JPH06506092A - sound reproduction system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] sound reproduction system Technical field This invention relates to the reproduction and transmission of sound using three or more loudspeakers.

Background technology When playing stereophonic sound using two loudspeakers, the sound exists between the positions of the two loudspeakers. It has long been known that incomplete hallucinations of spurious hallucinatory sound images can be obtained.

For listeners in the ideal three-dimensional seating position within the listening area, the sound appears to be a phantom illusion. The higher frequencies of the image travel farther from the midpoint between the two loudspeakers than the lower frequencies. The sharpness of the image becomes incomplete. For listeners far from the ideal three-dimensional seating position In this case, all hallucinatory sound images are transmitted to the nearest of the two loudspeakers. ideal As the listener in the target seating position rotates his or her head, the hallucinations become less Rotate the position accordingly.

These drawbacks degrade the naturalness of stereophonic hallucinations, cause listening fatigue, and reduce enjoyment. and allow several listeners in the room to all enjoy a good stereophonic illusion. make it difficult These disadvantages are that stereophonic sound is not suitable for television, video or film. program, using audiovisual and Son-Ko-Lumiere expressions and stereophonic sound effects. Visual projections, such as in theatrical performances and live amplified music performances. This is especially important when associated with statues. The apparent direction of the visual image and its relation to the previous The discrepancy in angle between the Experience has shown that it can be uncomfortable.

These disadvantages are overcome by the use of three or more loudspeakers distributed across a three-dimensional sound stage. It has long been known that it can be alleviated or improved by . These loudspeakers feature an improved stereophonic illusion, one for each loudspeaker system. It can carry independent transmission channel signals, or mix processes or is obtained from fewer transmission channel signals using a matrix process. signal can be supplied. The invention uses an improved matrix process and obtaining improved apparent hallucinatory images.

In the prior art, a stereophonic audio transmission signal is supplied to a stereophonic configuration of a loudspeaker. The result obtained by adding another loudspeaker between each adjacent combination of loudspeakers is and then to that other loudspeaker, perhaps with a different predetermined gain, an adjacent combination of loudspeakers. This can sometimes be improved by supplying an average of the transmitted signal supplied to It is known that there is.

For example, signal and R are two loudspeakers on each left and right side of a stereophonic sound system. If communications are used to feed the loudspeakers of +R) can be supplemented with another central loudspeaker. However, k is a predetermined It is the amplitude gain. In the prior art literature, various values of the predetermined gain are often set to =0. 7 and = 2, but there is a theory that improves all aspects of stereophonic hallucinations. There are no imaginary values. Generally, the larger the value of k, the more the listener is within the listening area. Changing the position improves the stability of the central imaginary hallucinatory image, but the hallucinatory three-dimensional sound that is played It is known that the overall apparent width of the Hibiki stage becomes narrower.

This above-mentioned prior art proposal is sometimes known as the "bridged center loudspeaker method". However, it not only improves the hallucinatory stereophonic sound effect, albeit imperfectly, but also improves the original stereophonic sound signal. Depending on the nature of the recording or mixing techniques used to generate the number and R. It gives a varying degree of improvement. Suitable for stereophonic sound reproduction using two or more loudspeakers transmitted signals, for example those obtained from widely spaced microphones. , the signal obtained from multiple spatially coincident directional microphones pointing in different directions. issue, by electrically simulating the stereoacoustic positions of multiple microphone source signals. and signals obtained from various combinations of these above-mentioned techniques. It will be appreciated that a number of different methods have been proposed and used to do this.

Any method of reproducing such a signal through a larger number of loudspeakers It is desired that it work well with all different recording techniques. Bridge type central loudspeaker The recording technique used to generate the original stereophonic sound signal results in It is important to note that the Experience has shown that it can vary greatly depending on the situation.

Another particular drawback of the bridged center loudspeaker method is This means that the previous recording level balance is not saved. provided in the listening room The total reproduction energy for the supplied "1" is proportional to the sum of the squares of the outputs of the loudspeakers in the room, The total energy L2-H2 of the two-loudspeaker stereophonic signal explained in the above example is The total reproduction energy L2+R2+ [+ k(L+R)]2 is not equal to and is not proportional to.

Other prior art proposals originally included amplification of the left and right sides of a two-loudspeaker three-dimensional sound system. The center Add (L+R) to the left and right to the loudspeaker, which is a signal proportional to the average of the two original signals. and the right loudspeaker is proportional to the difference between the original left and right signals, and whose polarity is opposite. Supplied by supplying 2 (L4) to each signal + and 2 (R-L) to + The three-dimensional sound center, which is reproduced as if it were a stereophonic sound source, is emitted only by the central loudspeaker. It provides a very stable reproduction of images and also uses a process known as an acoustic matrix. This provides a reasonable impression of left and right directivity for listeners in ideal three-dimensional seating positions. This allows the sound traveling from different loudspeakers to the two ears to be heard at the left and right positions. They strengthen and cancel each other out in such a way as to produce interphase characteristics.

However, with the Hughes SR3 method, all sounds come from both the left and right loudspeakers equally. Since the directional illusion is played with a new energy, the directional illusion is purely due to the phase relationship between the loudspeaker outputs. It has the disadvantage that it is created by the person in charge. Under these conditions, the acoustic matrix creates the impression of left-right directionality only in narrow listening areas, creating an ideal three-dimensional seating position. Poor reproduction at higher frequencies, especially above about 2kHz Gives the illusion of breadth.

Numerous other methods exist in the prior art for n2xn, linear matrix circuits or n1 a first plurality of loudspeakers using means having inputs and outputs; It is proposed to feed the signal to a second, larger plurality of loudspeakers. There is. Most of this conventional technology covers a 360° angular range around the listener. The so-called 4-channel system or surround sound that attempts to reproduce directional sound Although some of this prior art has been applied to reproduces the directional effect over the front sector in the direction relative to the 180° drop angle. It has also been applied to stereophonic sound systems that aim to

The first plurality of loudspeakers are operated by a second plurality of larger three-dimensional sound loudspeakers. All prior art systems for stereophonic signal reproduction that attempt to reproduce directional stereophonic signals produce directional incomplete illusions. give. The ear and brain produce directional hallucinations from stimuli in ways that are not fully understood. However, many aspects of the perception of directional effects are determined by four physical quantities at the position of the listener's head. can be reasonably well described.

These four quantities are the sound pressure, which is a scalar quantity, and the acoustic velocity, which is a vector quantity depending on the direction. acoustic energy, which is a scalar quantity, and the direction and magnitude of the energy flow in the sound field. The sound intensity is a vector quantity that describes the sound intensity.

The ratio of sound velocity to sound pressure is limited to a limited frequency band below frequencies of approximately 700 Hz. We predict the localization of sound across the range according to the theory of acoustic localization based on interaural phase cues. gives a vector quantity that can be used to calculate. Sound intensity versus acoustic energy The ratio also increases at higher frequencies, typically between 700Hz and 5kHz. , which can be used to localize the sound, but arrive from different loudspeakers. When the sound is transmitted at different distances from the listener and the loudspeaker is at different distances from the listener, the path length differs by many wavelengths. As shown, when the phases are severely uncorrelated, it is difficult to predict local limits at low frequencies. You can also use

The acoustic localization theory based on the ratio of acoustic velocity to sound pressure is called the velocity vector localization theory. However, those based on the ratio of sound intensity to acoustic energy are based on energy vector theory. It is said. In a first approximation, for a listener equidistant from all loudspeakers, approximately 70 At frequencies below 0Hz, sound is localized in the direction of the acoustic velocity vector, and at higher frequencies In wavenumber and for loudspeaker locations at extremely unequal distances from the listener, the sound changes in intensity. localized in the direction of the vector. Please note that the 700Hz frequency is a wide indication and As a practical matter, the two theories of sound localization have some applicability over a wide range of frequencies. It will be appreciated that a range of numbers exists.

A large number of illusory directions have sound velocities that differ significantly from each other even in ideal three-dimensional seating positions. Producing the vector direction of the degree and sound intensity is the traditional two-tone loudspeaker stereophonic method. This is a drawback of many existing stereophonic sound reproduction methods, including: of the acoustic velocity vector Differences in direction and sound intensity are the result of stereophonic sound signals resulting from playback by three or more loudspeakers. The numbers are often quite small.

Three-dimensional sound is transmitted by the first loudspeaker with a three-dimensional sound configuration via a second, larger loudspeaker. Prior art methods of reproducing a first plurality of signals for reproducing sound have one or more drawbacks. These include changes in recording level balance between sounds within a stereophonic recording, changes in sound velocity, The angular difference between the degree vector direction and the sound intensity vector direction, and the stereophonic sound Contains inappropriate stage playback width.

In the prior art, matrix methods combine a first plurality of loudspeaker feed signals with a second, even louder, signal. It is used not only to feed multiple loudspeakers, but also to preserve stereophonic sound effects, for transmission or recording purposes, and intended for reproduction by a second plurality of loudspeakers. A third plurality of transmission channel signals for the purpose of supplying the signal from the loudspeaker supply signal to the third plurality of transmission channels shown in FIG. It can also be used to generate 3 multiple transmission channel signals. first The process of obtaining a third plurality of transmitted signals from a plurality of loudspeaker-supplied signals generally involves encoding and transmitting the second plurality of loudspeaker supply signals from the third plurality of transmission channel signals. The process of obtaining is called decoding.

Such a system of matrix encoding and decoding has four channels in the prior art. Widely used in conjunction with , surround sound, and amplifier systems has been done. Some such systems have many different possibilities for the first plurality. possible values, many different values for the second plurality, and many different values for the third plurality. It is hierarchical in the sense that it considers different values, but it ensures the following desirable properties: I keep it.

(i) the first and second plurality are equal and the third value is less than the first plurality; For example, the second loudspeaker feed signal is the same as the first loudspeaker feed signal (possible overall game (Aside from the changes in the inn).

(i) the second loudspeaker supply signal is not less than the smaller of the first and second plurality; is invariant for a given selection of the first and second plurality relative to the selection of the third plurality; Ru.

(iii) encoding the first plurality of loudspeaker supply signals into the third plurality of transmission channels; and then decoded into a second plurality of loudspeaker feed signals and then a fourth plurality of transmissions. Once encoded into a channel and then decoded into a fifth plurality of loudspeaker feed signals, The result is that the first plurality of loudspeaker feed signals are transferred to the sixth plurality (second, third and fourth plurality). transmission channel equal to the smallest of the numbers) and then the fifth This is the same as decoding into several loudspeaker supply signals.

(1v) the first plurality of loudspeaker supply signals to the third plurality of transmission channels not smaller than the first plurality of loudspeaker supply signals; a second plurality of loudspeaker feed signals that are larger than the first plurality; The result of decoding into The second loudspeaker arrangement preserves or improves the quality of the reproduction.

This kind of hierarchical system of encoding and decoding processes multiple loudspeaker-supplied signals The procedure does not depend on whether the signal was originally generated for any other number of loudspeakers or not. The point is that it does not depend on whether the signal has passed through intermediate stages of encoding and decoding. is operationally desirable. Possible multiple 2 to encompass the range of both stereoacoustic sectors ．． For stereophonic sound using different numbers of loudspeaker feed signals, including 2.4, or 5 It will be recognized that various proposals exist.

In different applications of stereophonic sound, different loudspeaker supply signals are operationally convenient. can be conventional or customary. For example, sound broadcasting and recording or the broadcasting and recording of most sounds made for playback of compact discs. is made in the two-loudspeaker style, but records made in the 1950s are made in the three-loudspeaker style. It was done. Many recordings made for standard television also use the two-loudspeaker style. However, many movie audio tracks have three-dimensional Recorded in vocal or four-loudspeaker mode. High definition television (HD, TV) ), three or four loudspeakers may be used for the front three-dimensional sound stage. and have been proposed for use with different systems of HDTV or) Multiple different selections can be made by different broadcasts using the same IDTV system. is possible.

A hierarchical system for encoding and decoding stereophonic sound is It greatly facilitates the job of signal conversion to be reproduced by the other, and each recording or Utilizes three-dimensional acoustic materials made by other organizations that use multiple organizations with different transmission systems. Allows you to make several of its own options while allowing you to get more. Similarly, the final listener You can also choose which of the loudspeakers you want to use.

The OMX system of surround sound reproduction is a known prior art hierarchical system. However, it is not optimized for front stage stereophonic sound. Effective stereophonic sound The problem of designing hierarchical systems has hitherto remained unsolved. This is sa For round sounds, you can take advantage of the rotational symmetry of the desired sound stage. However, stereophonic sound has a much smaller degree of mathematical symmetry, which makes it difficult to use a hierarchical system. When the problem of finding a system, especially when taking into account the subjective quality of the directional result, In other words, when requirement (iv) listed in the hierarchy system requirements is taken into account. This is because it is much more difficult to solve if

Most stereophonic loudspeaker configurations have at least approximately left/right symmetry. sand In other words, the loudspeaker is placed symmetrically to the right side in the forward direction with respect to the loudspeaker placed on the left side in the forward direction. There is a second loudspeaker placed in the same position. The same goes for the opposite. Actual upper lip exact left/right The loudspeaker supply signal may deviate from the symmetry of the loudspeaker arrangement. It is customary to design by assuming such exact symmetry. normal deviation from symmetry It turns out that in cases of slight deviation, the subjective results remain reasonably satisfactory. ing. "Front", "Front", "Left", and "Right" directions in this document References to are purely a matter of convenience, and the "front" or "forward" direction is in fact empty. Any convenient direction chosen between the Any opposite selected orthogonal to that direction specified as "face" or "front" It can also be a direction.

Disclosure of invention One aspect of the invention seeks to supply a first plurality of loudspeakers forming a stereophonic configuration. supplying the first plurality of signals to a second plurality of loudspeakers forming a second stereophonic sound configuration; A second even larger loudspeaker supply signal suitable for the reproduction hallucination sound stage width The sound station of a hallucinatory imaginary sound image is transmitted to the listener across a wide listening area. provides a means for converting the matrix in a manner that substantially preserves or improves quality. do.

Another aspect includes a first loudspeaker intended to supply a first plurality of loudspeakers forming a stereophonic sound configuration. suitable for supplying a plurality of signals to a second plurality of loudspeakers forming a second stereophonic sound arrangement; A second, larger loudspeaker feed signal with local quality and internal converting maps in a way that substantially preserves or improves the level balance of different sounds. Provide tricks means.

Another aspect includes the first plurality of stereophonic loudspeakers to be reproduced by the first plurality of stereophonic loudspeakers. Encoding one signal into a third plurality of m transmission, storage, and recording channels and the third plurality of m channel signals are connected to the first plurality of m channel signals. The resulting system is not less than when said second plurality is larger than said first plurality. decrypt the system in a manner that ensures that the system achieves the above first or second object of the invention. Transmission, storage, recording, and playback of multiplex loudspeaker stereophonic sound generating 02 signals Provides a matrix system that performs.

