JP2003531555A - Multi-channel surround sound mastering and playback method for preserving 3D spatial harmonics - Google Patents

Abstract

(57) [Summary] From multiple mono or directional sound signals that reproduce sound fields via spatial harmonics that substantially match the spatial harmonics of the original sound field, via multiple separate loudspeakers. A method of recording or transmitting a sound field. In order to preserve the spatial harmonics, the monaural sound source is arranged to use the contributions of all speaker channels during mastering. If the specific configuration of the speaker is different from what is assumed during mastering, the speaker signal should be such that the spatial harmonics of the sound field reproduced by the different speaker configurations match the harmonics of the original sound field. Rematrixed at home, theater or other sound playback locations. Alternative methods include recording or transmitting directional microphone signals, or their spatial harmonic components, and then matrixing these signals at the sound playback location, taking into account specific speaker placements. . These methods are described for both the two-dimensional sound field and the more general three-dimensional case, the latter being based on the use of spherical harmonics.

Description

Detailed Description of the Invention

CROSS REFERENCE TO RELATED APPLICATIONS This application is a continuation-in-part application of Serial No. 08 / 936,636, filed Sep. 24, 1997, which is incorporated herein by this reference.

[0002]

TECHNICAL FIELD OF THE INVENTION

The present invention relates generally to electronic sound transmission, recording and playback methods, and more particularly to improvements in surround sound technology.

[0003]

[Prior art]

Improvements in sound reproduction quality and presence have been steadily made for decades. Stereo (2 channel) recording and playback via spatially separated loudspeakers, compared to previous mono (1 channel) sound playback,
The realism of the playback sound is greatly improved. More recently, audio signals are
It is encoded in two channels to drive four or more loudspeakers arranged around the listener. This surround sound further enhances the realism of the reproduced sound. Multi-channel (more than 3 channels) recordings are used in most movie soundtracks, which can be dramatic in theaters with a sound system that includes loudspeakers placed around the wall to surround the audience. Audio effect is brought about. A standard for multi-channel recording on a small optical CDS (compact disc), which is expected to become very popular for home use, is currently emerging. Recent DVD (Digital Video Disc) standards include P in CD, which may or may not include video.
It defines multiple channels of CM (Pulse Code Modulation) audio.

In theory, the most accurate reproduction of the audio wavefront would be obtained by recording and reproducing acoustic holograms. However, it will be necessary to record tens of thousands, or even millions of separate channels. In order to accurately reproduce the original acoustic wavefront, the two-dimensional array of speakers should be placed around the home or theater at intervals not exceeding half of the highest frequency wavelength to be reproduced, that is, less than 1 cm. Will need to be placed. A separate channel would have to be recorded for each of this very large number of speakers, which would also involve the use of many microphones during the recording process. Therefore, such accurate reproduction of the acoustic wavefront is not entirely practical for audio reproduction systems used in homes, theaters and the like.

When the desired reproduction is three-dimensional and the loudspeakers are no longer coplanar, these complications increase correspondingly, making this kind of reproduction even more impractical. The extension to three dimensions allows special effects in the mastering of movie or music recordings, as well as when not limited to the original sound source plane. For example, even in the case of recording a musician on a flat stage, the resulting ambient sound environment will have a three-dimensional character due to the echoes and variations in the instrument arrangements that can be captured and played. It is more difficult to quantify than sound source localization, but the inclusion of three dimensions increases this sense of "spread" and depth in the sound field, even when the actual sound source is coplanar. .

[0006]

[Problems to be Solved by the Invention]

Therefore, with nearly the same number of loudspeakers currently used in surround sound systems, it is possible to provide improved and immersive sound reproduction techniques with multi-channel recording as specified in the emerging emerging audio standards. It is the main and general purpose of the present invention.

It is another object of the present invention to provide a method and / or system for playing recorded or transmitted multi-channel sound in a home, theater, or other listening location, by which the user can: , The electronic matrix can be set at the listening position for the particular placement of the loudspeakers used there.

It is a further object of the invention to extend these techniques and methods to the capture and reproduction of a three-dimensional sound field when the loudspeakers are placed in a non-coplanar arrangement.

[0009]

[Means for Solving the Problems]

These and additional objects are realized by the present invention, in which the acoustic field is simply and generally ensured by multiple signals through four or more loudspeakers arranged to surround the listening area. And regenerated, the signal is processed to regenerate a specified number of spatial harmonics of the acquired acoustic field with virtually any arrangement of speakers around the listening area. This enhances the ambience of sound reproduction without imposing any particular constraint on the position of the loudspeaker.

It is not necessary that the loudspeakers be placed in a particular pattern before the system can reproduce the specified number of spatial harmonics, rather any existing loudspeaker positions will be present in the playback layout. Used as a parameter in electronic encoding and / or decoding of multi-channel sound signals to produce advantageous results. If one or more speakers are moved, these parameters are changed to preserve the spatial harmonics of the reproduced sound. The use of 5 channels and 5 speakers is described below to illustrate various aspects of the present invention.

According to one particular aspect of the invention, the individual monophonic sounds are mixed together using a matrix, which matrix forms an angle with the monophonic sounds when forming a recording or audio transmission. Presence is improved when these are played back through the intended speaker placement around the listener. Rather than simply sending a given mono sound to the two channels that drive the speaker on each side of the sound's position, it plays the sound at the desired spatial harmonics, as is currently done by standard panning techniques. Because of this, all channels are potentially included. A typical application is in mastering recordings of multiple musicians playing together. First, the sounds of each instrument are recorded separately and then mixed to be placed around the listening area during playback. By using all channels to maintain spatial harmonics, the reproduced sound field is closer to the existing sound field in the room where the musician is playing.

In accordance with another particular aspect of the invention, multi-channel sound is used at home to accommodate speaker arrangements different from those envisioned when initially mastered.
Can be re-matrixed in theaters or replay locations. The desired spatial harmonics are accurately reproduced using different actual placements of speakers. This allows free speaker placement, especially important in the home where speaker placement is often constrained, without losing the improved sound sensation.

According to a further particular aspect of the invention, the sound field is first acquired with directional information using a plurality of directional microphones. The spatial harmonic signals, as a result of the microphone output, or the initial partial matrixing of the microphone output, are transmitted to the listening position by recording or a separate channel. The transmitted signal, together with a certain number of spatial harmonics matched with the spatial harmonics of the recording position, then reproduces the recorded sound field, at home or at another listening location, at the actual speaker position. It is re-matrixed in consideration.

These various aspects can use two-dimensional or three-dimensional spatial harmonics. In the two-dimensional case, whether the original recording was based on two-dimensional spatial harmonics or by three-dimensional projection onto the speaker plane, the sound wavefront is reproduced primarily by the co-planar loudspeaker arrangement. It In 3D playback, one or more speakers are placed at a different height than the 2D surface. Similarly, the three-dimensional sound field is acquired by a non-coplanar arrangement of multidirectional microphones.

