to provide a 3D sound system with a device configuration that compactly houses the recording unit and reproduction unit.

Regarding the low-frequency characteristics of the speaker, an opening 51 may be provided near the three sides defined by wall A (47), wall B (48), and wall C (49), and a horn type may be adopted that utilizes the space. A passive radiator type low-frequency compensation circuit may also be provided.

Furthermore, by providing a light source in the space and using the opening 51 as a light passage, it can also function as indirect lighting. This makes it possible to create an interactive 3D sound system using sound waves and light.

1: Acoustic recording space, 2: Microphone, 3: Acoustic signal, 4: Ambisonics encoder, 5: Ambisonics audio signal, 6: Acoustic reproduction space, 7: Ambisonics decoder, 8: Reproduced audio signal, 9: Speaker, 10: Recording unit, 11: Acoustic recording polyhedron, 12: Reproduction unit, 13: Acoustic reproduction polyhedron, 14, 15, 28, 29, 30, 37, 52: Approximate type, 16: Microphone stand, 17: Microphone holder, 18: Acoustic element core, 19: Acoustic element, 20: Cubic octahedron, 21, 22, 23: Rotation axis, 24: 1 type of identity transformation, 25: 8 types of rotation, 26: 9 types of rotation, 27: Six types of rotation, 31, 34: Center of gravity, 32: Dual microphone, 33: Dual acoustic recording polyhedron, 35: Dual speaker, 36: Dual acoustic reproduction polyhedron, 38: Sphere of the same radius from the acoustic center, 39: Omnidirectional microphone, 40: Speaker stand, 41: Elastic center, 42: Acoustic compliance, 43: Acoustic inertance, 44: Almost congruent, 45: Operation, 46: Speaker enclosure, 47: Wall A, 48: Wall B, 49: Wall C, 50: Speaker unit, 51: Opening, 53: Opposite relationship through the center, 152: Rhombic dodecahedron, 252: Vertex of the room, 352: Mirror surface.

Claims (2)

A recording unit in which 12 microphones and 14 dual microphones are arranged three-dimensionally in an acoustic recording space;
A reproduction unit in which 12 speakers and 14 dual speakers are three-dimensionally arranged in an acoustic reproduction space,
The reproduction unit reproduces the sound recording space based on the sound signal recorded by the recording unit,
a sound recording polyhedron formed by connecting the arrangement points of the 12 microphones and a sound reproduction polyhedron formed by connecting the arrangement points of the 12 speakers are substantially the same type and have a cuboctahedral shape;
a dual sound recording polyhedron formed by connecting the arrangement points of the 14 dual microphones and a dual sound reproduction polyhedron formed by connecting the arrangement points of the 14 dual speakers are substantially the same type and have a rhombic dodecahedron shape;
A stereophonic system, wherein the rhombic dodecahedron is a dual to a cuboctahedron that is generally isomorphic to the cuboctahedron. the sound recording space and the sound reproduction space are the same space, and the sound recording polyhedron and the sound reproduction polyhedron are approximately congruent;
2. The stereophonic system according to claim 1 , wherein said dual sound recording polyhedron and said dual sound reproduction polyhedron are approximately congruent .

Images