Other aspects include transmitting the first signal by each of the third plurality of transmission, storage, or recording channels. Transmission and recording of the first plurality of signals for which three-dimensional sound is to be reproduced by a plurality of loudspeakers. to create a stereophonic configuration for recording or storage and encompassing a fan of the playback range. decoding a second plurality of signals to be played by a second plurality of loudspeakers; provides a hierarchical system in the sense described above.

Another aspect is that as the listener moves around the listening area, the center of the stereophonic sound stage The stability of hallucinatory imaginary sound images in the vicinity can be improved by providing a wide playback space for listeners across the listening range. While maintaining the width of the stage, two loudspeakers were added to three or more to improve the sound. To provide a means for reproducing a stereophonic sound signal to be reproduced by a device.

Another aspect is the apparent visual image and stereoscopic images relative to the visual images to ensure improved alignment of image directions and audible imaginary hallucinatory sounds Provides a means to reproduce sound.

Other aspects include the listener in an ideal three-dimensional seating position and the between or within the originally intended loudspeaker direction for both listeners who are far away from your location. To provide a high-quality directional sound image in the direction of a sound source that is applied to something in the way.

According to one aspect of the invention, stereophonic sound is produced for reproduction over n5 loudspeakers. the first audio signal encoded as When n2>n, the stereophonic sound is reproduced over the loudspeaker of n211. matrix converter Rn2. n, and ma The trix transformer R converts the total playback energy and directivity effects of the encoded audio signal into , is constructed to be substantially conserved even within the entire proportionality constant, which may be frequency dependent. A matrix converter is provided, characterized in that it is an energy conserving matrix. be done.

The matrix converter can be, for example, a transmission encoder, as explained later, or It may form part of a playback decoder. This may be a suitable form known to those skilled in the art. by software in digital signal processing equipment or by hardware in the analog domain. It can be implemented by a hardwired network.

According to the first aspect of the present invention, the first three-dimensional sound structure is provided over the range of the first fan-shaped area. a loudspeaker supply that is intended to be reproduced by a first plurality of loudspeakers installed in a configuration; over a second sector in response to said first plurality of signals representative of feed signals; A second, larger loudspeaker installed in a second stereophonic configuration generating said second plurality of output signals representative of loudspeaker supplied signals to be reproduced; Matrix regeneration decoding means are provided, the matrix means being frequency dependent. The first standoff caused by the second three-dimensional sound configuration is The body-acoustic configuration conserves virtually all of the intended playback energy. and the matrix means further includes the first stereophonic sound configuration according to the second three-dimensional sound configuration. To substantially preserve or improve the hallucinatory stereophonic sound effects intended by body acoustic configurations. It has become so.

According to the second aspect of the present invention, the first three-dimensional acoustic structure is provided over the range of the first fan-shaped area. the loudspeaker to be reproduced by the first plurality of loudspeakers installed in a configuration; across a second sector in response to the first plurality of signals representing device-supplied signals. a second, larger plurality of loudspeakers installed in a second three-dimensional sound configuration; matrix reproduction for generating said second plurality of output signals to be reproduced by decoding means are provided, said matrix means may be frequency dependent The first stereophonic sound due to the second stereophonic sound configuration extends into the entire proportionality constant. The configuration is designed to conserve virtually all of the regenerated energy that is intended. and said matrix means further comprises a second proportion which may be frequency dependent. Inside the constant, the direction angle from a predetermined imaginary forward direction at a predetermined preferred listening position and reproduced by the matrix means according to the second three-dimensional sound configuration, measured by the angular position of the velocity vector intended by said first stereophonic configuration when , and said matrix means is further adapted to substantially conserve frequency-dependent in the second stereophonic sound configuration to the interior of the third proportionality constant that may exist. When reproduced by the matrix means according to the first three-dimensional sound configuration, The angular position of the intended sound intensity vector is substantially preserved.

In a preferred embodiment of the invention, when said first plurality is equal to 2, said third proportionality constant is The number depends on the frequency.

and said matrix regeneration decoding means preferably includes all left inputs and outputs. If a force attempts to exchange with its right counterpart, the matrix regeneration decoding means It is also symmetrical in the sense that the result given remains virtually unchanged. .

In another preferred embodiment of the invention, the audio frequency section has frequencies over several octaves. The angular arrangement l of the reproduction speed peptization in wave numbers is such that the matrix means The angular distribution of the sound intensity vector when generating the signal to be reproduced by the body acoustic configuration It is configured to be substantially the same as the

In a preferred embodiment of the invention, said third proportionality constant is for audio frequencies higher than 5kHz. Within the frequency band, within the audio band at frequencies between 700) 1z and 3kHz configured to be larger. the third large proportional constant above 5kHz; The number is particularly desirable when said first plurality is equal to two.

In another preferred embodiment of the invention, when said first plurality is equal to two, said first the book representing the difference in said first loudspeaker supply signal intended for body acoustic configuration; In a preferred embodiment of the invention, the ratio of said second proportionality constant to said third proportionality constant is It should be in the range of 10 to 2.

According to other aspects, a first component having components W, X, and Y or a linear combination thereof To reproduce ambient acoustically encoded audio signals by 02 loudspeakers. A transformation matrix is provided for converting the signal into a second stereoacoustically encoded signal. Ru. However, n2 is an integer greater than or equal to 3, and the transformation matrix is The first signal Mdec consisting of the sum of the tropic component W and the first velocity component All said stations configured to receive a signal S consisting of another velocity component Y at a force n2×2 transformation matrix means by plane and the difference Tde of said components W and X It consists of means for outputting still other signal components obtained from c.

In this situation, the sum and difference components are clearly present, and the matrix is also pseudo-lateralized. 0, including both cases where the same operation is performed on the numbers.

Thus, as elsewhere in this specification, the referenced matrix is partitioned A series of functionally equivalent matrices can be formed, or one equivalent can be incorporated into a matrix, and all such configurations are within the scope of the present invention. One thing will be understood.

According to the third aspect of the present invention, the second three-dimensional sound structure is provided over the range of the second fan-shaped area. A second loudspeaker to be played by a second plurality of loudspeakers arranged in a configuration. generating a second plurality of signals not less than a third plurality representing loudspeaker supply signals of a transmission matrix in response to a third plurality of transmission channel signals greater than two; means are provided. wherein the transmission channel signal spans a first sector; to a first plurality of loudspeakers arranged in a first three-dimensional sound configuration. represents a first stereophonic loudspeaker supply signal, wherein the first plurality is the first stereophonic loudspeaker supply signal 2, said transmission matrix decoding means The feed signal is within the overall gain and equalization of the first stereophonic loudspeaker feed. and the first plurality is substantially identical to the second plurality of signals. is smaller and not larger than the third plurality, the transmission matrix decoding The means correspond to the first loudspeaker supply signal contemplated by the first or second aspect of the invention. This constitutes reproduction matrix decoding means.

According to that third aspect of the invention, said transmission channel signal is transmitted to said third plurality of channels. For each first plurality not greater than, precisely said first plurality of transmission channels. The signal is such that it can be substantially zero, and the second one is greater than the first plurality. for a first plurality of said transmission matrix channels; the transmission matrix decoding wherein the channel signal is substantially zero for the first plurality; A transmission channel input to the means for which the transmission channel signal is is a subset of the transmission channel inputs that are substantially zero for the first plurality. .

According to the fourth aspect of the present invention, a three-dimensional acoustic configuration is arranged over the fan-shaped area. 2 representing the loudspeaker supply signal to be supplied to the plurality of loudspeakers installed. used with means of signal transmission, recording, or storage in response to multiple signals that are large A transmission matrix decoding method suitable for generating multiple transmission channel signals a stage is provided, whereby the inverse of the transmission matrix decoding means of the present invention is provided. This constitutes transmission matrix decoding means according to the third aspect.

In a form of its fourth aspect of the invention, a reverse transmission mat according to its third aspect of the invention The Rix decoding means is according to a preferred embodiment of the third aspect of the invention and is as described above. represents a loudspeaker supply signal to be reproduced by the first plurality of loudspeakers of the leading one; a transmission channel signal less than said first plurality and greater than two, which is substantially zero; Further transmission matrix decoding means necessary to generate the signal are also included in the fourth aspect of the invention. It is also a transmission matrix decoding means.

In this preferred embodiment according to the fourth aspect of the invention, different first multiples are encoded by the encoding means. a third plurality of different transmission channel signals generated in response to a number of loudspeaker supply signals; , and an associated third plurality of different transmission channel signals obtained by the inverse decoder. a second plurality of decoded loudspeaker supply signals obtained from a coded signal in the sense defined above; and a hierarchical system of decoding is reliably constructed.

According to the fifth aspect of the present invention, the first three-dimensional sound structure extends over the range of the first fan-shaped area. a loudspeaker supply that is intended to be reproduced by a first plurality of loudspeakers installed in a configuration; encoding said first plurality of signals representative of feed signals into a third plurality of transmission channel signals; and transmitting the third plurality of transmission channel signals over a second sector. reproduced by a second plurality of loudspeakers installed in a second three-dimensional sound configuration. the second plurality of output signals representative of the loudspeaker supply signal to be transmitted; said transmission matrix encoding means used in conjunction with matrix decoding means; constituting the reproduction matrix decoding means according to the first or second aspect of the present invention; A matrix system is provided.

According to a sixth aspect of the invention, in response to the third plurality of transmission channel signals, For use with a transmission channel signal generated by a transmission matrix encoding means are installed in a second three-dimensional sound configuration over a second fan-shaped area. said second plurality of loudspeakers representing loudspeaker-supplied signals to be reproduced by a second plurality of loudspeakers; A plurality of output signals are generated and the resulting system is a matrix according to the fifth aspect of the invention. transmission matrix decoding techniques for constructing a box encoding and decoding system steps are provided.

According to the seventh aspect of the present invention, the first three-dimensional sound structure is provided over the range of the first fan-shaped area. a loudspeaker supply that is intended to be reproduced by a first plurality of loudspeakers installed in a configuration; a transmission matrix in response to one or more of said first plurality of signals representing feed signals; a first number greater than two and not less than said first plurality for use with decoding means; 3, and the resulting system is the fifth station of the present invention. Transmission matrices to form a matrix encoding and decoding system lix encoding means are provided.

According to an eighth aspect of the invention, some of the loudspeakers have a more limited bass range than other loudspeakers. The present invention is intended for use with a loudspeaker (or loudspeaker system) having live capability. Matrix decoding means according to the first, second, third or sixth aspect of , whereby the matrix decoding means has a more limited low frequency at low frequencies. said loudspeaker or loudspeaker system having sound reproduction capability; to provide good bass.

According to the ninth aspect of the invention, all of the listening positions are not necessarily all at the same distance from the preferred listening position. Delay compensation means for output signals intended to be fed to uninstalled reproduction loudspeakers The first, second, third, and sixth aspects of the present invention incorporated into or used in connection with , or matrix decoding means according to the eighth aspect, whereby said delay A delay compensation means ensures that signals from all loudspeakers arrive at said listening position at substantially the same time. reach.

In a preferred embodiment according to the eighth aspect of the invention, the intended three-dimensional sound configuration of the reproduction loudspeaker are substantially bilaterally symmetrical, and the preferred listening position is located on the axis of bilateral symmetry. .

According to the tenth aspect of the present invention, a three-dimensional acoustic configuration is formed over the fan-shaped area. a loudspeaker supply signal to be reproduced by the first plurality of installed loudspeakers; a first plurality of signals representing a larger third plurality of transmission channel signals. Transmission encoding means are provided for encoding the range of the fourth sector. larger than the third plurality, the third plurality being installed in a three-dimensional acoustic configuration over the surrounding area. a fourth plurality of loudspeakers that are not large and are larger than the first plurality; the first plurality of signals providing the fourth plurality of signals representative of loudspeaker supply signals; a regeneration matrix according to the first, second, third or sixth aspect of the invention in response to The encoding means is followed by the encoding means. the third plurality of transmission channel signals in response to the third plurality of transmission channel signals; Encoding means according to the fourth or seventh aspect of the invention are provided.

According to an eleventh aspect of the present invention, the first left-right symmetry is provided over the range of the first fan-shaped area. to be reproduced by a first plurality of loudspeakers installed in a three-dimensional sound configuration. across a second sector in response to said first plurality of signals proportional to said first plurality of signals. and a second larger complex installed in a second symmetrical stereophonic configuration. generating said second plurality of signals proportional to the signal to be reproduced by a number of loudspeakers; Reproduction matrix decoding means is provided, the matrix decoding means comprising: for each pair of signals for a pair of symmetrically installed loudspeakers of the first configuration. input sum and difference matrix means, all of said sum signals and said first configuration; in response to any of said first plurality of signals proportional to a central loudspeaker supply signal to a first number not less than the number of signals of the first output signal. first linear or matrix means for supplying all of said difference signals; means for generating a second number of output difference signals not less than the number of said difference signals in response to said difference signal; a second number in which the first number and the second number are added up to the second plurality; linear or matrix means, and one for each symmetrical amplification of said second configuration. one of the first output signals and one of the output difference signals; from the second plurality of signals for the associated loudspeaker pair of the second configuration in response; It consists of output sum and output difference matrix means for generating a signal, thereby of said second plurality of signals proportional to a central loudspeaker supply signal for said second configuration; which is obtained from the output of one of said first linear or matrix means.

Other aspects, embodiments, objects, and advantages of the invention will be apparent from the description.

Embodiments of the invention will now be explained by way of example with reference to the accompanying drawings, in which: FIG.

Brief description of the drawing Figures 18-1g show examples of loudspeaker arrangements that may be used with the present invention.

2 and 3 show simplified block diagrams of matrix regeneration decoding means according to the invention. vinegar.

FIG. 4 shows a regenerative decoder that generates three output signals from two input signals.

FIG. 5 shows a frequency-dependent version of the decoder of FIG.

FIGS. 6 and 7 show how the transmission signal is encoded between the 2 loudspeaker playback signal and the 3 loudspeaker playback signal. 1 shows a block schematic diagram of a system for encoding and decoding.

Figure 8 encodes two signals into three transmission channels and converts the two signals into three loudspeaker signals. Figure 3 shows frequency dependent means of decoding.

FIG. 9 shows a matrix consisting of two other matrix regeneration decoding means connected in series. 6 shows the Rix playback decoding means.

Figure 1O shows a stereoscopic signal for multiple loudspeakers consisting of a larger number of loudspeakers. FIG. 2 is a schematic diagram showing how to mix and decode for stereoscopic reproduction.

Figure 11 shows a system for encoding and decoding stereoscopic signals with transmission channel signals. It is a schematic diagram.

Figure 12 shows a transmission encoder consisting of a regenerative decoder and another transmission decoder connected in series. vinegar.

FIG. 13 shows a transmission decoder consisting of a series connection of another transmission decoder and a regenerative decoder.

Figure 14 shows a hierarchy of transmission encoders that accept signals for different stereophonic loudspeakers. FIG.

FIG. 15 shows the method of FIG. 14 for decoding the transmitted signal into signals for a plurality of three-dimensional loudspeakers. This is a schematic diagram of the opposite hierarchy.

FIG. 16 shows a transmission encoder and decoder according to FIGS. 14 and 15 and the present invention. 2 is a flowchart showing a procedure for designing a hierarchical system.