Additional objects, features and advantages of various aspects of the invention will be apparent from the preferred embodiments of the invention, which are to be construed in conjunction with the accompanying drawings. Should be.

[0016]

DETAILED DESCRIPTION OF THE INVENTION

We will discuss from the method of spatial harmonics in the two-dimensional plane. Some of the results of this methodology are: That is, (1) a surround sound recording method that can be used to supply an arbitrary number of speakers, (2) a monaural sound panning method for accurately creating a predetermined set of spatial harmonics, and (3) surround sound. A method of storing or transmitting on a channel such that two of the channels are standard stereo mixes and a third channel can be used to reproduce a surround supply that preserves the original spatial harmonics.

Following the discussion of two dimensions, we extend this same theory to three dimensions. In two dimensions, spatial harmonics are based on a single variable, the Fourier sine and cosine series of angle φ. Unfortunately, the 3D version of math is not as clean and compact as it is for 2D. There is no particularly good way to reduce complexity, and thus 2D
Present the version first.

To extend the method of spatial harmonics to three dimensions, we next briefly discuss the Legendre function and spherical harmonics. In a sense, this is a generalization of the Fourier sine and cosine series. The Fourier series is a function of one angle, φ. The series is periodic.
This can be thought of as a representation of a function on a circle. The spherical harmonic is defined on the sphere and is a function of two angles, θ and φ. φ is an azimuth angle defined when 0 ° is straight ahead, 90 ° is left, and 180 ° is straight backward. θ is the declination (upper and lower), 0 ° is straight overhead and 90 °
Is a horizontal plane and 180 ° is straight down. These are shown in FIG. 9 for the point (θ, φ). While the range of θ is 0 to 180 °, φ
Note that the range is from 0 to 360 ° (or -180 ° to 180 °).

Spatial Harmonics in Two Dimension FIG. 1 shows a person 11 in the center of the listening area, which is surrounded by loudspeakers SP1, SP2, SP3, SP4 and SP5. Is aimed at sending those sounds to the center. An angular coordinate system has been established for the purposes of this application. The front direction of the listener 11 facing the front speaker SP1 is (θ ₁ , φ ₁ ) = (90 °, 0 °) as a reference.
Considered to be located in. Remaining speakers SP2 (front left), SP3 (rear left)
, SP4 (rear right) and SP5 (front right) have angular positions (θ ₂ , φ ₂ ), (θ ₃ , φ ₃ ), (θ ₄ , φ ₄ ), and (θ ₅ , φ ₅ ). Here, the loudspeakers are placed in a typical arrangement that defines a surface that is substantially planar, 1
One example is a horizontal plane at θ = 90 °, parallel to the floor of the room where the speaker is placed. In this situation, each of θ ₁ −θ ₅ is 90 ° and these θ are not explicitly represented for the moment and are omitted from FIG. One or more loudspeakers need not be taller than one or more other loudspeakers, but may be made to fit a limited space. One or more cases where θ _i ≠ 90 ° are discussed below.

It is desirable to place the monaural sound 13, such as from a single instrument, in a position where there is no speaker at angle φ ₀ from the zero reference. Usually, there are other mono sounds that are desired to be co-located at other angles, but for simplicity of explanation only source 13 is shown here. For multi-instrument music sources, for example, individual instrument sounds are typically placed at various angles φ ₀ around the listening area during the mastering process. The sound of each instrument is captured by one or more microphones recording mono on at least one separate channel. These mono recordings act as sound sources during the mastering process. Instead, mastering is
Can be run in real time from a separate instrument microphone.

Prior to describing the mastering process, FIG. 2A is provided to illustrate the concept of spatial frequency.
See ~ D. FIG. 2A shows the space surrounding the listening area from the viewpoint of the angular position. The five positions of each of the speakers SP1, SP2, SP3, SP4 and SP5 are shown without changing the desired position of the sound source 13. Sound source 13
Can be viewed as a spatial impulse that can be expressed as a Fourier expansion as Where m is an integer of the individual spatial harmonics, 0-M harmonics are reconstructed and is the coefficient of one component of each individual harmonic, and a _m is one component of each harmonic. And b _m is the coefficient of the orthogonal component of each harmonic. Therefore, the value a ₀ is the zero-order value of the spatial function.

The spatial zero-order is shown in FIG. 2B, and has the same size around the entire space that rises and falls with the size of the spatial impulse sound source 13. FIG. 2C shows the first-order spatial function, which has one complete cycle around the space, whereas it has a maximum at the angle of the impulse 13. As illustrated in FIG. 2D, the quadratic spatial function has two complete cycles around the space. Mathematically, the spatial impulse 13 is accurately represented by a large number of orders, but due to the fact that only a few loudspeakers are used, the number of spatial harmonics that can be included in the reproduced sound field is Has a limit. If the number of speakers is (1 + 2n) or more (n is the number of harmonics to be reproduced here), there are zero to n spatial harmonics of the reproduced sound field in the original sound field. Can be played virtually exactly as the ones. Conversely,
The spatial harmonics that can be accurately reproduced are zero to n harmonics, where n is the largest integer that is less than or equal to half an integer less than the number of speakers placed around the listening area. Alternatively, a number below this maximum number of possible spatial harmonics may be chosen to be reproduced in a particular system.

One particular aspect of the present invention is illustrated in FIG. 3, which schematically illustrates some features of a sound console used to master multi-channel recordings. In this example, five signals S1, S2, S3, S4
And S5, probably in digital form, are recorded on five separate channels of a suitable recording medium such as tape. Each of these signals will drive a separate speaker. It is illustrated that the two monaural sound sources 17 and 19 are mixed into the recorded signals S1-S5. These sound sources 17 and 1
9 can be, for example, live or recorded signals of various musical instruments that are mixed together. One or both of the sound sources 17 and 19 may be synthetically generated or naturally recorded sound effects, voices and the like.
In practice, usually more than two such signals are used to make the recording. The individual signals can be added to the recording tracks one at a time or mixed together for simultaneous recording.

Illustrated in FIG. 3 is a monaural sound “positioning” technique. That is, the apparent position of each of the sound sources 17 and 19 when the recording is played back via the surround sound system is set during the mastering process, as described above with respect to FIG. Nowadays, the normal panning technique of the mastering console sends the mono sound only to two of the recorded signals S1-S5, which causes the sound source listener to the speakers on either side of the desired position for that sound. With relative amplitudes that determine the apparent position relative to. But it lacks some kind of presence. Therefore, as shown in FIG. 3, each sound source is directed to each of the five channels of a signal having a certain number of spatial harmonics, at least zero and first harmonics, of the sound field emanating from that location. Supplied with the relative gains set to build the set. One or more channels may still not be able to receive a portion of a particular signal at all, which is a result of the preservation of a certain number of spatial harmonics, where the signal is artificially only on two channels. Not because it's limited.