Figure 17 shows a 4x2 matrix regenerative decoder according to the invention.

FIG. 18 shows a schematic diagram of a 4×3 matrix regeneration decoding method according to the present invention.

FIG. 19 shows a schematic diagram of an R2 x n matrix regenerative decoder according to the present invention.

FIG. 20 shows the rectangular and polar coordinates of the loudspeaker relative to the listener.

Figures 21 to 25 are stereoscopic images for reproduction by two loudspeakers. and 2 channels by 3×2 matrix regenerative decoder (FIGS. 23 to 25). 2 shows a graph of parameters describing the localization quality of channel stereo.

Figure 26 shows the use of delay compensation means to compensate for different loudspeaker distances.

FIG. 27 shows a multi-speaker stereophonic portable reproduction device according to the present invention.

Figures 28 and 29 show an audiovisual multi-speaker stereophonic sound system used with the present invention. shows.

FIG. 30 shows a multi-channel system using an amplifier control unit incorporating a matrix decoder. FIG. 1 is a schematic diagram of a peaker stereophonic sound system.

Figure 31 shows a multi-speed configuration in which the amplifier control unit supplies the matrix decoder. 1 is a schematic diagram of a three-dimensional sound system.

Figure 32 illustrates the use of the present invention in a multi-speaker stereophonic public address system. show.

FIG. 33 shows an arrangement of loudspeakers in a vehicle for use with the present invention.

Figure 34a is a three-speaker signal generator for B-style signals.

Figure 34b is an n-speaker decoder for B-style signals.

Figure 34c is the rotation matrix used in the decoder of Figures 34a and 34b. .

Figure 35 shows the encoding of the directional acoustic coding system to the transmission signal and the decoding from the transmission signal. Indicates the number.

FIG. 36 shows the relationship between the transformation matrix and the transmission coding matrix.

Figure 37 shows a hierarchical structure that can be connected in series for stereo and surround sound. shows.

BEST MODE FOR CARRYING OUT THE INVENTION 3D sound loudspeaker configuration A sector (3) in the direction in front of the listener (4) suitable for use in connection with the present invention. ) A typical stereophonic configuration using symmetrical loudspeakers is shown in Figures 1a to 1. Shown in g.

FIG. 1a shows an example of a 4-channel system that can be used for single-tone reproduction of a stereophonic signal. ) shows a typical single tone loudspeaker C2 in front of. Figure 1b shows the left and right enlargements, respectively. A typical two-speaker configuration consisting of voice organs L2 and R2 is shown. Figure 10 shows each Typical 3 speakers with left, center, and right loudspeaker 9, C5 and R3 Show the configuration. Figure 16 shows each loudspeaker L4 in front of the listener (4) from left to right; Figure 18 shows a typical four-speaker configuration with speakers 5, R6 and R4. (4) forward left or four right loudspeakers L6, 7th, C2, R7 and R6 respectively; A typical 5-speaker configuration with .

In all these configurations, for various numerical subscripts P, the symbol Cp is set to the ideal position. used to indicate the (imaginary) central loudspeaker in the forward direction to the listener (4) at the Then, measure p counterclockwise to the (imaginary) left of hand (5) well in front of hand (4). is used to indicate a loudspeaker installed in the direction of angle θp, with Rp well in front (5 ) is used to indicate a loudspeaker placed symmetrically to the right in the direction of angle θp to the (imaginary) right of use In the diagrams from Figure 1b to Figure 18, all loudspeakers are It is placed equidistant from position (4) and faces towards the listener's position (4).

However, other arrangements are possible, and by way of example, Figure 1f shows three loudspeakers located at the listener ( 4), but the two outer loudspeakers are located equidistant from each left, center, and tilted to cross in front of the listener (4) An alternative preferred three-speaker configuration is shown with right loudspeakers R5, C5, and R1. are doing. Figure 1g shows another alternative three-speaker configuration, with the outer loudspeaker and and R5 is tilted as before, but connect the central loudspeaker C5 and R3. It is in the center of the line and therefore closer to the listener (4).

The angle θp between the loudspeakers is determined according to the convenience of three-dimensional sound expression or the required stage width. can be selected over a wide range of values. However, the position of the arrogant hand (4) If the angle between adjacent loudspeakers is too large, the quality of the imaginary hallucination image will be poor. It is generally known that The angular width and the overall image quality are satisfactory. As shown, there are no sharp contours between the angular width and the angular width giving unsatisfactory image quality. , for two-speaker stereo, the playback sector (3) has a range greater than 35° (70°). It has been found that a larger C2 (giving a total angular width of ) gives poor image quality. There is. For three-speaker reproduction, θ is preferably 45° (90° reproduction fan area (3) angle Even wider, encompassing a sector (3) of 120° or more, but not larger than The wide stage width can be used with 4 or more loudspeakers to achieve satisfactory results. . In general, the sector of playback range (3) using four or more loudspeakers exceeds 180°. but if it does, a slightly larger angle, e.g. 210° or 225° can be used. However, the inclusion angle slightly more than 180° In such a three-dimensional sound configuration, the outermost loudspeaker is inserted after the loudspeaker (4). The angle is so large that it is impossible to form a stable image after the listener. Book The invention provides three-dimensional sound that does not include the flat area (3) in the range that does not include the stable image at the excluded angular position. Applicable only to configurations and encompasses a 360° surround sound stage It is not applicable to loudspeaker configurations that can be used.

Although the present invention is not limited to the specific values of the angle θp shown in FIGS. 1b to 1g, , the values below are useful exemplary reference values that can be used in practical stereophonic configurations. be.

C2=35°, θ,=45°, C4=50°, θ,=+, θ,=16+', C6 =54°, and θ7; +θ6=27°. More generally, a given ideal listener The angle between adjacent loudspeakers at position (4) is as in the case of the exemplary reference value shown above. , it is convenient to choose a loudspeaker arrangement that is the same as that of all other adjacent pairs. Very often.

Matrix decoding means Using practical techniques and equipment available to the scoring/balancing technician, Certain stereophonic loudspeaker configurations, such as those shown in Figures 1b to 1f, The desired directional hallucination can be generated over the range of the available fan area (3). It can produce, record, store, or transmit stereophonic audio signals.

It is an object of the present invention to have a larger number of loudspeakers such as those shown in Figures 1c to 1e. This configuration substantially maintains or improves this required stereophonic sound effect. be.

A general method of doing this according to the invention is shown in FIG. 2, according to which, for example, Stereophonic microphone configuration, output from mixing desk, tape or disc playback A stereophonic signal source (which can be the output from a machine, a broadcast receiver, or a telecommunications link) 1) of the original first plurality (20) nt signals from the first stereophonic sound structure; The signal representing a loudspeaker supply signal suitable for a second stereophonic audio signal representing a loudspeaker supply signal suitable for the second stereophonic sound configuration (50); This second plurality of signals is shown in FIG. Phase 1 Nox means (2) is shown to be supplied directly from power 1 to the loudspeaker (50). However, such a supply to a general loudspeaker requires amplification and and gain adjustment stage, overall volume and tone control, loudspeaker and room characteristics an equalizer for, a time delay that adjusts the time each loudspeaker power reaches the listener; any necessary or desired connections, such as cables or infrared links. It is understood that a new intermediate stage is included.

The R2X nl reproduction matrix decoding means (2) converts each of the 02 output signals into 01 input signals (20). n2xn of these linear mixtures One coefficient is called a "mad 1 nox coefficient". These linear mixtures are independent of frequency. linear mixing force (or alternatively frequency dependent). – If it is frequency dependent, the Mad 1 Nox coefficient is a complex gain that is a function of frequency. It is. In a preferred embodiment of the present invention when the Mad1 Knox coefficient hζ is frequency dependent, , matrix coefficient 1, two or three relatively wide audio frequency bands is a real number greater than or equal to 1 and independent of frequency, and the transition frequency between these frequency bands is It changes significantly only in the wavenumber domain.

Instead of describing the input (20) or output (40) signal directly with the loudspeaker supply signal, The signals Lp and R1) to be supplied to the loudspeakers, denoted by the same symbol, are often rM sJ It's time again to write in what is called the "sum and difference" form 7! Convenient when you are bored And it's extremely useful.

rMsJ or "sum and difference" signals Ml) and sp each have an amplitude gain of 2 -1/2 = Q, sum of Lp and R1) with 7071 Mp=2-I/2(Lp+Rp) and difference 5p=2-+72 (Lp-Rp) It is defined as. The amplitude gain of 0.7071 was chosen as a matter of convenience. It is something that Other gains come at the cost of complicating the detailed description of the invention. can be selected. Perform this above mentioned MS or sum and difference process The matrix means will be referred to as "rMS matrix" means.

Signals Hp and Sp in MS form apply a second identical MS matrix means. By doing so, the equation %formula%) can be reconverted directly or into a "left/right" format. Sum and difference techniques has been used in stereo technology since British patent 394.325 in 1931. , for example, MS stereo microphone technology and FM stereo multiplex broadcast Ze It is widely known in connection with the Nith/GE system.

In the MS form, we combine the signals of form) 1p or cp with the ``sum'' signal, the signal of form sp. shall be considered a "difference" signal. A two-speaker stereophonic signal in MS format, and R2 are represented as M2 and S2, and the three speaker signals in MS form are 3, C5, and and R5 are represented as Hl, C5, and S, and the four speaker signals in MS form are 4, 5th, R5, and R4 are represented by M4, H6, S4, and S5, and MS type 5 speaker signals L6, 7th, C5, R7, and R6 are connected to M6, M7, C5, It is often convenient to denote by S6 and S7.

The reproduction matrix decoding means (2) is described as being performed in MS format. Sometimes it's convenient. By using the MS matrix equation described above, this Such a description describes the operation of matrix means (2) with respect to signals of left/right form. easily converted into The present invention is applicable to either left/right or MS configurations. accepts signals (20) represented in both, and either in left/right or MS form. matrix means (2) for generating a signal (40) represented by can be used. If the output (40) is generated in MS form, the output signal (40) The connection to the reproduction loudspeaker (50) involves the necessary other MS matrix stages. will be understood.

According to one form of the invention, the reproduction matrix decoding means (2) Stage (2) is arranged so that the output signal (40) of the matrix means (2) is transmitted to the second volume of the loudspeaker. When reproduced by the acoustic configuration (50), it is supplied to the target first three-dimensional acoustic configuration. It is necessary that substantially all of the energy of the input signal (20) applied should be conserved. explanation For simplicity and without limiting the invention, the signal supplied to the loudspeaker is the same as the signal emitted into the room by the loudspeaker, within the gain constant. Assuming that all loudspeakers have the same characteristics and flat frequency response, It is convenient to In this case, the total energy released into the room at each moment is separately is the sum of the squares of the loudspeaker supply signals, which also reduces it to L22+Rp2=Mp Since it can be easily shown as “+Sp”, it can be expressed in total energy or MS form. equal to the sum of the squares of the signals.

To maintain level balance between different components due to original stereo effect , aesthetic reasons and the ability of the ear to distinguish distance effects of sound sources are initially related to direct sound. Because it is partly concerned with maintaining an accurate level balance between the reflections of It is generally desirable to conserve energy both in terms of

There are many different recording and mixing techniques that save virtually all the playback energy. can be used to prepare original stereoscopic effects, and some stereophonic signals By giving a component a gain that differs by, say, more than 3 dB from the other gain. A regeneration matrix decoding means that substantially deviates from conserving total energy. (2) accidentally canceling out or enhancing some signal components of an unpleasant audible type; Until now, it has not been recognized that this is possible. For example, The time delay between the stereo components of the Compensation that creates a stereoscopic effect, as in the case of stereo or time-delayed stereo panning. With help, we can transform this non-invariance of the energy gain into a position-dependent comb frame. Filter coloration occurs so that some frequencies of the sound are distinct from other nearby frequencies. It is possible to obtain different gains.

Therefore, to ensure usability using a wide range of recording or blending techniques, It is desirable that the reproduction matrix decoding means (2) should be substantially energy conserving. However, there is no gain or tonal integrity that affects all component stereo signals equally. It will be appreciated that physical adjustments may be tolerated for reasons of convenience or desired effect. vinegar using reproduction matrix decoding means (2) between different components of the teleo signal; It is desirable that the gain variation at the frequency generated by the This variation in gain should be less than 2 dB to obtain high quality results. Ideally, it should be less than 1 dB.

From 2 speakers to 3 speakers A first embodiment of the invention will now be described with reference to FIGS. 3 and 4. In this case , stereophonic sound signals L2 and R2 for two loudspeakers as in Fig. 1b, Fig. 3 Loudspeaker for 3 loudspeakers as in Figure 10 or Figure 1f as shown in the schematic diagram of It is desirable to convert the supply signals to 1, C1, and R3. 3X2 regeneration matrix The signal decoding means (2) receives an input signal (20 from R2 and R2 of the form s=( +sinφ) (R20R2)+ + W (R2R2) Rs= (+sinφ) (R2+R2)+W(R2R2)Cs=(2-1/2cosφ) (R2+R2) produces an output signal (40). where the given angle parameter φ is preferably selected in the range of 15° to 75°, and the predetermined width parameter W is ideal. is chosen to be very close to the value w = 1, and has a reasonably wide sound stage. In order to obtain a reproduction of It is fixed so that interference occurs, that is, W is larger than sinφ.

Using simple algebra and trigonometry, in the ideally preferred case w=1 Ls2+ You can confirm that “C32+R32=L22+R2” is obtained, so try again. The raw matrix decoding means (2) are energy conserving.

Figure 4 shows a 3×2 reproduction matrix that satisfies the decoding means equation of the above 3×2 matrix means. 2 shows a schematic diagram of the box decoding means (2). The first MS matrix means (31) is receives input signals L2 (21) and R, (22) and produces signals M2 and S2; Ru. The difference signal s2 is provided with an optional smooth gain adjustment (32) w to control the playback stage. The required adjustment of the page width is made to produce the signal s3. The sum signal M2 is a constant power for gain adjustment. pairs or sine/cosine potentiometers or their squares up to 1 of gain adjustment that produces two outputs with gains COSφ and sinφ. It is sent to a circuit (33) like this. This circuit (33) is a fixed pair of gain adjustment stages or can be constructed from a fixed resistor circuit for a constant value of the parameter φ, and can be constructed from an adjustable circuit that produces a constant power output. Gain COSφ The signal C3 with The signal H1 having insinφ is fed to the MS matrix means together with S. , a three-speaker stereophonic configuration as shown in Fig. 3, Fig. 1c, Fig. 1f, or Fig. 1g ( Signals Ls (41) and Rs (43) suitable for feeding to the outer loudspeaker of 50) occurs.