The relative contribution of the source 17 signal to the five separate channels S1-S5 is represented by the individual variable gain amplifiers 21, 22, 23, 24 and 25. The gains g ₁ , g ₂ , g ₃ , g ₄ and g ₅ of each of these amplifiers are controlled by the control processor 2.
It is set in the circuit 27 by the control signal from 9. Similarly, the sound signal of the source 19 is fed to the respective amplifiers 31, 32, 33,
Sent through 34 and 35. The respective gains g ₁ ′, g ₂ ′, g ₃ ′, g ₄ ′ and g ₅ ′ of the amplifiers 31-35 are also set by the control processor 29 via the circuit 37. These sets of gains are calculated by the control processor 29 from inputs from the sound engineer via the control board 45. These inputs are source 1
It includes the desired placement angle Φ (FIG. 1) of the sounds from 7 and 19 and the assumed speaker set placement angle φ ₁ -φ ₅ . The calculated parameters can also optionally be provided and recorded through circuit 47. The individual outputs of each amplifier 21-25 are fed to the amplifier 3 by the respective summing nodes 39, 40, 41, 42 and 43.
Combined with the outputs of 1-35, it provides five channel signals S1-S5. These signals S1-S5 are eventually reproduced via one of the respective speakers SP1-SP5.

The control processor 29 includes a DSP (digital signal processor) that operates to solve the system of equations from the input information to calculate a set of relative gains for each of the mono sources. The principle set of linear equations that can be solved for each separately placed source placement can be expressed as: Where φ ₀ represents the angle of the desired apparent position of the sound, φ _i and φ _j represent the angular positions corresponding to the loudspeaker placement for the individual channels, and each of i and j is 1 Has an integer value from 0 to the number of channels, m represents the spatial harmonics, which can range from 0 to the number of harmonics that are matched to the harmonics of the original sound field during playback, N being the number of channels Is the total number, g _i represents the relative gain of the individual channels, i goes from 1 to the number of channels. It is for this set of relative gains that the equation is solved. The use of the subscripts i and j follows the usual mathematical notation for matrices, where i is the number of rows of terms in the matrix and j is the number of columns.

In a particular example where the number of channels N and the number of speakers is also 5, and the zero-order and first-order spatial harmonics are accurately reproduced, the above linear equation can be expressed as the following matrix. The general matrix, solve for the desired set of relative gain g ₁ -g _5.

This is a rank-3 matrix, which means that there are many relative gain values that satisfy it. Another constraint is added to provide a unique gain set. One such constraint is that the second spatial harmonic is zero, which changes the lower two rows of the matrix above.

Another constraint that may be placed on the solution of the general matrix is the velocity vector (about 750-150).
That is, frequencies below the transition frequency in the range of 0 Hz) and power vectors (at frequencies above this transition) must be substantially aligned. As is known, the human ear discriminates sound direction by different mechanisms in the frequency range above and below the transition. Thus, the apparent location of the sound, which potentially extends into both frequency ranges, is made to appear to the ear to come from the same location. This is obtained by equating the equations for the angular directions of each of these vectors as follows: The definition of the velocity vector direction is to the left of the equal sign and the definition of the power vector is to the right.
For the power vector, taking the square of the gain term is an approximation of the model of how the human ear reacts to high frequency ranges and so may vary slightly between individuals.

Once a set of relative gains has been calculated by the control processor 29 for each of the sounds to be placed around the listener 11, the resulting signals S1-S5 are reproduced from the recording 15 and the speaker SP1. -One of SP5 can be driven individually. If the loudspeakers are located exactly at or very close to the angular positions φ ₁ -φ ₅ around the listener 11 that were assumed when calculating each sound source, all sound source positions are It would seem that the engineer was in the place where they tried to place those sources. 0th order, 1 included in these calculations
The next and any higher spatial harmonics will also be faithfully reproduced.

However, physical constraints in the home, theatre, or other places where the recordings are to be played often limit where the speakers of the sound system can be placed. The spatialization of individual sources may not be optimal if they are placed around the listening area at an angle different from what was assumed during recording. Therefore, according to another aspect of the invention, the signals S1-S5 are re-matrixed by the listener's sound system, as illustrated in FIG. The sound channels S1-S5 reproduced from the recording 15 are, in a particular embodiment, initially due to the harmonic matrix 51:
It is converted into spatial harmonic signals a ₀ (zero-order harmonic), a ₁ and b ₁ (first-order harmonic). The first harmonic signals a ₁ and b ₁ are orthogonal to each other.

If higher than the 0th and 1st spatial harmonics are to be preserved, two additional quadrature signals are generated by the matrix 51 for each further harmonic. These harmonic signals then act as inputs to the speaker matrix 53, which the speaker matrix signals S1 ', S2', S3 ', S4'.
And converted to a modified set of S5 ', so that these signals provide an improved sense of the reproduced sound intended when the recording 15 was initially mastered, assuming various speaker positions. Drives a uniquely placed speaker. This is achieved by the relative gains set in matrices 51 and 53 via respective gain control circuits 55 and 57 from control processor 59. The processor 59 includes mastering parameters recorded and reproduced by the soundtrack, first assumed speaker angles φ ₁ , φ ₂ , φ ₃ , φ ₄ , and φ ₅ ,
And these gains are calculated from the corresponding actual speaker angles β ₁ , β ₂ , β ₃ , β ₄ and β ₅ provided by the listener to the control processor via the control board 61.

The algorithm of the harmonic matrix 51 is illustrated using 15 variable gain amplifiers arranged in 5 sets of 3. Three of the amplifiers are connected to receive each of the sound signals S1-S5 reproduced from the recording. Amplifier 6
3, 64 and 65 receive the S1 signal, amplifiers 67, 68 and 69 receive the S2 signal, and so on. The output from one amplifier of each of these five groups is connected to summing node 81, which has an a ₀ output signal, and the output from another amplifier of each of these five groups outputs an a ₁ output signal. Adder node 8
3 and the output from the third amplifier of each group is the third summing node 85.
, And the output of this node is the b1 signal.

The matrix 51 has the following intermediate values from only the audio signals S 1 -S 5 reproduced from the recording 15 and the speaker angles φ ₁ , φ ₂ , φ ₃ , φ ₄ , and φ ₅ assumed during mastering. Calculate the signals a ₀ , a ₁ and b ₁ . Therefore, in the description of this algorithm, shown as matrix 51,
Amplifiers 63, 67, 70, 73 and 76 have unity gain and amplifiers 64, 68
, 71, 74 and 77 have a gain less than that which is the cosine function of the assumed speaker angle, and amplifiers 65, 69, 72, 75 and 78 which are less than the sine function of the assumed speaker angle. Has a small gain.

The matrix 53 takes these signals and creates new signals S1 ′, S2 ′, S3 ′.
, S4 ′ and S5 ′ to drive a speaker having a unique position surrounding the listening area. In the description of the process shown in FIG. 4, five amplifiers 87-91 that receive the signal a ₀ , five amplifiers 92-97 that receive the signal a ₁ and five amplifiers 98-103 that receive the signal b ₁ are described. Included are 15 variable gain amplifiers 87-103 grouped together. The output of the unique one of each of these three groups supplies the input to summing node 105, the output of another of each of these groups supplies the input to summing node 107, and the other amplifiers are shown. As such, their outputs are connected to nodes 109, 111 and 113 in a similar manner.