There is no single value of the angle parameter φ that gives the best subjective result in all cases. stomach. For example, a small value of φ of about 20" provides good stability of the center stage hallucination image. However, the reproduction of the three-dimensional stage width is poor and uncertain, and the ideal three-dimensional seating position relies heavily on acoustic mixing for the listener (4) at Edge-edge hallucinations produce a high degree of image movement relative to imaginary images. Large parameter of about 60° The value of φ increases the stage edge frame at the cost of poor center stage image stability. Provides stable and wide reproduction of sky images. Typically values of φ between 32° and 55° improve Compromise reproduction in which the stability of the central image is traded against the deterioration of the stability of the stage line image. give. It is typically found that values of φ between 35° and 45° are generally preferred. However, the width and stability of the stage line image are still not ideal.

However, values of φ within the range of 15° to 75° will result in frequencies lower than about 700 Hz. and the value of φ at about 35° is about 4kHz. It has been found to give reasonable stability of image and width at frequencies up to. Approximately 5k At frequencies above Hz, a good sense of stage width and central image stability is typical. is given for a value of φ of approximately 55°. For most stereophonic signals, medium The stability of the central image is determined primarily by the frequency between about 300) IZ and 5 kHz. However, frequencies above about 5kHz are important in creating the feeling of wide stage width. be.

The authors have determined that the value φ = 35.26° or its vicinity at frequencies up to a maximum of 5k)lz. , and at frequencies above about 5kHz, φ = 54.74° or around it is reasonable. A teacher's hand (4) in an imaginary three-dimensional seating position and a pair of teacher's hands crossing a large listening area. We found that this method gave the most generally satisfactory results. Typically, re compared to a two-speaker stereo that covers the same sector of the live range (3). The degree of angular image movement with respect to the direction of the loudspeaker as the position changes is related to the central hallucinatory image. The angular image shift relative to the loudspeaker direction for the stage edge hallucination image is reduced to approximately 173. The magnitude is significantly similar to that for central hallucinations. The authors use approximately 5kHz. The exact value of the transition frequency is not important, but transition frequencies lower than 4kHz are poor I found that it gave good results. The low and high frequency values of parameter φ and The transition between It turns out that it's quite gradual and not sudden.

FIG. 5 shows one embodiment of the frequency-dependent matrix variant of the invention. input signal MS forms M2 and S2 of L2 (21) and 11!2 (22) are MS matrices. The difference signal S2 is generated by the box means (31) and the difference signal S2 is directly connected (37) as before. to an optional width gain adjustment (32). The sum signal M2 consists of two sets of signals. It passes through a band splitting filter (34) which divides it into frequency components. Typically this f The filter is composed of a low-pass filter (34a) and a high-pass filter (34b), Their outputs are added to their input M2. For reference, these filters are , a complementary-order filter with a crossover frequency of about 5 or 6'kHz, or An RC filter may be used, but for a sharper transition rate, use a second-order or higher-order filter. can be achieved.

The high frequency signal component of M2 from the band dividing means (34b) is transmitted to the constant power gain adjusting means ( 33b) to produce gains COSφ□ and sinφ8 as shown.

However, φ8 is the required high frequency value of the angle parameter φ (typically about 55°). Bandwidth The low frequency signal component from the dividing filter means (34a) is transmitted to another constant power gain adjusting means (34a). 3a) to produce gains COSφ8 and sinφS as shown. child where φ - the desired mid- and low-frequency values of the parameter φ (typically around 35°) . The sinφ outputs of these gain adjustment means (33) are sent to the addition means (36). The caSφ output of these gain adjustment means (33) is outputted by another addition means. (35) to produce signal C5. Signals H5 and S are sent to the second MS matrix is sent to box means (39) for left and right signals and produces R5. three signal Ls(41), C.

(42), and R, (43) are the three spindles in Fig. 3, Fig. 1c, Fig. If, or Fig. 1g. is a loudspeaker supply signal suitable for use in speaker configurations.

Various variations of FIG. 5 will be apparent to those skilled in the art. For example, a band division filter (3 4> can be placed after the constant power gain adjustment stage (33) instead of before. band A band-splitting filter (34) is configured such that its output is an all-pass response rather than may be used to substantially add to the The parallel all-pass filter (37) with For example, as shown in Figure 5, in series with the S2 signal path, It should be installed.

Certain preferred embodiments have the same phase characteristics to eliminate phase differences between all paths. Filters (34a), (34b), and (37) are used. This is it For example, a low-pass means (34a) consisting of two cascade-connected low-pass stages and a reverse polarity A high frequency means (34b) consisting of two cascade-connected next high frequency stages is provided. By using a first-order all-pass circuit (37) in which the filters have the same time constant can be achieved.

The frequency-dependent variety of the present invention has a band-splitting circuit (34) at its input or an all-pass response. a band-splitting circuit consisting of filter means producing three or more outputs which are substantially summed with each other; can be extended to the case of circuit (34), whose sinusoidal gain output is The cosine gain output is sent to the calculation means (36) to generate a signal base, and the cosine gain output is sent to the addition means (36). (35) to produce a signal C5 corresponding number of constant power gain adjustment stages (33) Supply 3 or more. Such varieties of the invention may be used, for example, at frequencies below about 200 Hz. One value φ, of the parameter φ at low frequencies, approximately 200) between 1z and approximately 5kHz a second value of φ at φ−, and a third high frequency value of φ8 above about 5 kHz. It can be used to select.

As before, values of approximately φ, = 35' and φ8 = 55° were found to be satisfactory. Therefore, the value of φL at very low frequencies is relatively important for stereophonic effects. I know there isn't. The value of φL is adjusted in the range from Oo to 90° and Satisfactory results can be achieved taking into account the performance of the three loudspeakers at the frequency Ru.

In general, the bass response spread of very small loudspeakers is poor, and convenience, cost, and For reasons of space, appearance, or physical size, the bass response is extended as shown in Figure 3. It may be desirable to use only one or two of the three loudspeakers. During ~ If the bass response of central loudspeaker C5 is poor, or if it is inferior to R5, φ around 90°, The value of can be used to minimize the bass delivered to the center loudspeaker.

If instead only the central loudspeaker had an extended bass response, the value φL near 0° would be The bass signal to the two loudspeakers is minimized. Similarly, in addition to three small loudspeakers, A system that uses one "superwoofer" for the bass frequencies is It works best if you use a small φ and get the supply to the superwoofer from the 03 signal. do.

If the three loudspeakers have substantially the same bass response, as shown in FIG. φ,=φ. or instead use a value φ, close to 54.74°. can be done. The latter value typically includes the most A strong bass, the central bass, is played with the same energy from all three loudspeakers. This maximizes the bass power handling capability of the loudspeaker configuration and provides the best subjective bass response. It has the advantage of being large.

In applications where the loudspeakers have different bass response characteristics, the phase response between the three loudspeakers is In order to compensate for the answer difference, the outputs (41), (4 It may be desirable to incorporate phase adjustment means in positions 2) and (43).

If φ, = φ8 (as when the two are equal to about 55°), the lower part of Fig. 5 filter means (34a) for frequencies between approximately 200Hz and 5k)lz; The high-pass filter means (34b) can be replaced by a complementary bandpass filter. can be replaced by stop filter means.

The above transition frequency of 200Hz is for illustration only, and φ=φL and φ=φ. The transition between It is understood that it can be even higher.

In any of the above 3X2 matrix decoding systems according to the present invention, if necessary, It is possible to make inW frequency dependent. This is a frequency lower than 600Hz can be particularly advantageous. At such frequencies, even lower frequencies such as 1. It is sometimes found that increasing the width by a factor of four increases the spatial quality of the recording.

Hierarchical 3-channel transmission system Referring to FIGS. 6 and 7, the 3×2 regeneration matrix force equation and procedure described above will be explained. A hierarchical system based on stage (2) will be explained. This system is for each Two signals from a 2-speaker or 3-speaker stereo source (1b or lc, respectively) No. (21b.

22b) or three signals (21c, 22c, 23c) as a transmission matrix code. encoded by encoding means (7 or 7b) to produce a transmission channel signal (60); , this signal can be transmitted over, for example, electrical wires, broadcast or telecommunications channels, tape or digital Consists of disk recording and playback channels, digital storage channels, etc. is transmitted via a transmission channel that can be signal means (9 or 9b), and two speaker signals (41b and 4 2b) or 8 speaker supply signals (41c% 42C- and 43c) .

Figure 6 shows three speakers supplying signals 5, C3 and R5 and receiving them through transmission means (8). A 3x3 transmission transmission map that generates the transmission channel signals R, R, and T that are transmitted by trixing means (7), and reconfiguring three speaker supply signals L3, C5, and A 3x3 transmission transmission matrix child stage (9) is shown for generating R5. Figure 6 also shows two steps. Picker left and right signals L2 and Ft2 are signaled and transmitted as =L2 and R-R2 Direct transmission is shown. This kind of 2-speaker transmission is a 3x3 transmission decoding means (9) and the resulting 3×2 recycled playback map according to the third aspect of the invention. Trix child stage (2) performs 3×2 decoding according to the above 3×2 matrix decoding equation It is necessary to be a vessel. Also in accordance with the present invention, a 3×3 transmission transmission matrix element Stage (9) is the reverse of the 3x3 transmission matrixing means (7), and has three speakers. It is necessary that the supplied signal be restored virtually unchanged after three channels of transmission. be.

A suitable equation describing the transmission matrix decoding means (9) according to the invention is s= +(sinφ'+w')L++(sinφ'-w')R+ (2-""cosφ ”)k’TR5=+(sinφ’-w’)L++(sinφ’+w’)R+ ( 2-"2cosφ")k'TC3=(2"2cosφ')(L+R)-(si nφ”)k’T, where the angular parameters φ’ and φ” are 15° and 75°. The width parameter W′ is preferably equal to l, In any case, it is desirable that the third channel gain be larger than sinφ′. The input parameter ``can be equal to 1 or some other predetermined non-zero value.

The third transmission channel is replaced by the 0x sign, and the L and R transmission channels are For two-speaker stereo signals L2 and R2, a 3X3 transmission decoder (9 ) can be used, for example, in a 3×2 reproduction mat of the form previously described with reference to FIGS. 3 and 4. It will be seen that it operates as a Rix child stage (2).

The inverse 3x3 transmission coding matrix means (7) then has the following form according to the invention: The inverse of the above equation must be satisfied.

The form of the equation for the 3x3 transmission matrix (7) is a 3x2 regeneration/decoding matrix. It will be appreciated that this will depend on the requirements of the invention regarding the

The angle parameter φ゛ (which determines the result of the 3 × 2 reproduction signal matrix) and It is desirable that φ” and φ” should be equal to 1. The values of φ° and φ” in the equation are preferably between 32° and 55°, 45° is a highly preferred choice. W゛The preferred value is very close to or equal to 1. Yes. If φ′=φ” and w′=′=1, then 3×3 transmission power and sign The form of the equation is the same, and the same matrix is used for both encoding (7) and decoding (9). xx means can be used.

In the conventional technology, two signals L2 and R2 are used for the purpose of two-speaker three-dimensional sound reproduction. are left and right transmission signals respectively in left/right form =S2, R=R2 or in MS form as sum and difference signals M=82=2− 172 (L+R) and S = 32 = 2-1/2 (L-R) As shown in FIG. 7, it is possible to transmit using the MS transmission encoding means (7a). What can be done and how the reproduction 2 speaker supply signals L2 and R2 can be converted to an inverse MS matrix It can be restored by means (9a) of S, but -8 is transmitted instead of S. I know what I can do. Similarly, the above three-channel hierarchical transmission Transmitting system signals as signals M, S and T instead in MS format is 1. However, M = 2-1/2 (L + R), S = 2-1/2 (L-R ) and 5% R1 and T are predefined coded from 3, C3 and R5. This is the signal that is being used. Figure 7 shows the present invention when using MS transmission channel signals. A schematic diagram of a hierarchical transmission system is shown.

The MS form of the above 3X3 transmission decoding and encoding equations is M3=(sinφ ’) M + (cosφ one’tC3= (cosφ two M-(cosφ″)k’TM =(cos(φ′-φ”))-1[(sinφ”)M3 + (cosφ”)c s], which means that the equations for encoding and decoding a hierarchical symmetric transmission signal are generally This shows the fact that it has the simplest appearance in MS form.

By using the above transmission 3x3 decoding and 3x3 encoding equations, the 3-speed transmission The three-speaker stereophonic sound reproduction device was originally intended to be a three-speaker transmission signal. The manner in which the effect is received and decoded by the 3x2 reproduction reproduction matrix unit of the present invention. Receives a two-speaker stereo signal transmitted by. This allows 2 speakers and 3 The material generated from the loudspeaker stereophonic sound source can be processed by the listener without having to change the decoding device (9). Without the need for a program such as that shown in FIGS. 6 and 7 by the adder means (7o) Can be freely mixed together in creation.

Accurately rectifies a 2-channel signal from material emanating from a 3-speaker stereo source (LC) A two-speaker stereo listener who receives R or M and S has the parameter φ゛. , φ", and W' to obtain satisfactory two-speaker expressions for the preferred values described above. It becomes.

If the left and right extreme source images are smaller, the left and right extreme source images are and is played back at a position wider than the extreme position on the right.

For the above 3×3 transmission and decoding equations, the angle parameter φ′ is given as The disadvantage of using a constant value is that decoding of two channels by three speakers is the optimal frequency. This means that it does not have any form of dependence. Use frequency-dependent encoding parameters Although it is possible to do so, there are two disadvantages to this. (i) 2 channel transmission The signal and R are frequency dependent, so two-speaker reproduction is an optimal fit. It doesn't belong to me. and (ii) frequency-dependent standardization could further improve subjective results. make future modifications impossible.

If only two transmission channels are accepted, the transmission decoding matrix is Switching or adjusting to provide a decoder with frequency-dependent values of meter φ I can do that.

Alternatively, stereo source material originating from two or three channels If you want to mix both the two-speaker stereo signals L2 and R2 together, First, the 3-speaker configuration is achieved using a 3x2 matrix reproduction/decoding means as shown in Figure 5. converted and then fed into a 3x3 encoding matrix (7) to create three transmission signals. Ru. By this means, the decoded signal threshold obtained after transmission matrix decoding (9) , C5, and R3 are frequency-dependent matrix regenerative decoders like the one in FIG. will be the same as if used by the final listener.

If a frequency-dependent 3×2 regenerative decoder is used before the transmission encoder (7), the frequency-dependent The transmitted signal produces %R1 and T, which is This is a disadvantage in the hands. However, in preferred variants of transmission encoding and decoding, this The disadvantages can be very small as shown below.

Let the transmission parameters be such that φ°=φ” and w’=’=1. For a frequency-dependent 3x2 regeneration matrix decoder such as the one in Figure 5, Li=+(sinφ+1)L2++(sinφ−1)R2Rs=+(s inφ−1)L2−+(sinφ+1)R2C5=(2−”2cosφ)(L 2+Rz) where φ varies with frequency, typically between 35° and 55°. transmission After matrix encoding (7), the transmission channel signal, R and T are L = + (cos (φ-φ')) (L2+R2)+'r (L2 R2) R=+( cos (φ−φ’)) (L2+R2)’r (L2 R2)T=( 2"2sin(φ-φ')) (L2+R2), so 1φ-φ If the angle is small, for example less than 25 degrees, then approximately equal to L2 and R2, respectively, as required for compatibility with . φ′= If φ”=45° and φ changes with frequency between 35° and 55°, then 1φ −φ′1≦lO°, therefore coslO′ = 0.9848≦cos(φ−φ ゛)≦1゜This has little effect on the transmitted and R signal, and the resulting left/right interference The signal can transmit L=L2 and R=R2 at frequencies as low as -42 dB, and j A third chip equal to R7 passed through a frequency dependent gain equal to an (φ − φ゛) The signal T of the channel can be transmitted. Referring to FIG. 8, in practice, L2 and M2 is obtained by the MS matrix (31) in response to R2 and R2, and then it is T can be obtained by obtaining T through the router (38). at low frequency 4' = 45' and φ = 35°, and for φ = 55° at high frequencies, Filter (38) with gain (38b) equal to janlo° = 0.176 and low Therefore, at low frequency φ = 45° - 10° = 35°, and at high frequency An all-pass circuit (38a) with a frequency of φ = 45° + 10' = 55° is provided. I can do that.

In FIG. 8, the input stereo signals L2 and R2 (21, 22) are and the difference signal S2 is sent to an optional width gain controller (32) (optional). optionally) to generate a (optionally) modified difference signal S (62). Ru. The sum signal H2 from the MS matrix (31) generates the signal M (61). and is also sent to the filter means (38) described above to produce a third signal T. (63) is generated. Three signals M%S, T are 3-channel transmission in MS form number, and you want to perform physical acoustic frequency-dependent and (optional) width control. supplying a signal obtained from a two-speaker stereo source to a transmission system according to the invention; It can be used for. The part (7) of FIG. 8 explained so far is 3 according to the present invention. A ×2 transmission transmission matrix is constructed. Ensure phase alignment of the three channels. If it is necessary to maintain the M and S (or S and R) signal paths (61 , 62) to output the T-channel filter means (38). It is possible to generate the required phase difference with

In addition, a 3-speaker transmission manager is used for each of the M, S, and T signals (61 to 63). Trix decoding means (9) are provided to perform the decoding as shown in l1lc, Figure if, and Figure 1g. None 3-speaker stereo signal (4 1 to 43), FIG. 8 shows an alternative frequency-dependent structure to that shown in FIG. 5 according to the present invention. A 3×2 reproduction matrix decoding means is configured.

The means shown in Figure 8 also incorporates switches (not shown) in the signal paths (61-63). As input 2, instead of input (2L 22) of R2 form, 2 channels of MS form channel and three channel transmissions. In addition, the T signal path (63 ) of the filter (38) obtained from the 2-channel input. Instead of a combined third channel signal at the output, a three channel transmission source, R , T may be accepted.

Frequency for generating loudspeaker supply signals for n loudspeakers greater than 3 according to the invention The number dependent n,X2 regeneration matrix decoding means in FIG. In place of the means (9), an nX3 transmission matrix decoding means of the type described later is placed. This can be achieved by changing.

Series connection of matrices There are many possible R2x nl regeneration matrix decoders according to the present invention, and A clear explanation of each case is an extremely difficult task. Therefore, the simpler It is convenient and useful to consider a "composite" decoder constructed by a series connection of . It has three sequentially larger plurality n1, nl, and R2, and also shown in FIG. As shown, one R3 Xnl regeneration matrix decoder (2a) according to the present invention n2Xn also according to the invention, having a regeneration matrix decoder (2b) If so, the result of connecting two decoders in cascade is 01 decoders from stereo source (1). input signal (20) is converted into n9 signals ( 20a), then n2xn, matrix (2b) converts n7 signals. (40) and the n2xnl regeneration matrix decoding according to the present invention (2) Configure.

In particular, each element decoder (2a) and (2b) The composite decoder (2) operates that way if energy is conserved. Element decoder ( 2a) and (2b) each substantially preserves or improves the intended stereo effect. If the composite decoder (2) operates as such, the elemental decoder (2a) and Each of (2b) is inside the proportionality constant as the playback speed vector at the ideal listening position. or the angular position of the reproduced sound intensity vector, the composite decoder ( 2) works like that.

A composite decoder based on two known decoders according to the invention uses two known element decoders. There is no need to physically realize and connect together the devices to configure them; instead, one of ordinary skill in the art Achieves the same final result as cascading two known decoders in an obvious way Can be configured as a single matrix circuit or means designed to Wear. In particular, the matrix coefficients of the n3Xnl matrix decoder (2a) are R3 x nt represented by matrix Rn3nl, lo2×n, matrix decoding The matrix coefficients of the device (2b) are expressed by the n2xn3 matrix Rn2n3. Then, the matrix coefficients of the composite decoder (2) are n2xn, which is more expressed.

Rn2n, = Rn2n3Rn3nl Therefore, for any plurality 02 greater than nl, n2xnl according to the present invention The regeneration matrix decoder performs (n+1)Xn regeneration according to the invention for each plurality n. As long as you know how to design a matrix decoder, as shown in the schematic diagram in Figure 10, It can be designed by increasing the number of series connections. Figure 10 is from Figure 18 to Figure 1e. Continuous signal sources (la to le ) is shown. (For completeness, the case of a single note is shown.) For n-1 to 4 The continuous (n+1)Xn reproduction matrix decoding means (2c to 2f) is (n+1) XnVnl trix Il+Iln', from n input signals, (n+1) loudspeakers (50 for each n+1) as schematically shown in 1O. (n+1) loudspeaker stereo signal supply suitable for supplying (a to 50e) produces (n+1) output signals representing . Figure IO can also be used for mixing means or additional means. Outline how to mix together the signals generated for different numbers of loudspeakers using stages. It schematically shows that the signal for one number of loudspeakers can be converted to a larger number according to the invention. This shows how it can be played back.

Although Figure 1O only shows stereo with up to 5 speakers, more For large matrices, e.g. 6x5 and 7x6, how can this schematic diagram be used? It is clear that it can also be extended to multiple loudspeakers. The most practical regeneration matrix In a binary decoder, most or all parts of the schematic diagram in Figure 1O are clearly implemented. 10, but such a decoder nevertheless can have an overall effect comparable to that of the No. 1 route.

General hierarchical transmission system A specific (n+1)Xn regeneration matrix decoder R1+2. according to the invention. . is 2 Before explaining how it can be designed for larger n, we The knowledge of the regenerative decoder is used to encode the transmission signal and the transmission signal according to the present invention. This section describes how to design a hierarchical system for decoding from I'll do it.

Figure 11 shows n1 signals (20) from speaker stereo i (1), mX n1 matrix Emn, ff1Xnj transmission matrix encoding A general system for encoding m transmission channel signals (60a) by means (7) These transmission channel signals are then connected to the selected transmission medium. n2x+ carried by body (8) and marked n2xa matrix Dn2m received as m signals (60b) supplied to the n transmission matrix decoding means (9). The three-dimensional sound structure extends over the fan-shaped area (3) at the position of the raised hand (4). 02 signals (40) representing the supply signals to the 02 loudspeakers forming the configuration (50). Occur. The encoding/transmission/decoding signal path (2) as a whole corresponds to the source signal (20). n2xn constitutes reproduction matrix decoding means.

This overall reproduction matrix decoder (2) comprises a second plurality of n2 loudspeakers. The supplied signal is larger than the first plurality n, and the transmitted signal of the third plurality m is smaller than the first plurality n. If not, the present invention should be followed. Also, the first and second plurality are equal , that is, when n,=j12, the third plurality m is not smaller than these, the regenerative signal equalization that can affect the overall gain and all signal paths equally. Regardless, it should be the same as originally intended.

According to matrix notation, these requirements are R2>n1 and 1≧n. When Dn2mEIln1 = Rn2n1 It can be written as However, the 8-port 2nl is the n2xn according to the present invention, the regeneration matrix is the matrix description of the box decoder, and for m≧n, DnmEmn= In=EnnDnn In this case, In is the nXn identity matrix, and in this case, the author's number is For convenience, we exclude all changes in overall gain and localization from these considerations (because , the n×n encoding and decoding matrices are mutually inverse.

In addition, as shown in FIG. 12, the present invention described as n3Xnl matrix Rn3nl fi, Xn, regeneration matrix decoder (2g) followed by ll according to the invention Xn3 Mad 1 Knox Ecn, Xn3 transmission matrix decoder (7g ), but these are also according to the present invention for %N ns thicker than nl. A transmission matrix decoder (7) should be configured. In other words, FIG. Both the reproduction matrix decoder (2g) and the transmission matrix decoder according to the invention llXn1 matrix Emn1 consisting of serial connection of A composite transmission matrix decoder (7) is shown. In this case bn1=Ecn Jn3n1 Similarly, FIG. 13 shows the composite transmission matrix decoder (9) according to the present invention Reproduction matrix decoder (2h) according to the invention of transmission matrix decoder (9h) This figure shows how it can be configured by connecting it in series. n1XIl matrix Dn The present invention follows the nlxm matrix transmission decoder (9h), denoted as , ffi. n4xn2 regeneration matrix marked as n4xn2mad IJtxRn, n2 by n4XIl transmission marked 14Xm Mad 1 Knox. It constitutes a matrix decoder (9). However, n41 is larger than n2 0°D n, ++ = Rn4n2Dn2IllI1 layer system transmission encoding and decoding Apart from the above requirements regarding the matrix, the preferred form of the invention is shown in FIGS. and imposes another advantageous requirement as shown in FIG. i.e. n channel The loudspeaker signal may be encoded into n transmission channels with each first plurality n and (n+1) transmission channels required for (n+1) speaker stereo transmission. In addition to n speaker stereo, there is also one channel named Tn+1. Now configures the n channels used to transmit the That is what you should do. FIG. 14 shows the n speaker system including the case of single sound n=1. Transmission channel signal nXn matrix En for the rheo source (la to le) The respective encoding means (7b) to (7e) marked as A schematic diagram of such a hierarchical system for encoding into channels is shown. Here in this book Transmission channel signals M%S1 and and T, T1=M, T2=S1 and T3='r, and T4 and T5 are Carry separate signals for 4-speaker stereo and 5-speaker stereo respectively used for. Figure 15 shows the corresponding power signal hierarchy, which is an nXn matrix. Each (n x n) decoder (9b) to (9e) marked as 0 songs is Transmission channel signals M, S1T%T4, and T, to n speaker stereo drive n signals representing loudspeaker supply signals for the corresponding loudspeakers.

As in the case of FIG. 1O, FIGS. 14 and 15 show an even larger number n channels. It can be expanded without limit to incorporate. Left/right signal L=2-1/2(M+ S) and R = 2-1/2 (M-8) in place of the MS-form signal 4 and Figure 15 variants require transmission and reception compatible with two-channel left and right signals. It is obvious when

In relation to FIG. 10, to construct Rn2n1 for any 02 greater than n, The (n+1)Xn regeneration matrix decoding matrix according to the present invention as used If there is knowledge of the and all matrix equations with the form shown in Figure 15. constructing a hierarchical system for encoding and decoding transmission channel signals of the preferred form described above; It is possible to undertake a systematic design procedure to Explain this design procedure However, this is summarized in the flowchart of FIG.

In general, the R22 and D22 shapes are used in the prior art to transmit two-speaker stereo. By traditional left/right or MS matrix encoding and decoding methods used for Given. At a given stage of the design procedure, for each plurality n' up to a maximum of n; n'Xn' decoding matrix Dn'n' and inverse n1xnl decoding matrix Let's imagine that we have determined the shape of SEn'n''=(Dn'n')-'. The known ( n+1)xn reproduction matrix R1゜,. If we give (n+1)X (n+1) f1 matrix Dn+j nil can be created as follows. Wear. D11+ representing the response to the first n transmission channels Tl to Tn 4 The first n columns of n+1 are (n+1)Xn matrix R11+1 nDnn form a convenient non-zero column vector whose last column is not a linear combination of the first n columns Select so that Next, Ell+1 l, il is the inverse of the decoding matrix (D Calculate as n+I n14)-'. Next, increase the value of n by 1 and proceed with the design. Repeat steps above.

In the above design procedure, the selection of the last column of Dn+1 n+1 is very arbitrary, but it is preferable In the example, it is limited to expedient earthworks. For example, if all matrices are A preferred regenerative decoding method is to have real frequency-dependent entries, as is preferred for The matrix is Rn + I n conserves the total signal energy, so their columns are simply We can use the fact that the matrix D nn is the last column of Dn++r++l in each stage, for example, a matrix Using the Durham-Schmidt orthogonalization process found in algebra textbooks, By simply constructing it to be that unit length vector orthogonal to the other n columns, It is possible to ensure an orthogonal matrix in each stage. This is the decryption matrix The matrix Dnn is orthogonal and the inverse n×n encoding matrix Enn=Dnn-' is the transposition of Dnn, i.e. if the D song has an end 1ba dij) then the end 1) matrix, which can be conveniently calculated as It can be shown that a layered encoding and decoding system results. φ゛＝φ” and the previously described 3X3 encoding and decoding matrix with '=1 in w'= was an example of an orthogonal encoding and decoding matrix obtained with this procedure.

More generally, the last column of D1141n+1 is Tn+t for odd n. Let be a linear combination of signals consisting only of the MS form sp, and T for an even number n Let l++ be a linear combination of signals consisting only of MS form Hp or buttons. Accordingly, the selection can be made to satisfy the left-right symmetry requirements.

energy conservation decoding The design of the n2xnl regeneration matrix decoder according to the present invention is divided into two main parts. It will be done. The first is that the decoder has a possible overall gain that affects all signal components equally. and the total energy of the stereo signal passing through them, apart from changes in equalization. This imposes an objective requirement that it should be preserved in terms of quality, and the second is that it should be preserved. Even more important is the need to substantially preserve or improve stereo directivity effects. visual or psychoacoustic requirements. Dealing first with energy conservation requirements It is convenient.

n2xn that describes the reproduction matrix decoder, matrix Rn2nl is that n A column is of unit length (i.e., the sum of the squares of the absolute values of the matrix coefficients for that column is 1). equal), and the columns are orthogonal in pairs (i.e., one column of the sum of the products of an entry and the complex conjugate of another corresponding entry is 0), and Only then will energy be conserved. In matrix terms, this is Rn2n 1 is the first 01 column of the nl x n2 identity matrix, or all has a real value, then the first entry of the n2xn2 orthogonal matrix This means that it is the n1 column of .

The general form of the nXn orthogonal matrix is known to mathematicians, and it is a rotation in n-dimensional space. The family of +(n-1)n parameters of such an nXn orthogonal matrix describing exist. All other nXn orthogonal matrices have the sign of the last column entry You can get from them by reversing them. The product of two orthogonal matrices consisting of Orthogonal. Therefore, if we use known results available in textbooks, we can There is no difficulty in finding examples of energy conservation matrices of the required form.

In particular, all 2x2 orthogonal matrices have an obvious shape for the angle parameterizer. and all 3x3 orthogonal matrices have the obvious form of rotation matrices. or in other cases with the above form with the sign of the last column reversed.

It can be shown that Here, a2+b2+c”=1, and φ is the axis (a, b.