The relative gains of amplifiers 87-103 are set to satisfy the following set of simultaneous equations that depend on the actual speaker angle β. Where N = 5 in this example, and consequently i and j have the values 1, 2, 3, 4 and 5. The result is that a home, theater or other user can "dial in" a specific angle taken by the position of the loudspeaker,
This can even be changed from time to time in order to maintain the improved spatial performance offered by the mastering technique.

The matrix equation of the above simultaneous equations for the actual speaker position angle β is as follows, where the condition of the second spatial harmonic equal to zero is also imposed. The value of the relative gain of the amplifier 87-103 is the result of solving the above matrix for the output signals S1'-S5 'of the circuit matrix 53 at the actual speaker position angles β ₁ -β ₅ and the coefficients a ₀ , a _1. And b ₁ are selected.

In the above description, the mastering and playback process was treated as including recording, as indicated by block 15 in each of FIGS. 3 and 4. However, these processes may also be used where there is real-time transmission of mastered sound to one or more playback locations via block 15.

The description with respect to FIGS. 3 and 4 is primarily directed to mastering a three-dimensional sound field, or at least contributes to individual monophonic sound sources. Referring to FIG. 5, a technique for mastering recording or sound transmission from a signal that represents a sound field in three dimensions is illustrated. Three microphones 121, 123 and 125
Are identical and are arranged with respect to the sound field and contain audio field information m ₁ , m ₂ and m ₃ containing sound field information enabling reproduction of the sound field in a set of surround sound speakers.
To produce. A signal is created by placing such a microphone in, for example, a symphony hole, from which a sound effect with realistic directivity can be reconstructed.

As shown at 127, these three signals can be delivered immediately by recording or transmission in three channels. The m ₁ , m ₂ and m ₃ signals are then regenerated, processed and regenerated in the home, theater and / or elsewhere. Playback system
It includes a microphone matrix circuit 129 and a speaker matrix circuit 131,
These are operated by the control processor 133 via respective circuits 135 and 137. This ensures that the signal S1-supplied to the loudspeaker is reproduced in order to accurately reproduce the original sound field with a particular unique arrangement of loudspeakers around the listening area.
It is possible to control and process the microphone signal at the listening position so as to optimize S5. The matrix 129 develops zero-order and first-order spatial harmonic signals a ₀ , a ₁ and b ₁ from the microphone signals m ₁ , m ₂ and m ₃ . Speaker matrix 13
1 takes in these signals and generates the individual speaker signals S1-S5 by the same algorithm as described for the matrix 53 of FIG. The control board 139 also allows the listening position user to specify the exact speaker positions for use by the matrix 131, as well as any other parameters required.

The arrangement of FIG. 6 is very similar to that of FIG. 5, except that it differs in the signal recorded or transmitted. Instead of recording or transmitting a microphone signal at 127 (FIG. 5), microphone matrixing 129 is performed at the sound generation location (FIG. 6),
The spatial harmonics a ₀ , a ₁ and b ₁ of the resulting sound field are recorded or transmitted at 127 '. The control processor 141 and the control board 143 are used in the mastering place. The control processor 145 and the control board 147 are used at the listening place. Figure 6
The advantage of this system is that it does not depend on the type and placement of the microphone used by the recorded or transmitted signal, so that this information does not need to be known at the listening location.

An example of the microphone matrix 129 of FIGS. 5 and 6 is shown in FIG. Each of the three microphone signals m ₁ , m ₂ and m ₃ is an input to a bank of three variable gain amplifiers. The signal m ₁ is applied to the amplifiers 151-153, the signal m ₂ is applied to the amplifiers 154-156, and the signal m ₃ is applied to the amplifiers 157-159. One output of each bank of amplifiers is
It is connected to a summing node which results in a null space harmonic signal a ₀ . Also, another one of the amplifier outputs of each bank is connected to summing node 163, resulting in the first spatial harmonic signal a ₁ . Further, the outputs of the third amplifiers in each bank are connected together at summing node 165 to provide the first harmonic signal b ₁ .

The gains of the amplifiers 151-159 are individually set by the control processor 133 or 141 (FIG. 5 or 6) via the circuit 135. With these gains,
The transfer function of the microphone matrix 129 is defined. The required transfer function depends on the type and placement of the microphones 121, 123 and 125 used. Figure 8
1 illustrates one particular microphone arrangement. The microphones can be the same, but need not be. Only one of the microphones can be omnidirectional. As a particular example, each is a pressure gradient microphone with a cardioid pattern. These are arranged in a Y-pattern, their main sensitivity axis being directed outward in the direction of the arrow. The microphones 121 and 125 are arranged at an angle α on the opposite side of the direction axis of the other microphone 123.

In this particular example, the microphone signal can be expressed as: where ν
Is the sound source angle with respect to the direction axis of the microphone 123. m ₁ = 1 + cos (ν-α) m ₂ = 1-cosν (9) m ₃ = 1 + cos (ν + α) The three spatial harmonic outputs of the matrix 129 are as follows for the three microphone signal inputs. Since these are linear equations, the gain of amplifiers 151-159 is a coefficient for each of the m ₁ , m ₂ and m ₃ terms of these equations.

For clarity of explanation, various sound processing algorithms have been described for analog circuits. Although some or all of the described matrices can be implemented in this manner, these can be used in digital sound mastering consoles commercially available when encoding signals for recording or transmission, and in the digital circuitry of playback devices. It is more convenient to implement the algorithm in the listening position. next,
The matrix is digitally formed within the device in response to supplied software or firmware code that implements the algorithms described above.

In both mastering and playback, the matrix is formed by parameters that include expected or actual speaker positions. There are few restrictions on the position of these speakers. Whatever the speaker position, they are considered as parameters in various algorithms. Improved presence is required by others, such as the use of diametrically placed speaker pairs, speakers placed in the corners of floors and ceilings of rectangular rooms, and other specific linear arrangements. It is obtained without the need for the specific speaker position suggested. Rather, the process of the present invention allows the loudspeakers to be first placed at the desired location around the listening area, then their positions are used as parameters in the signal processing to obtain the signals, which signals are then used by those loudspeakers. To play sound through a specified number of spatial harmonics that is substantially the same as the spatial harmonics of the original audio wavefront.

The spatially reproduced harmonics in the above example are the zero and first harmonics, but higher harmonics will also be reproduced if sufficient speakers are used for that purpose. obtain. Moreover, the signal processing is the same for all frequencies reproduced,
High quality systems extend from tens of hertz down to 20,000 Hz and above. Separate signal processing in the two frequency bands is not necessary.