C) is an angular parameter that describes the angle of rotation around C).

If A is an nXn orthogonal matrix and B is an mXm orthogonal matrix, then ( The n+1l)X(00m) matrix is also orthogonal. Symmetrical stereo loudspeaker arrangement In the case of a symmetrical regenerative decoder, the energy conservation matrix is in MS form. When expressed, the reproduction matrix allows the sum signal (i.e., in the form of Mp or cp ) must be converted into a sum signal, and the difference signal (i.e. also of the form Sp ) has a particularly simple form since it has to be converted into a difference signal.

Therefore, the energy-conserving symmetric 3x2 regeneration decoding matrix is expressed by the equation % must be satisfied, but the shape of the energy conservation symmetric 4×3 regeneration decoding bar is Must be satisfied. However, φ, and φ. is an angular parameter.

An energy-conserving symmetric 4×2 regeneration decoding matrix must satisfy an equation of the form Must be. Here φ42 and φ. is an angular parameter. This is φ4 If 2=φ−φ, then it is composed of the series connection of the above 3×2 and 4×3 decoders. A composite decoder (as shown in FIG. 9).

Figure 4 showed the shape of the energy conserving 3×2 regeneration decoding bar according to the equation shown above. will remind you of that. This form can be generalized to multiple other inputs and outputs. I can do that. For example, Figure 17 shows a 4x2 regenerated matrix according to the present invention and the above equation. 6 shows the Rix decoding means. 2 speaker stereo signals L2 and R2 are input It is converted into signals M2 and S2 by the MS matrix means (31). S2 An optional width gain adjustment means (32) may be passed. M2 and S2 each is then provided with a constant power or sine/cosine gain adjustment means, respectively (33c) and (33d). One output from each of these means (33) is is sent to the output MS matrix means (39c) to provide an output signal and produce R4. , the other output from each of the means (33) is connected to the second output MS matrix means (33). 9d) to produce output signals L5 and R5. These output signals L4 and L6 , R5, and R4 are suitable for a four-speaker stereo loudspeaker such as the one in Figure 1d. through the amplifier means, equalization means, preamplification means, and amplification means. can be done. If necessary, the respective sine/cosine gain adjustment means (33c) and and the angular parameters φ4□ and φ associated with (33d). In the case of 3×2, It is possible to make the frequency dependence by the method already explained in connection with the method (33) of 5. Wear.

FIG. 8 shows the above equation and the 4×3 reproduction matrix decoding means according to the present invention.

5, C5, and R1 are input signals for 3-speaker stereo playback. and be accepted. 1 and R5 are provided to the input MS matrix means (31). to obtain signals M3 and S, respectively. S, constant power or sine/cosine gain adjustment means (33e) to produce two output difference signals S4 and S5. Hl and and input C5 are 2×2 orthogonal rotation matrix means (3 3f). , M4 and S4 are the first output MS matrix means (39 e) and yields R4, M5 and S5 are the second output HS master. It passes through trix means (39f) to produce an output signal and R5. signal 5, R6 and R4 feed into a four-speaker stereo configuration such as the one in Figure 1d. Suitable for use as a feed signal. As associated with means (33) of Figure 5, band division If necessary using filter means in conjunction with means (33e) and (33f) Frequency-dependent values of the angle parameters φ3 and φD can be generated.

FIG. 19 shows the present invention, which generalizes the special cases shown in FIGS. 4, 17, and 18. The general form of the energy conserving symmetric reproduction matrix decoding means is shown below. Input MS The matrix means (31) comprises a first plurality of n loudspeakers for loudspeaker stereo. Supply signal (20) to the difference signal 5p (29) of a number n, ° equal to the integer part of +n+ and Hp or a button-shaped sum signal (28) of other numbers n1"=n, -n, +. exchange. The total energy of the sum signal (28) is substantially the total energy of the signal (28). Matrix A means (33g ), and the difference signal (29) has its total energy substantially equal to the total energy of the signal (29). Multiple n2° (which is equal to H2" or H2"-1) equal to the output signal ( 49) to matrix 38 means (33h). Sum signal (48) and The difference signals (49) are passed as a pair through the output MS matrix means (39) and n Generates an output (40) suitable for use as a loudspeaker supply signal for a two-speaker stereo. Ru. Here n2=n2'+n2''.

Matrix A and 8 means (33g) and (33h) are the means of FIG. 5 if necessary. frequency-dependent by means similar to those used in connection with (33) or by other means. It can be done. Energy conservation bilateral symmetry according to the invention not shown in FIG. Other embodiments of the n2xni regeneration matrix decoder are, for example, The functions of stages (31), (33g), (33h), and (39) will be clear to those skilled in the art. This is possible by separating and reassembling in a gentle manner.

Another example of the n2xn1 reproduction matrix encoding equation expressed in MS form is shown below. However, this is the matrix 8 means (33g) and the matrix 38 means Specify the equation for (33h). For example, 5×4 energy saving left and right pair The symmetric equation comprises a matrix A equation that can be parameterized into the form ing.

Here ago, b>o, C>Ol and a2+b2+c2=1, and φ4 is The angle is 1 parameter, and u, = b (a 24 H2) -1/2, μ 2=a(a2+b2 houses/2, u1=ac(a2÷b2)-1/2, LJ 2=b c(a2+b2)-1'2, and λ=(a24b2)-172. Also ma It takes the form of a trix B equation. Here, φ5 is an angle parameter. 4 speed If equal signals are supplied to all four loudspeakers in the configuration, aSb and C are Determine the relative energy reproduced by the five-speaker configuration.

In a similar manner, the 6×5 energy saving left/right regeneration decoding matrix equation in MS form is The first two columns of the 3x3 orthogonal matrix are used to create a 3x3 orthogonal matrix. Matrix “Matrix A” equation and 3X2 “Matrix B” equation will occur. These equations are characterized by having a total of six free parameters. Ru.

directional psychoacoustics The theoretical method necessary to ensure the correct subjective stereodirectional effect from the present invention To show how to determine the optimal values of the free parameters of the equations of an energy-preserving decoder I will summarize it below so that you can:

12120 were all placed at the same distance from the listener (4) in the ideal listening position. This shows the loudspeaker arrangement. A plurality of n loudspeakers numbered with index index i=1 to n exists and a given bitumen generally has a frequency-dependent and complex-valued gain i Suppose it is played from the second loudspeaker. The absolute value of complex quantity 2 is 121, and its real number The part is represented by ReZ, and the real coefficient of its imaginary part is represented by 1mZ.

In this way, for the above-mentioned sound source, the pressure gain at the position of the hand (4) is i = 1 and the energy gain at the listener (4) position is i=1 is proportional to. The (imaginary) forward direction (5) at the position of the listener (4) is the X axis. , the (imaginary) left direction (6) is the y-axis of the rectangular coordinates, and the direction of the listener (4) is , counterclockwise from the x-axis (i.e., toward the y-axis), as shown in Figure 20. Let's assume that the angle is measured as θ1. Then, the velocity gain for the above sound source is Each component along the x-axis and y-axis is VX=ΣGi cosθ1 i=1 A vector quantity v=(vx, vy) is defined. Here, θi is shown in Figure 20. is the directional angle of the i-th loudspeaker. The sound intensity gain for the above sound source is The respective components along the x-axis and y-axis of i=1 The vector quantity e=<ex, ey).

According to the energy vector volume localization theory, the local quality and direction of the listener's sound are Ratio of sound intensity gain vector to energy gain Cex/E, ey/ E) is determined by the magnitude γE and the direction angle θE. γE and θE are rectangular coordinates It can be calculated from the following equation by polar coordinate transformation.

7ECC8θE=ex/E 7EsinθE=ex/E However, γE≧0゜θE means that the apparent sound source, especially 700Hz and 5kHz, It represents the apparent direction of the sound source when it is facing the sound source with a frequency between In the case of , the localization varies greatly depending on the ratio of internal strengths. This direction is the same as the strength of the sound is the direction pointed by the vector. The quantity γE is the magnitude of the energy vector, and equal to l for natural sound sources, but greater than 1 for sounds originating from two or more loudspeakers. is a small value, which describes the stability of the hallucinatory sound image when the listener changes position. useful for.

In order to achieve stable localized quality of natural sound, it is desirable that γE be as close to the ideal value of 1 as possible. Delicious. As an empirical method, the degree of unnecessary image movement of Togi when the listener moves from the ideal position is Since it is approximately proportional to l-γE, if γE = 0.95, the image quality obtained is It is approximately 173 when γE=0°85.

At low frequencies below about 700 Hz for the central listener (4), the local limit is the velocity gain. The vector ratio (vx/P, vy/P) to the pressure gain of the invector changes significantly. In general, this vector has complex entries, but the internal phase The principal localization direction according to acoustic localization theory is its real part (Re(vx/P), Re(vy/P)). Similarly, the energy above For the case of and velocity direction angle θ9 as γyCO8θv=Re (v x/P) γvsinθv Defined by = Re (v y/P). Ideally, for natural sound station quality Therefore, the magnitude of the velocity vector γV should be very close to 1; At much higher and lower values, the image becomes unsettling as the listener's head rotates. becomes fixed. The direction θV is often [known as the Makita local J direction; It was named by the founder who introduced this localized parameter. Makit The a-direction θV follows the internal phase localization theory when the listener is facing the apparent sound source. describes the apparent localization at low frequencies below z. Ideally, Makita direction θV is the same as Binel vector direction θE to sharpen the image. Should.

The imaginary part of the speed ratio vector (ltr(vx/P), In(vy/P)) is Although it is called a "topological vector," it is mainly concerned with the direction of the image rather than its apparent direction. It affects subjective decency, imparting a generally unpleasant decency often referred to as "topological", This is also revealed by the widening of the image. Ideally, the size of the phasic vector should be The size should be kept as small as possible, preferably with a length of less than 0.2. should be. In the most preferred embodiment of the invention, the relative values of the matrix coefficients are It usually deviates only a small amount from the actual value, and such deviations are caused by transition frequencies. Since it is limited to a few bands, the effect of phasing for a listener in an ideal position is Usually reasonably small and can be ignored.

When the magnitude of the phasing is small, the Makita direction θV is the velocity gain base of the signal. Since it is generally true that the direction that the The two directions can be used interchangeably.

For a known signal gain Gi supplied to n loudspeakers, four local parameters The calculations of γV, γE1θv1 and θE can also be calculated using the values obtained from the decoder matrix. all from the ideal listening position (4) for a given loudspeaker signal supply, including This can be done using the above equation for a given loudspeaker configuration at a distance.

These four parameters determine the quality, orientation, and security of the relevant image over a wide listening area. Gives good qualitative instructions.

According to the invention, the first stereophonic sound consisting of n1 loudspeakers over the range of the fan-shaped area a reproduction matrix receiving loudspeaker supply signals of a first plurality n, intended for a sound configuration; The box tX means forms a second three-dimensional sound configuration over the range of the second fan-shaped area. The larger output signals to be supplied to the four loudspeakers are transmitted to four localized speakers. For example, if the values of the in such a way that it is realized or “improved” by including Or improve the stability of the image in the direction where the stability of the image was originally poor for the intended three-dimensional reproduction. It should occur in such a way that it becomes larger. Does the decoder meet these objectives? In order to determine whether or not the local parameters γV, γE, θV and θE are originally Sound through the intended loudspeaker configuration and the N2xNL matrix decoder It is necessary to calculate both the sounds resulting from the final intended configuration.

Ideally, the values of TV and γE are kept equal to 1 by matrix decoder regeneration. or should be very close to 1, and the reconstructed image directions θV and θE are substantially It should be preserved. In practice, it is necessary to maintain the values of γE and TV perfectly. It is not always possible to maintain accurate angular alignment of θV and θE. This is neither possible nor often desirable.

In particular, the stereophonic recording originally intended to encompass the first fan-shaped area in the range of angular width θ■ The recording is played back by a second fan-shaped area of a range that includes different angular widths θ0 at the listener's position. Sometimes it is desirable to do so. Therefore, a simple expansion of the three-dimensional sound localization range is often desired. desirable or acceptable. Generally, if the angular width becomes twice as wide, 1- The value of γE typically increases by a factor of R2 for k less than 1.

In the originally intended three-dimensional reproduction, the angular positions of θV and θE do not match exactly. In general, the local limits of speed and sound intensity are probably constants for each. kv and kI! , are substantially proportional to the original playback directions θV and θE. then, i.e., substantially θV’=kVθV and θE’=kEθE If so, it is acceptable according to the invention.

In practice, the constants kv and ka for different directions across the stereostage are Small variations are acceptable unless this results in fairly significant angular distortion of the sound image. It is possible. For example, if the reproduced angular configuration is given by a strict proportionality constant This is generally acceptable if it does not differ by more than 4° or 5° from Ru.

2 channel stereo localization Regarding the case of a 3×2 regeneration matrix device of the type described with reference to FIG. The application of the localization theory to the present invention will be explained next. For example, panpot positioning or by using spatially matched directional microphones. The respective gains G1=cos(45°-θ) and G2= Two speakers when encoding into two channels with cos (45° + θ) Consider a stereo signal and R2. The parameter θ is defined by the authors as “panning”. "pot angle" describes the three-dimensional position of the intended sound, and for the loudspeaker on the left, θ = 45°, θ = 0° in the center, θ = -45' in the right loudspeaker, and the intermediate value is Corresponds to an intermediate position.

Figure 21 shows the parameters when playing with the two-speaker configuration of Figure 1b when θ2 = 35°. The local parameters calculated and drawn using the equation shown above for the pot angle θ The values of TV, γE1θV, and θE are shown graphically. Generally θE is the center and not equal to but greater than θV except at the left and right extreme positions; It can be seen that θE is approximately twice as large as θV for images near the center. this angle Due to the discrepancy between the two speakers, the image quality becomes poor in transmission two-speaker stereo. other, Near the center of the stereostage, TV and γE drop to values much smaller than 1. , the image stability becomes poor.

In Figures 22 to 25, two-channel stereo signals are processed by a 3X2 reproduction matrix decoder. Figure 4 shows the local parameters when supplied by (in this case the width is set to w=1). 3 in Fig. 1c when θ, = 45° for various angle parameters φ according to This is shown for the loudspeaker arrangement. Figure 22 shows for φ=90°, i.e. , the central loudspeaker supply is O, so the reproduction is from two loudspeakers at θ2=45°. 3-speaker playback. This is suitable for TVs and TVs with a larger angular width. The results are the same as in FIG. 21, except that the deviation of γE from 1 is large. Figures 23 to 23 25 are the respective values φ=54.74°, φ=35.26°, and φ=1 9. Shows reproduction for 47°. θE gradually decreases as φ decreases However, it is noted that θV remains substantially the same. However, compared with Figure 21, Then, the reproduction localized angle is for all values of the angle parameter between 0゛ and 90゛. is substantially proportional to that given by two-speaker stereo, and the decoding in Figure 4 It can be seen that the device meets the requirements of the second aspect of the invention.