Three-Dimensional Representation What has been discussed so far was to present a method of two-dimensional spatial harmonics by placing the load speaker and the sound source in one plane. This same theory can be extended to three dimensions. In that case, four channels are required to transmit the 0th and 1st terms of the 3D spatial harmonic expansion. It has the same matrixing properties such that two channels carry a standard stereo mix, while the other two channels provide a feed for any number of speakers around the listener. Unfortunately, 3D
The math in the version case is not as clean and compact as in the 2D case. There is no good way to reduce complexity.

In order to extend the method of spatial harmonics to three dimensions, it is necessary to briefly discuss the Legendre function and the spherical harmonic function. In a sense, this is a generalization of the Fourier sine and cosine series. The Fourier series is a function of one angle, φ. This series is periodic and can be used to represent functions on a circle. Just as the Fourier sine and cosine series are the complete set of orthogonal functions on a circle, spherical harmonics are the complete set of orthogonal functions defined at the surface of a sphere. As such, any function on the sphere can be represented by a spherical harmonic in the generalized Fourier series.

The spherical harmonic function is a function of two coordinates on the sphere, that is, the angles θ and φ. These are shown in FIG. 9, where the points on the sphere are represented by the pair (θ, φ), where φ is the azimuth. 0 degrees is straight ahead. 90 ° is on the left. 180
° is straight back. θ is a declination (up and down). 0 degrees is straight up. 90 ° is the horizontal plane and 180 ° is straight down. The range of θ is 0-18
Whereas 0 is 0 °, the range of φ is 0 to 360 ° (or −180 ° to 180 °).
). In the two-dimensional discussion, the angular variable θ is flattened,
It has been equated to 90 °. More generally, both angles are included. For example, the positions of the loudspeakers SP1, SP2, SP3, SP4 and SP5 in FIG. 1 are the angle pairs (θ ₁ , φ ₁ ), (θ ₂ , φ ₂ ), (θ ₃ , φ ₃ ) here.
, (Θ ₄ , φ ₄ ) and (θ ₅ , φ ₅ ), where θ ₁ is between 0 ° and 18
Located somewhere in the 0 ° range. 1 and 8 can be considered as a coplanar arrangement of the shown elements or a projection of a three-dimensional situation onto a particular planar subspace.

The general definition of spherical harmonics begins with the Legendre polynomial, which is defined as From these, the Legendre related functions can be defined, which are defined as follows. In the formula, P ₀ (cos θ) = 1, P ₁ (cos θ) = cos θ, P ₁ ¹ (cos θ) =
-Sin θ, and so on. Both Legendre polynomials and related functions are orthogonal (but not orthonormal). I gave these specific definitions because some authors have defined them somewhat difficult. If one of the alternative definitions is used, the equation below must be modified accordingly.

Although these are polynomials, they can be transformed into periodic functions by the following substitutions. μ≡cos θ (13) From these, the expansion of the function in polar coordinates can be performed as follows. Function _{P n (cosθ), cosmφP n} m (cosθ), and sinmφP _n ^m (
cos θ) is called a spherical harmonic function. This expansion is equivalent to the Fourier series of equation (1), but it is relatively troublesome to actually derive it. One approach is to fix the value of θ, ie 90 °. The remaining terms are grouped into the equivalent of the Fourier sine and cosine series. Coefficient _{(A n, A nm, B} nm) , the generalized coefficients in Equation _{(1) (a 0, a} m, b m) of the n ≠ 0.

For the function just defined on the circle, there are 1 + 2T coefficients for the series containing harmonics of orders 0 to T. For spherical harmonic expansions, if harmonics up to order T are included, the total number of coefficients is (T + 1) ² and the sphere is 2
Square appears because it is a dimensional surface. Therefore, to store the harmonics up to the first order,
Instead of the three terms a ₀ , a ₁ and b ₁ , four terms A ₀ , A ₁ , A ₁₁ and B _{1 1} are now needed.

When applied to sound, this can be viewed as the sound pressure on the surface of a microscopic sphere at a point in space centered at the listener's location. This extension can be used as a guide through the generation of pan matrix and microphone processing for sounds that can come from any direction around the listener.

As in the 2D discussion, the function on the sphere that we are trying to approximate is considered to be a unit impulse in the direction (θ ₀ , φ ₀ ) to the listener, and the additional coordinate θ is now explicit. become. For compactness, μ ₀ is defined as follows. μ ₀ ≡ cos θ ₀ (15) The extension of the unit impulse in that direction can be calculated as follows. For many different position (θ ₀ , φ ₀ ) multipoint or astigmatic sources, this function is replaced by the sum of these points or the integral of the distribution, respectively.

Although the discussion here is made using three-dimensional harmonics generated from spherical coordinates,
Other sets of orthogonal functions in three dimensions could be used as well. The corresponding orthogonal function would instead be used in equation (16) and other equations. For example, if the geometry of the three-dimensional loudspeaker placement in the listening area fits into a unique coordinate system, or the microscopic surface around the point corresponding to the listener is aspheric due to microphone placement or characteristics. If so, one of the spheroidal coordinate systems and the corresponding orthogonal expansion will be used.

Returning to FIG. 1, N speakers are placed around the listener at an angle of (θ ₁ , φ ₁ ), (θ ₂ , φ ₂ ), ..., (θ _N , φ _N ). , N = 5 exemplary values and each θ _i = 90 ° is no longer used. The gain, g _i , for each of the loudspeakers is determined so that the resulting sound field around the point at the center matches the desired sound field (f ₀ (θ, φ) above) as well as possible. These gains are obtained by minimizing the integrated squared difference between the resulting sound field and the desired sound field. The result of this optimization is the following matrix equation, which switches the left and right sides to generalize equation (2). BG = S (17) where G is a column vector of speaker gains. G ^T = [g ₁ ... G _N ] (18) The components of the matrix B can be calculated as follows. In _{addition, S = [b 10 ... b} N0] T (20)

It should be noted that equation (19) is similar to the expansion in equation (16) for a unit impulse in a constant direction, except for the term (-1) ^m . The initial sum is written without a cap, but in reality this is a finite sum. The rank of the matrix B depends on how many expansion terms are stored. If 0th and 1st
If the terms were preserved, the rank of B would be 4. If another term is taken,
The number of floors will be 9. The rank of B also determines the minimum number of speakers needed to match that many expansion terms.

Any number of loudspeakers can be used, but if the number of loudspeakers is not the perfect square number (T + 1) ² corresponding to the Tth harmonic, then the system of equations is pending. There are various ways to solve an indeterminate system. One method is to use the pseudo-inverse of matrix B to solve the system. This is equivalent to choosing the minimum norm solution, which gives a completely acceptable solution. Another method is to augment the system with an equation that nulls some higher harmonics. This involves taking the minimum number of rows of B that store the rank, and then adding a row of the form [Ρ _{n + 1} (μ ₁ ) ... Ρ _{n + 1} (μ _N )] = [0] (21a) or [cos φ ₁ Ρ ^m _{n + 1} (μ ₁ ) ... cos φ _N Ρ ^m _{n + 1} (μ _N )] = [0] (21b) or [sin φ ₁ Ρ ^m _{n + 1} (μ ₁ ) ... sin φ _N Ρ ^m _{n + 1} (μ _N )] = [0] (21c) These equations are the above equations. A generalization of the process used to transform (3) into equation (4). Exactly which of these is taken does not have much effect. Each additional row increases the rank of the matrix until the maximum rank is reached.

Therefore, the inventor has derived the matrix equations needed to create the speaker gain for panning a single (monophonic) source into a multiple speaker that preserves some spatial harmonics in three dimensions. .

FIGS. 3 and 4 illustrate the mastering and reconstruction process in the case of the coplanar example of two mono sources mixed into five signals, which signals are transmitted through the first order. Converted to spatial harmonics and finally matrixed into a modified set of signals. As mentioned there, although the common multi-channel configuration is the 5.1 format of movie and home movie soundtracks, although it is convenient to choose 5 signals to be recorded and 5 modified signals as output. ,
Any of these particular choices could be taken in various ways. Alternative multi-channel recording and playback methods are described, for example, in James A. et al. 20 by Moorer
Co-pending February 17, 2000, US Patent Application No. 09 / 505,556 "CD
Regeneration Enlargement ", which is incorporated herein by this reference.

The arrangements of FIGS. 3 and 4 are applied to include three-dimensional harmonics, the main modification being that instead of the (1 + 2T) signal, if harmonics up to T are preserved, this time Means that the (T + 1) ² signal becomes the output of the harmonic matrix 51. Therefore, to store the harmonics via the first order, four terms (A ₀ , A ₁ , A ₁₁ , B ₁₁ ) are needed instead of three terms (a ₀ , a ₁ , b ₁ ). Become. Further, the control processor 59 in turn, instead of the individual azimuth angles φ _i and β _j , pairs of assumed speaker angles (θ _i , φ _i ) and actual speaker angles (γ _j , β _j). ), The gain must be calculated, and (γ _j , β _j ) is again provided via the control board 61. Finally, one convenient choice for 3D is the six signals S1-S6 in the non-coplanar case.
And using a modified set of six signals S1'-S6 '. In any case, spherical harmonics require at least four non-coplanar speakers, just as in the 2D case at least three non-collinear speakers are needed. This is because at least four non-coplanar points are needed to define a sphere, and three non-collinear points define a circle on the plane.

The reason why 6 loudspeakers is a convenient choice is that it mixes 4 or 5 tracks recorded or transmitted on medium 15 for coplanar placement, leaving the remaining 2
This is because one to one track can be used for a speaker placed away from the plane. This allows the listener to access and use 4 or 5 coplanar tracks without the need for an elevated loudspeaker or a playback device for spherical harmonics, while providing a complete 3D for the listener. The remaining tracks along with the playback capability are still available on the media. This is similar to the situation described above for the 2D case where two channels can be used in conventional stereo reproduction, but additional channels are available for sound field reproduction. In the 6-channel 3D case, two channels are used for the stereo mix, two channels are boosted for four-channel surround sound recording, and the last two channels further enhance playback through the six channels. Available to provide a three-dimensional sound field. The listener will be able to access the required number of channels from the stored medium, eg as described in the co-pending application “CD Playback Extensions” included by reference above.

Returning to FIG. 3, the modification in this example is to include one additional amplifier for each mono source and one added to provide the additional signal S 6 to the medium 15. In that case, the control board 29 also provides additional gain for each of the sources,
All gains are derived not only from the azimuth position of the assumed speaker placement, but also from the declination. Similarly, in FIG. 4, each of the six signals S1-S6 feeds four amplifiers in matrix 51 to produce the four outputs in this example using the 0th and 1st harmonics, thus A ₀ , A ₁ , A ₁₁ , B ₁₁ (or, more generally, four independent linear combinations of these), one out of four summing nodes. The matrix 53 has six amplifiers for each of these four harmonics, producing a set of six modified signals S1'-S6 '. Also in this case, not only the azimuth position of the actual speaker arrangement but also the declination angle is used. More generally, the control board 61 could also provide the control processor 59 with radius information for any speaker not on the same spherical surface as the other speakers.
In that case, the control processor 59 could use this information matrix 53 to produce a corresponding correction signal and introduce delays to compensate for different radii, compensate for wavefront spread, or both.

In this arrangement, the equivalent of equation (6) above is as follows. Having four of the speakers, S1-S4, be a typical coplanar arrangement parallel to the floor of the room, θ ₁ −θ ₄ = 90 °, equation (6 ′) becomes fairly simple. . Further, by having a perfect three-dimensional representation, two-dimensional projection of the listening area onto any other plane can be realized by fixing θs and φs.

Standard directional microphones have pickup patterns that can be expressed as zero and first order spatial spherical harmonics. The equation for the standard sound pressure gradient microphone pattern is: Where Θ and Φ are angles in spherical coordinates of the microphone main axis. That is, these are the directions in which the microphone is "pointing". Equation (22) is a more general form of equation (9). The equations are up to the overall factors of both, C = 1/2, θ = Θ = 90 °, φ = ν, and Φ = for each microphone m ₁ , m ₂ , or m _3.
Corresponds to equation (22) with α, 0, or −α. The constant C is called the "directivity" of the microphone and is determined by the type of microphone. C is for an omnidirectional microphone and zero for an "8" microphone. Depending on the median value, cardioid (1/2), hypercardioid (1/4), supercardioid (3
/ 8), and standard pickup patterns such as subcardioid (3/4) are obtained. With four microphones, the zero and first spatial harmonics of the 3D sound field can be recovered as follows. This equation corresponds to the 2D zero-order and first-order spatial harmonics of equation (10). The spatial harmonic coefficients on the left side of this equation are sometimes referred to as W, Y, Z, and X in commercially available sound field microphones. The expression of the three-dimensional sound field by these four coefficients is
Sometimes referred to as "B-format". (This term is merely to distinguish it from the direct micro-delivery, which is sometimes referred to as "A-format".)

The terms m ₁ , ..., _{M M} refer to M sound pressure gradient microphones at angles (Θ ₁ , Φ ₁ ), ..., (Θ _M , Φ ₁ ). The matrix D can be defined by the reciprocal thereof as follows.

Each row of this matrix is one directional pattern of microphones. Four microphones unambiguously determine all coefficients for the 0th and 1st order terms of the spherical harmonic expansion. The angles of the microphones must be distinguishable (no two microphones pointing in the same direction) and non-coplanar (because it provides information in only one angular dimension instead of two). In these cases, the matrix is in good condition,
There are reciprocals.