From these calculations, when φ = 50.36°, two speakers are used as shown in Figure 21. giving almost exactly the same local angles θV and θE as stereo, and T V also remains almost unchanged, and γE becomes slightly smaller at the left and right extremes, but other is the same over a wide range, so the three-speaker recovery with φ = 50.36° This device virtually preserves the local quality of two-speaker stereo, including its shortcomings. It becomes clear that there is.

As shown in Figure 24, as φ decreases to 35.26', The values of γE and TV approach 1 to improve image stability, and γE is Image stability that is nearly constant across the board and approximately the same at all panpot angular positions give.

Furthermore, θV and θE are virtually equal across most of the stereo stage. This improves image quality. θE becomes too narrow only near the left and right extreme positions Ru. The local parameters shown in Figure 24 explain why a φ of approximately 35° is generally preferred; This also explains why the regeneration stage is too narrow.

As shown in Figure 25, when φ is reduced to 19.47°, even for images near the center, The reproduced θE is too narrow, but the center values of TV and γE are extremely close to 1, The stability of the central image obtained is good. However, the stage edge value of γε is poor. This indicates that the stability of the stage edge is poor.

We hold a uniformly incomplete local parameter for the image position where all values of φ exist. The fact that This is particularly desirable for decoders operating with two-channel stereo inputs. explains the reason.

save decoder It is possible to design an n2xn, regeneration matrix decoder that satisfies the perspective of stereo localization. There are two main purposes it can be used for. On the other hand, the originally intended velocity vector and sound The stage width can be adjusted by preserving the angular configuration of the intensity vector to within one overall proportionality constant. The objective can be to take into account changes in This kind of decoder is It tends to preserve other localized quality as well, such as TV and γE. be.

As seen above, the 3X2 decoder with φ = 50.36° is a conservation decoder in this sense. Yes, it preserves all the drawbacks of a two-speaker stereo.

Another, less well-defined purpose is to improve replay hallucinations. General This means that the playback directions θV' and θE' are set to φ=35. as shown in FIG. virtually equally for most of the playback directions, as the 26° 3×2 decoder did. This means using different values of the proportionality constants kv and ke so that . Also , especially for directions different from 1, perhaps other Re-increasing the value of γE at the cost of reducing γE somewhat in the direction It is possible to use a raw decoder. Such an "improved decoder" is, for example, γ It can be designed to ensure that E is approximately constant in all directions.

The hidden intentions behind the "storage decoder" are pretty well revealed, but the "modification" The intent of "good decoder" is generally a trade-off between conflicting psychoacoustic requirements. There is, and therefore somewhat less obvious. However, a large number of different plays Extensive calculations of the regenerating local parameters of the decoder show that all reasonably improved decoders It is clear that the decoding parameters are not very significantly different from those of the encoder. Therefore, once the problem of designing a conserving decoder is solved, an improved decoder can be created. only requires slight adjustment of parameters.

In principle, the design of the conserving decoder can be applied to very diverse n, loudspeaker stereo supply signals. then calculate the local parameters for the energy conserving n2×nj decoding matrix Calculate the local parameters of the resulting signal for each possible value of the lix parameter. The decoder parameters with the 0th order are very laborious work as it involves calculating the It is necessary to find out whether the meter substantially preserves the required local parameters. This way Such a search is difficult for a 3×2 decoder with only one free parameter φ. However, in more complex cases it becomes difficult and the search is reduced to each possible first and second It is necessary to do this again for the loudspeaker in the three-dimensional sound configuration. Local parameters γV, γ E1θV and θE are highly nonlinear functions of the decoding matrix, and the stereo orientation n to the speaker stereo signal, which can be used to create a stereo effect. There are also many possible speaker gains Gi.

However, the authors have developed various patterns that reduce this design procedure to manageable proportions. I found it. First, we only need to examine the case n2=n1+1. In other cases, Figure 9 and cascading such decoders as noted above in connection with Figure 1O. This is because it can be obtained by Second, the authors The Kuss equation describes the entire loudspeaker arrangement if the relative values of the inter-speaker angles remain unchanged. We found that it changes only slightly as the angular width changes. Thirdly, the brush et al. also found that the conserving decoder matrix can be be relatively sensitive to slight variations and ensure that the angles between all adjacent pairs of loudspeakers in the configuration are the same. It has been found that it can be assumed for the purpose of designing the decoding matrix that

Therefore, apart from a slight "fine match" of the decoding matrix, the investigation in Figure 1 b to 1e and exemplary reference values θ2=35°, θ,=45″, θ4 Is it limited to =50°, θ6=54°, C7=27', and θs"164"? can. Another problem is that, as noted earlier, the large number of different recording and mixing techniques A large number of possible stereophonic sounds that can be fed to n2 loudspeakers, obtained by Select a broadly representative signal to calculate the local parameters. This means that there are difficulties. The gain coefficient Gi of stereophonic sound produced by n loudspeakers is Form a vector (at....On) of n11-dimensional quantities and represent a stereoscopic signal The possible combinations of such signals encompass this region of n11 dimensional quantities. we calculates the values of the regeneration local parameters for representative combinations of points that broadly cover this region. I want to.

In practice, we have found the following choices to be good ones: Next n, Spi - Select the gain of the stereo signal. Try to play from exactly one loudspeaker signal, i.e. for which θV=θE; θ11 and TV=γE=1 Select an isolated loudspeaker supply signal, for which θV=θB=4rCθi−θj) and TV=γE=COS+(θi−θj) with equal polarity and gain. Select the signal to be reproduced by the loudspeakers of set i and set j, which have the following characteristics. this By substantially preserving the intended local parameters of the stereophonic signal from the It can be seen that the localization of the stereoscopic signal is also substantially preserved. These −)nl(n1+ 1) Stereo test signals also help to access the locality of the "improved decoder" .

For these signals, the local parameters of velocity and energy are the same, In particular, θV=θE. This is because all gains Gi are 0 or 1. This is because Gi=Gi2. Therefore, (n+1)Xnt energy conservation and regeneration matrix The reproduction direction parameters θV′ and θE′ specify the lix Rn+In. is equal to the signal. Energy conservation (n+1)Xnl matrix There are +n1 (nl + 1) free parameters that describe This system of nonlinear equations θv゛=θE゛ for (nl + 1) test signals is It should determine the decoder matrix. These equations are highly nonlinear However, it can be done numerically using a computer using numerical hill climbing methods that solve systems of nonlinear equations. I can solve it.

In the case of left/right symmetry, the size of the equation system becomes smaller. For example, for θi=0 , the corresponding θV′ and θE゛ are both equal to O, and θi = −θJ. The peaking symmetry axis is in the direction of Oo. In addition, θv°=θE゛ for the test signal If so, this condition also holds for the mirror image test signal.

For example, if n,=2, then θ for each L2 and R2 Just find the 3×2 decoding angle parameter φ such that v゛=θE° is 1 and O. stomach. If 111=3, then for each (C3, C5, R3) signal decoder parameters to obtain θv゛=θE° (1, 0, 0) and (1, l, 0). We need to find the value of the data. If n, = 4, then each For the (L4, Ls, R5, R4) signal, θv゛20E゛ is (l, 0.0.0 ), (0,1,0,0), (1,1,0,O) and (1,0,1,O) are obtained. Find the values of the 5×41 matrix parameters a, b, c, φ, and φ5. and if n, = 5, then (7) (L6, L7 , C1, R7, R6), -IV'=θE゛ is (0, ■, 0.0.0), (0,1,0,0,0), (L1.0.0.0), (0,1,1,0,0,) , (1,0,1,0,0) and (1,0,O1■,0), energy conservation left We have to find the values of the six free parameters of a right-symmetric 6x5 decoding matrix. The same applies to the case of a larger value of n.

n, loudspeaker as shown in FIGS. 1b to 1e with exemplary reference values of angle θp. (n, + 1) x nl energy preserving symmetrical decoder for teleo arrangement Achieving a "conservative decoder" in the above sense using a numerical procedure that solves the equations for I found the following decryption parameters.

About 3x2 decoder φ=　50.36゜ About 4x3 decoder φ=10.57° and φ. = 28.64' and About 5x4 decoder a=0.6164, b=0.6558, c-0,4359, φ4=51.6 4° and φS = 9.64° left/right configuration, these decoders have the following matrices: satisfies the Kuss equation.

Regarding the 3×2 reproduction reproduction matrix device of the storage decoder J according to the present invention, the present invention Regarding the 4×3 reproduction reproduction matrix device of the “storage decoder” according to the present invention, For the 5X4 playback matrix decoder in "Storage Decoder", the 4X3 and 5 above To form a composite decoder as shown in Figure 9 from the The resulting 5X3 "storage decoder" satisfies the matrix equation and has a similar complex The decoder equations can be reduced to 4X2 and 5X2 by multiplying the appropriate matrices. can be formed from the above equations for the case of . But I've already seen it In the case of a two-speaker stereo signal source, preserving the original effect is a two-speaker stereo signal source. Considering Castele and Leo's essential flaws, it is rare that it is the most desirable I+1 thing. Ru.

In the above method for the loudspeaker arrangements of FIGS. 1b to 1e with exemplary reference values θp The effect of the above-mentioned conserving decoder on the calculated local parameters is shown below by a series of tables. show.

Table 1 shows the 3X2 preservation (Summer) compared to the original 2 speaker values for various input signal gains. The calculated values of the local parameters by the signal generator are shown.

Table 1 Gain 2 Speaker Parameters 3 Subikano 1 Lameter L2R, yV eV rE θE rV eV rE θE1 01.0000 35.00 1.0 000 35.00 1.0000 35.38 0.940435.381 10.8192 0.00 0.8192 0.00 0.8153 0.00 0.8263 0.00 Table 2 shows the original three speaker values for various input signal gains. Show the calculated value of the local parameter by the above 4 × 3 storage device compared with t. .

Table 2 Gain 3 Speaker parameters 4 Subikano (Lameter L3 C, R3rV eV 7E θE rV eV 7E θE1 0 0 1.0000 45. 00 1.0000 45.00 0.9805 45.0B 0.96904 5.0B1 1 0 0.9239 22.50 0.9239 22.50 0.9282 22.32 0.925422.321 0 1 0.7071 0.00 0.70?1 0.00 0.6924 0.00 0.6534 0.000 1 0 1.0000 0.00 1.0000 0.00 1 ．． 0303 0.00 0.9474 0.00 Table 3 shows various input signal gains. The local parameters of the above 5×4 storage device compared with the original 4 speaker values. Indicates calculated value.

Table 3 Gain 4 speaker parameters 5 speaker parameters L4LSR5R41V θx 7F+ θB TV θv IE θE1000 1.0000 50. 00 1.0000 50.00 0.9996 50.95 0.97935 0.950100 1.0000 16.67 1.0000 16.67 1 ．． 0009 16.32 0.960616.321100 0.9580 3 3.33 0.9580 33.33 0.9549 33.75 0.954 633.831010 0.8355 16.67 0.8355 16.67 0.8328 17.56 0.779017.510110 0.9580 0.00 0.9580 0.00 0.9606 0.00 0.9613 0.001001 0.6428 0.00 0.6428 0.00 0. 629B 0.00 0.6377 0.00 Table 4 shows various input signal gains. The calculation of the local parameters by the above 5×3 storage device compared with the original 3 speaker values. The calculated value is shown.

Table 4 Gain 3 speaker parameters 5 speaker parameters L5C, R3rV eV IE θE TV eV yE OF.

1 0 0 1.0000 45.00 1.0000 45.00 0.97 64 45.76 0.968345.740 1 0 1.0000 0.0 0 1.0000 0.00 1.0378 0.00 0.9515 0.0 01 1 0 0.9239 22.50 0.9239 22.50 0.9 272 22.70 0.922622.411 0 1 0.70?1 0. 00 0.7071 0.00 0.6812 0.00 0.6475 0. From Tables 1 to 4, all of these storage decoders are 2 for the proportionality constant kv and for the angular arrangement of eV and θE substantially equal to 1. Other inputs that may be found equipped and used to achieve the desired stereophonic illusion It can be seen that the signal gain also substantially preserves the local angle of the earth to within about 2 degrees. Up As you can see from the table, the values for TV and IE are not saved exactly, but they are In general there is a close similarity in the output of the storage decoder. IE is very busy at input point If the direction of the sound is close to the direction of the loudspeaker, IE can be close to 1. It is impossible to avoid some IE reduction because Ru. However, IE actually increases It is noteworthy that

The numbers shown above are approximations and may vary somewhat depending on the angular width of the loudspeaker placement. be understood. In practice, even if the matrix coefficient changes by about 0.02, it is very Changes significantly greater than this do not appear to be significant, e.g. 0.05 or Variations of 0.1 and 1 are acceptable in many applications.

In practice, angular rotations of the matrix up to 6° (i.e. orthogonal rotations resulting in angular rotations) multiplying the matrix) has a substantial effect on the “storage decoder” according to the present invention. I don't think so. In particular, the angle parameters of the decoder can be up to 6° from a given value. The direction of the (a, b, c) vectors can also change by 6° without any practical effect. It can change.

Two (n,+1) speaker configurations with (n+1)Xnt playback, playback, storage, and decoding capabilities can be designed using the stereo test signal method described above. Table 5 shows the angle θ in Figure 1d. The parameters φ, and φ of the 4×3 conserver for various values of 4 and θ.

This is what is listed.

Table 5 θ4 θ5 φ3 φD 45 9 9.07 33.39 45 15 10.40 28.32 50 10 9.0B 32.72 50 164 10.57 28.6460 12 9.16 31.64 60 15 9.89 30.75 60 20 10.98 29.42 60 24 11.73 28.37 60 30 12.65 26.70 ? 5 15 9.49 30.92 ? 5 25 11.76 31.01 As specified earlier, the values of the decoder parameters are It can be seen that the precise angle arrangement does not change much. Also, all these For the decoder, the direction θV and the direction of the reproduced velocity vector and sound intensity vector are and θE are substantially proportional to those intended by the original three-speaker arrangement.

The storage decoder according to the invention can, if necessary, decode according to the angular arrangement of the loudspeakers used. can incorporate means for adjusting instrument parameters, or can be used.

improved decoder Especially for two-speaker stereo, and to a somewhat lesser extent three- or four-speaker In the case of speaker stereo, the directionality of speed and sound intensity, localized image stability, and In terms of consistency, it is difficult to achieve the desired stereophonic illusion using the intended loudspeaker arrangement. can prove that it is impossible. In such situations, the main The reproduction can be improved by using a brighter decoder with even more loudspeakers. . This allows you to change the decoder parameters from their stored decoder values calculated above. This can be done by

For decoders with two-speaker stereo inputs, the required changes can be quite large. For example, the angle parameter φ is 50.36°, which is the "save decoder" that we have already seen. may be reduced by 15° or 20° from the value of . The n2×2 improved decoder is , as in Fig. 9, the 3×2 improved improver and its successor (from the n2×3 conserving decoder or by forming a composite decoder configured such that can be obtained by using a decoder that has the same overall effect as Ru.