Corresponding changes are also required in FIGS. 5, 6 and 7. In FIGS. 5 and 6, the number of microphones is four, which corresponds to m ₁ -m ₄ in equation (23), and further four harmonics (A ₀ , A ₁ , A ₁₁ , B ₁₁ , or 4). Independent linear combination) replaces three terms (a ₀ , a ₁ , b ₁ ). The number of output signals is also adjusted. In the example used above, S6 or S6 'is included. Furthermore, the alignment of each microphone is specified by a pair of parameters, the angles of the principal axes (Θ, Φ), and the signal S1
Each of -S6 or S1'-S6 'has a declination as well as an azimuth. Correspondingly, the microphone matrix of FIG. 7 will have four sets of four amplifiers.

One possible arrangement of the four microphones in equations (23) and (24) is to place m ₁ -m ₃ on the equatorial plane and m ₄ on the north pole of the sphere, as in FIG. Is. This is (Θ ₁ , Φ ₁ ), (Θ ₃ , Φ ₃ ) = (90 °, ± α), (Θ ₂ , Φ ₂ ) = (90
, 180 °), and Θ4 = 0 °. Another option, as shown in FIG. 10, intended to place the microphone with two microphones facing backward, m ₁ 121
Is (90 °, α), m ₂ 123 is (90 ° + δ, 180 °), and m ₃ 125 is (90
°, −α), and m ₄ 126 are (90 ° −δ, 180 °). α = δ = 6
If 0 ° is taken, it will be a regular tetrahedron arrangement.

In some applications, one of the microphones may be placed at various radii for practical reasons, in which case some delay or lead of the corresponding signal should be introduced. Is. For example, if the rear-facing microphone m ₂ in FIG. 8 is moved backwards, there will be about 1 recording per 1 foot to ensure propagation time.
ms will advance.

Equation (23) is valid for any set of four microphones, again assuming that only one of them is omnidirectional. Looking at this equation for two different sets of microphones, the directional pattern of the pickup can be modified by matrixing these four signals. The starting point is
Equations (23) and (24) and their corresponding matrix D for two different sets of microphones. The actual microphone and matrix are the letters m and D, with the re-matrixed "virtual" amount shown by Tilde.
Indicated by.

Given the formulation of equations (23) and (24), these microphone feeds can be transformed into a set of “virtual” microphone feeds as follows:

The matrix D (note: Tilde D) represents the orientation and angle of a “virtual” microphone. The result of this is the sound that would have been recorded if a virtual microphone was present in the recording section instead of the microphone used. This gives
Recordings can be made with "general" sound field microphones and then they can later be matrixed into any set of microphones. For example, only the first two virtual microphones, m ₁ (note: tilde m ₁ ) and m ₂ (note: tilde m ₂ ) could be chosen and used as a stereo pair for standard CD recording. Then m ₃
(Note: Tilde m ₃ ) could be added to the above planar surround sound recording along with m ₄ (Note: Tilde m ₄ ) used for full 3D realization.

Any non-degenerate transform of these four mic feeds can be used to create any other set of mic feeds, or exactly the zero and first spatial harmonics of the original sound field. It can be used to generate speaker feeds for any number (4 or more) of speakers that can be played. In other words, sound field microphone technology
It can be used to adjust the directional characteristics and angle of the microphone after recording is complete. Therefore, by adding a third rearward facing microphone for 2D and a fourth non-coplanar microphone for 3D, the microphone can be modified with a simple matrix operation. Regardless of whether the data is intended to be released in a multi-channel format, the recording of the third rear-facing channel increases the freedom of stereo release, and the recording of the fourth non-coplanar channel provides stereo and More freedom in flat surround sound.

To matrix the microphone feed into multiple speakers, the right side of the matrix equation (17) is reformulated for panning as follows: and The matrix R ₁ is simply the 0th and 1st order spherical harmonics evaluated at the speaker position. Care must be taken to include the term (-1) ^m . Because it is a direct result of the least-squares optimization needed to derive these equations.

Returning to the recording of the sound field, three or four channels of (preferably uncompressed) audio data corresponding to the 2D and 3D sound fields respectively are stored on a disk or other medium, and In a simple way, it is re-matrixed to stereo or surround. By equation (25) (or its 2D reduction), there are countless lossless non-degenerate transformations of the four channels into the other four channels. Therefore,
Instead of storing spatial harmonics, two channels could store the proper stereo mix, a third could store the channel for 2D surround and a fourth channel for 3D surround mix. Ah In addition to audio, the matrix or its inverse may also be stored on the medium. For stereo demonstrations,
The player simply ignores the 3rd and 4th channels of the audio and the other 2
Play as the left and right feeds. For a 2D surround demonstration, the reciprocal of the matrix is used to derive the 0th and 1st order 2D spatial harmonics from the first three channels. From the spatial harmonics, a matrix such as the planar projection of equation (8) or equation (17) is formed and the speaker feed is calculated. For the 3D surround demonstration, the 3D harmonics are derived using all four channels and then form the matrix of equation (17) to calculate the speaker feed.

While various aspects of the invention have been described with respect to preferred embodiments thereof, it will be understood that the invention is entitled to protection within the full scope of the appended claims.

[Brief description of drawings]

[Figure 1]   FIG. 5 is a plan view of an arrangement of multiple loudspeakers surrounding a listening area.

[Fig. 2]   2 illustrates the acoustic spatial frequency of the sound reproduction arrangement of FIG.

FIG. 3 is a block diagram of a matrixing system for locating monaural sound.

FIG. 4 is a block diagram for rematrixing the signal matrixed in FIG. 3 to take into account different speaker positions than were envisioned when initially matrixing the signal.

FIG. 5 is a block diagram showing an alternative arrangement for obtaining and playing sound from a multi-directional microphone.

FIG. 6 is a block diagram illustrating an alternative arrangement for obtaining and playing sound from a multi-directional microphone.

[Figure 7]   7 shows further details of the microphone matrix block of FIGS. 5 and 6.

8 shows an arrangement of three microphones as audio signal sources for the system of FIGS. 5 and 6. FIG.

[Figure 9]   The arrangement of spherical coordinates is illustrated.

[Figure 10]   3 shows an angular arrangement for a three-dimensional array of four microphones.

─────────────────────────────────────────────────── ─── Continued front page    (81) Designated countries EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, I T, LU, MC, NL, PT, SE), OA (BF, BJ , CF, CG, CI, CM, GA, GN, GW, ML, MR, NE, SN, TD, TG), AP (GH, GM, K E, LS, MW, MZ, SD, SL, SZ, TZ, UG , ZW), EA (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM), AE, AG, AL, AM, AT, AU, AZ, BA, BB, BG, BR, BY, BZ, C A, CH, CN, CR, CU, CZ, DE, DK, DM , DZ, EE, ES, FI, GB, GD, GE, GH, GM, HR, HU, ID, IL, IN, IS, JP, K E, KG, KP, KR, KZ, LC, LK, LR, LS , LT, LU, LV, MA, MD, MG, MK, MN, MW, MX, MZ, NO, NZ, PL, PT, RO, R U, SD, SE, SG, SI, SK, SL, TJ, TM , TR, TT, TZ, UA, UG, UZ, VN, YU, ZA, ZW F-term (reference) 5D020 AD04

Claims (30)

[Claims] 1. A method of processing a sound field for reproducing a sound field over a predetermined frequency range via a surround sound system having at least four channels individually supplying one of at least four speakers. Where the steps of: acquiring a multiple signal of a sound field, and sending the acquired sound field signal to individual channels of a plurality of channels with a set of relative gains over the entire frequency domain, the set comprising: Determined by solving a relational expression, which includes (1) selected speaker positions around the listening area, but not constrained by normal geometric, coplanar patterns, and ( 2) Substantially preserving each of the three-dimensional spatial harmonics of the sound field, so that the sound field reproduced from the speaker located at the selected position is: A method of substantially reproducing a plurality of three-dimensional spatial harmonics of an acquired sound field. 2. The method of claim 1, wherein the plurality of substantially conserved three-dimensional spatial harmonics comprises only zero-order and first-order harmonics. 3. A plurality of substantially conserved three-dimensional spatial harmonics are from zero order to n.
The method of claim 1, comprising the following harmonics, n being an integer less than or equal to an integer less than the square root of the number of speakers. 4. The step of obtaining multiple signals of a sound field comprises obtaining multiple mono signals of a sound to be placed at a specific position around a listening area, the relational expression including such specific positions, The method of claim 1, wherein the sound field reproduced from the speaker further comprises a monaural sound at the particular location. 5. The method of claim 1, wherein the step of acquiring multiple signals of a sound field comprises positioning a multi-directional microphone in the sound field. 6. The method of claim 1, wherein the set of relative gains is determined, at least in part, by a relational expression that includes an assumed speaker position around a listening area. 7. The method of claim 1, wherein the set of relative gains is at least partially determined at a location adjacent to the listening area by a relational expression that includes actual speaker positions around the listening area. 8. The method of claim 1, wherein the set of relative gains is further determined by substantially aligning the velocity and power vectors. 9. The method of claim 1, wherein the set of relative gains is further determined by minimizing the second or higher orders of the plurality of three-dimensional spatial harmonics. 10. The method according to claim 1, wherein the surround sound system has exactly 6 channels, each of which feeds a separate one of exactly 6 loudspeakers. 11. The at least one of the exactly six speakers is arranged to be non-coplanar with another of the exactly six speakers.
The method described in. 12. A method of simulating a desired apparent three-dimensional position of a sound in a multi-channel surround sound system, the method comprising: monaurally acquiring the sound whose three-dimensional position is to be simulated; and Driving the monophonic sound to a plurality of individual channels with a set of relative gains, said set providing a relational expression of the desired apparent position declination and azimuth of the sound at and around a point. Is determined by solving for each set of positions corresponding to the expected position of the loudspeaker driven by each of the multi-channel signals, said relational expression being at least the zero and first three-dimensional harmonics of the sound, Is played through the speaker at the expected position as if it were a monaural As command is either actually present at the position of the apparent, formed by substantially preserved, method. 13. The speakers are actually arranged such that at least one of them has an actual position that is different from the expected position, and the relative position is determined by solving a second relational expression that includes the actual position of the speaker. Including the step of calculating a modified set of gains, wherein at least the zeroth and first order three-dimensional harmonics of the sound, as if the monophonic sound actually existed in said apparent position, when the sound was played through the speaker at the expected position. 13. The method of claim 12, wherein the method is stored as if it were done. 14. The set of relative gains is further determined by substantially aligning the velocity and power vectors of the sound field reproduced through the speaker,
The method according to claim 12 or 13. 15. The set of relative gains is further determined by minimizing second and higher order three-dimensional spatial harmonics of the sound field reproduced through the speaker. The method described in any one of. 16. The method according to claim 12, wherein the number of channels is 4 or more. 17. The method according to claim 12, wherein the number of channels is exactly 6.
The method described in any one of. 18. The method of claim 16, wherein at least one of the expected locations of the speaker is non-coplanar with other expected locations of the speaker. 19. A method of reproducing a three-dimensional sound field via four or more speakers arranged around a listening area, comprising: acquiring a plurality of electric signals representing the sound field; Processing the signal to produce a signal of at least the zero and first three-dimensional spatial harmonics of the sound field, and solving the relational expression involving the term of the actual position of the speaker Three-dimensional spatial harmonic processing to determine the relative gains provided individually,
When solving the above relational expression, at least the zero order and the one order of the sound field reproduced through the speaker, which correspond to the zero-dimensional and first-order three-dimensional harmonics of the acquired sound field, respectively.
A method of substantially preserving the following three-dimensional harmonics. 20. The method of claim 19, further comprising recording and reproducing a plurality of electrical signals representative of a sound field. 21. The method of claim 19, further comprising the steps of recording and reproducing sound field harmonic signals. 22. The method according to claim 19, wherein the sound field is reproduced via exactly 6 loudspeakers. 23. At least one of the exactly six speakers is arranged to be non-coplanar with another of the exactly six speakers.
The method described in. 24. An input for receiving at least four audio signals of an original sound field intended to be reproduced by each of the at least four loudspeakers at a particular assumed position surrounding the listening area, and the assumed position. Is a sound reproduction system having outputs for driving at least four speakers at specific actual positions surrounding different listening areas, the information of the speakers at specific actual positions including declination and azimuth angle. An electronically accepting input, and electronically providing another signal from the input signal to the output in response to the input actual speaker position including the declination and azimuth and the assumed speaker position. A matrix, the output unit drives a speaker to reproduce a sound field by a plurality of three-dimensional spatial harmonics, the harmonics being the original sound. Substantially coincides individually to each the same number of 3-dimensional space harmonics in the system. 25. The matrix is a first part for generating individual signals corresponding to a plurality of three-dimensional spatial harmonics from assumed speaker position information and an input signal, and the three-dimensional spatial harmonics signal and an actual speaker. 25. The sound system of claim 24, further comprising a second portion that produces individual signals for the actual speakers from the location information. 26. The sound system of claim 24 or 25, wherein the plurality of matched three-dimensional spatial harmonics comprises zeroth and first harmonics. 27. The sound system of claim 24 or 25, wherein the plurality of matched three-dimensional spatial harmonics includes only zero and first harmonics. 28. A sound system according to claim 24 or 25, wherein the plurality of speakers at the actual speaker position comprises exactly six. 29. The sound system of claim 25, wherein at least one of the actual speaker positions is arranged to be non-coplanar with another of the actual speaker positions. 30. An input section for receiving an audio signal of an original three-dimensional sound field and an output section for driving at least four loudspeakers at specific actual positions surrounding a listening area to reproduce the sound field. A sound system including an electronically implemented input that accepts speaker actual position information and an output that responds to the input actual speaker position information and an input signal. A matrix, the output section drives a speaker to reproduce a sound field by a plurality of three-dimensional spatial harmonics, the harmonic being substantially the same number of three-dimensional spatial harmonics in the original sound field. A system that matches the corresponding ones individually.