The 4×2 improved improver has the angles shown in Table 5 for the 4-speaker configuration shown in Figure 1d. degree parameter φ. typically about 28+゛; The angle parameter φ42 is substantially equal to 35 at frequencies between about 400 Hz and about 5 kHz. °−φ, (typically about 25°), and frequencies above about 5 kHz may be substantially 55°-φ, (typically about 45°). However, φ , are as shown in Table 5.

This kind of frequency-dependent 4×2 improved decoder is as in the case of FIG. 17, but as shown in FIG. φ4□ sine/cosine gain adjustment using associated band splitting means (34) such as Stage (33c) can be frequency-dependent and implemented. φD sine in Figure 17/ The cosine (33d) can be made frequency dependent as well if desired. like this At the decoder, the bass energy is φ4. close to 90° and φ. loudspeaker near O゛, and to R4, and φ42 at low bass frequencies. and φ near 0゛. to near 90° and give priority to loudspeakers L5 and R5. can be supplied.

A frequency-dependent 4×2 improved decoder is used instead of the output matrix means (9) as described later. Replace the 4×3 transmission matrix decoder means such that the angular parameter φ′ ・45°, φ3 and φ. substantially as shown in Table 5, and implemented as shown in Figure 8. It can also be done.

Most "improved decoders" with n, = 3 or more channel inputs use conservative decoding. The required changes in decoder parameters from the decoder values are much smaller. In general, the decoder With acceptable small changes in the parameters, the velocity vector and the velocity vector of the sound It is known that the values of ν and h are modified to the proportionality constant of the angular arrangement of torque, and the modification A good decoder will increase the value of kE somewhat in the intermediate frequency range to increase γE somewhat. or reduce it, for example to about 0.8kv or 0.9kv, and do not change the kv significantly. It is designed to increase kt somewhat above 5kHz. With this strategy The maximum sense of width is preserved above 5kHz, while the listener's shift at intermediate frequencies is Image stability with motion is improved.

Stereo input gain (i.e., single speaker and speaker pair Numerical calculation of reproduction local parameters, especially θV゛ and θE゛, for each signal) Therefore, it is helpful to optimize the values of decoder parameters for the improved decoder. It is a useful means of Decoder parameters for 4×3 and 5×4 improved decoders Such a theoretical calculation of the local parameters changes the values of the decoder that preserves them. and a wide variety of program materials created by a variety of recording and mixing techniques. is best determined by a combination of subjective tests of generally only a few Changes from degree-preserving decoder parameters are required for 4x3 and 5x4 cases. I know that.

A composite improved decoder is constructed by connecting two improved decoders in cascade or This can be implemented by connecting a storage decoder after the decoder. Such complex A combination decoder is designed to achieve the same result as a cascaded decoder by methods known to those skilled in the art. It can be implemented as a single decoder designed to

All subjectively desirable matrix regeneration decoding methods according to the present invention are left/right style or or direct loudspeaker feed mode, for example. at least a few octas including an intermediate frequency range from 500) 1z to 3k) lz Over the curve, there is a dominant matrix with some matrix coefficients larger than others. It has a polarity opposite to that of the xx coefficient. Such a reverse polarity auxiliary mat The Rix coefficient helps stabilize the image and make different auditory localization mechanisms more consistent. I have the effect of doing that. In preferred cases, coefficients with substantially opposite polarity The magnitude of is less than or equal to 215 of the magnitude of the dominant matrix coefficient.

For the improved decoder, the parameters φ and φ,2 are preferably within 25°. , parameters φ3, φ0, φ4, φ5, and vector parameters a, b, c) are preferably are within 15° of their storage decoder values shown above.

transmission hierarchy According to the flowchart in Figure 16, first set the parameters in the case of 3×2.4×3 and 5×4. The transmission map according to the present invention is constructed from an energy conserving regeneration matrix decoder with The obvious equations for the trix encoder and transmission matrix decoder are show. In this case, another channel coefficient of Dn+I n+1 is assumed to be of unit length. and Dn+1 and Dn+1 to be orthogonal to the other n columns. This lowers The equations for orthogonal matrix transmission decoding and encoding expressed in the MS form as shown below are obtained. It will be done.

Single note C,=M, M=C.

2 speaker stereo M2=M, 52=S (decoding) M=M2.5=32 (sign code) 3 speaker stereo For decryption For encoding 4 speaker stereo For encoding 5 speaker stereo For decryption However, the 3×3 matrix A is given as follows.

Here λ = (a 2 + 1) 2) I / 2, μ, = b / λ, μ2 = a / λ , LJ 1 = ac/λ, LJ 2 = bc/λ, a 24 bQ4c2 For encoding Here, AT is the transpose (aji) of the matrix A=(aij).

These transmission coding and Preferred values for the decoding and decoding equations are substantially Therefore, φ'=45°. Therefore, preferably 32゛≦φ′≦55゜ 4°≦φ, ≦17° 22'≦φ0≦35゜ 45°≦φ, ≦58゛ 3゛≦φ5≦16' The vectors (a, b, c) are of unit length, and the directions within 6° (0, 6 164,0°6558.0.4359). Values within these angular limits are preferred. Although suitable, 15° or 25° of the central values given for these parameters Wider angular limits within 10° are within the scope of the present invention.

Very suitable values for these parameters are substantially φ°=45, φs”10 ．． 57°, φ, = 28.64', φ4 = 51.64°, φ5 = 9.64°, and (a, b, c) = (0,6164,0,6558,0,4359) be.

14 and 15 and transmission encoding and decoding matrix means according to the invention In the practical application of this hierarchy, the transmitted signals M, S, T% T4, and T5 have The amplitude level of the signal within each transmission channel is the peak level of that channel and For each given non-zero amplitude gain to match the noise characteristics, ′,k2 It is recognized that ゛ can be given to ゛, k3'', k41, and ゛. Another amplitude gain such as ki'-1 can be applied to each channel signal at the decoding stage.

The gain ki' can be positive or negative, or can have a complex value; can be frequency dependent if equalization of the transmission channel is required. one Generally, the transmission channel signals M, S, T.

Since T4 and T5 are of gradually decreasing average signal energy, the associated The channel gain ki can be selected to be a gradually increasing value. is commonly found.

Transmission coding expressed in left/right form for the highly favorable parameter values mentioned above. and the decoding equation has the explicit values shown below.

Single note CI = 2-1/2 (L + R), L = R = 2-1/2C.

2 speaker stereo L2=L, R2=R: L=L2, R=R23 speakers stereo For transmission matrix decoding For transmission matrix encoding 4 speaker stereo For transmission matrix decoding For transmission matrix encoding 5 speaker stereo For transmission matrix decoding For transmission matrix encoding Using the above equation, the encoding from the loudspeaker-supplied signal of the first plurality n, and decoding to a second plurality n2 not smaller than nl is the input signal for nj = n2 and decoder parameters φ゛, φ3, φD, φ4, φ1, and (a, b If a very suitable value of %C) is used as in the numerical equation above, Figures 1b- A conservative decoder (if nl > 2) or an improved decoder ( An exemplary reference value of θp is given for nl=nz).

For other loudspeaker arrangements, such as those with other values of θp, these equations are more accurate. may not show the effectiveness of an optimal conservative or improved decoder) Note that it is always close. However, it is difficult to optimally adapt the decoding result to a particular loudspeaker configuration. LGf, code (matched to the loudspeaker configuration actually used, not to the parameter) (Rameters φ' (for 3-speaker reproduction), φ and φD (for 4-speaker reproduction) φ4, φ5, and (a, b, c) (for 5 speaker playback) Based on the modified values of the transmission matrix (summer) can be used. Such Since the modified transmission decoding matrix number is orthogonal, such a decoder still The energy conservation results are shown as . Such modified transmission maps are suitable for other loudspeaker configurations. The Trix decoder is within the scope of the present invention and is compatible with the transmission decoding map for which it was originally intended. Since it is very close to the trick, when nl = n2, it is essentially Doi's loudspeaker - child. The feed signal is saved. Instead, a transmission matrix with the same parameters as the decoder The regeneration matrix preserving decoder or improved decoder according to the present invention was later added to the decoder. can be connected. Transmission matrix decoder and subsequent (regeneration matrix) The matrix decoder can be assembled into a single matrix means by methods obvious to those skilled in the art. I can do that.

Other possible non-orthogonal transmission matrix encoders and transmissions for use in hierarchical systems The matrix decoder preserves or improves the reproduction decoder parameters φ′, φ3, φ can be configured for the desired values of φ, φ5, and (a%b,c), can be designed using the design procedure described in connection with the flowchart in Figure 16, and within the scope of the invention.

n speaker signals transmitted by m channels greater than n according to the equation If we use a transmission decoder that uses the supply, an additional non-zero transmission signal Tn+1.・ ...T■, but this requires that the fundamental signals TI, ...Tn remain substantially unchanged. For example, psychoacoustics provided to listeners using more than m transmission channels. To improve the analytical results, the frequency-dependent matrix of n input channels is Using additional channel signal Tn+1. ...prevents synthesis of Tl11 It's not a thing. - An example has already been described in connection with FIG. 8 for the case n=2.

In the more general case, a frequency-dependent n×m transmission matrix encoder can supply the output of an existing mXn regeneration matrix improved decoder, or equivalent A frequency-dependent matrix code to achieve the effect of such a composite encoder It can be supplied by means of conversion.

4X4 or 5X5 matrix as described above when T4 and T5 are set to 0 The explicit 4x3 or 5x3 transmission transmission matrix force equation obtained from the equation is When a 4×2 or 5×2 frequency-dependent improved regenerative decoder is required, the decoder shown in FIG. It can be used as an output means (9).

Delay compensation The design theory presented above is that all loudspeakers are directed in the same direction from the listener in the ideal position. The following loudspeaker configuration is assumed. The present invention can be applied to the configuration of Figure 1g or the n-speaker configuration. preferably along a straight line or along a non-circular path or at its center. This equidistant requirement holds, such as those placed along a circular path that are not within the listening area. Can be used for loudspeaker configurations that are not equipped. The result in such cases is generally Although less satisfactory than the equidistant loudspeaker configuration, it is still generally acceptable.

However, for optimal results, these loudspeakers should be set to The signal supply (94) is closest to the take position (4), and its time delay is as shown in Fig. 1g. The arrival of sound from that loudspeaker and the farthest loudspeaker as shown in Figure 26 for the arrangement The matrix reconstruction is performed by a time delay means (93) equal to the time difference between the arrival of the sound from It is desirable to obtain it from No. 2 (2).

In general, matrix decoders are used for all loudspeakers or at preferred listening positions. provide time delay means for all loudspeakers except the one furthest from the Any loudspeaker supply that may be incorporated or used in conjunction with it The time delay applied to the signal delays the impulses passing through the decoder to the preferred listening position. Arrival times are virtually guaranteed to be the same for all loudspeakers.

Such delay compensation means include analog charge-coupled delay lines and digital delay techniques. can be provided using any available time delay technique, including. digi The installation of digital delay compensation is a matrix that is carried out using digital signal processing technology. It is particularly simple for the decoder means.

In a preferred embodiment of the delay compensated fJf signal generator according to the present invention, the intended loudspeaker structure The formation is substantially symmetrical, making it suitable for a! It is on the axis of symmetry. In this application, Delay compensation does not attempt to compensate for listening positions that are far from the preferred position; compensation for the actual loudspeaker configuration and the large listening area in which listeners may be present. I am simply thinking about general relationships.

low frequency correction The ear hears lower frequencies below about 200Hz, especially below 100Hz, and higher frequencies. less locally sensitive at lower frequencies. Therefore, the matrix decoder according to the invention may deviate from the strict requirements of the present invention at such low frequencies.

When used with loudspeakers, some of which have limited bass reproduction capabilities, the present invention The decoder by φ゛ or φ at low frequencies to redistribute bass energy between the various loudspeakers. , φ3, φD, φ4, φ5, and modified matrix reconstructions such as (a, b, c). parameter can be incorporated. Such a decoder can detect the low pitch of different loudspeakers. Incorporating or used in conjunction with phase compensation means to attempt to compensate for differences in phase response. The low frequency localized queue is kept as large as possible.

Portable multi-speaker stereo system Widely popular forms of stereophonic sound equipment include cassette tape recorders, radio reception means, and a signal source (1), such as a compact disc player, amplification and control. In everyday language, it incorporates control means and loudspeakers into one unit. It is called a blowing instrument J. - It is a solid-state portable device. This type of device is especially suitable for a pair of removal It is equipped with a loudspeaker unit that can be mounted to allow this device to be used for even wider stereo applications. For best effect, attach the loudspeaker to or from the main housing unit or The loudspeaker can be separated and used for playback.

According to one aspect of the invention, each frequency range includes a range from 400 Hz to 5 k) Iz. – stereophonic sound reproduction using at least three loudspeakers covering the audio frequency range. A portable or movable system is provided, which is a single unit It can be carried as a stereo sound source signal, and can be carried as a stereo sound source signal. trix decoding means for generating a supply signal for said loudspeaker system; and thereby at least one said loudspeaker system is connected to said portable or movable system. A gasket that is securely attached to or integrated into the main housing unit of the stem. , and thereby two of the other loudspeaker systems provided in the main housing. can be mounted very close to each other and the main housing. said main housing unit so that it can be used at a distance from the unit. movable or removable relative to the system, said two different said When a loudspeaker system is installed in the main housing unit, stereophonic sound reproduction It is desirable that the arrangement be such that it can also be used for Also the matric It is also preferred that the step decoding means be in accordance with the invention as previously described.

A device of this type preferably comprises a signal source or receiver integrated into said main housing unit. an optional width tone responsive to a two-channel stereo source signal that may be provided by the means; 3x2 or 4x2 matrix decoding means of the type previously described with node means. We are prepared.

FIG. 27 shows, as an example, the above-mentioned type of portable device for multi-speaker stereo reproduction. .

The main housing unit (81) includes a cassette player (11), a radio receiver ( lj) and/or a signal source such as a compact disc player (1k), Controls (82) such as volume, equalization, width and source selection knobs; and preferably a central loudspeaker installed at the center of the front of the housing unit (81). The main housing unit (81) also includes the main housing unit (81). Amplifying means incorporated inside the unit (81) or inside the loudspeaker envelope (85) (not shown), all of which have a main frequency range of 400Hz to 5k) Central loudspeaker (52) and left (51) and right covering a frequency range including z 3×2 or 3×3 responsive to a stereo signal source feeding the loudspeaker of (53) Matrix decoder means (2) or (9) (not shown) are also provided.

The left (51) and right (53) loudspeaker envelopes (85) are connected to the main housing unit. Although shown attached to the main housing unit (81), the main housing unit (81) and Audio signal cables can be removed from each other and installed spaced apart. (84) or other audio signal communication such as an audio infrared link. remains connected by communication means.

The left and right loudspeaker envelopes (85) are fitted with catches (83), clips, hooks, and bases. Lucrow or other fastening or attachment means to said main housing unit. It can be attached and then removed. Alternatively or additionally, an enclosure The device (85) slides or otherwise attaches to the main housing unit (81). is movable (e.g. by rotation or pantograph action) and has an arm or link. directly connected to said main housing unit (B1) while still being physically connected by a A movable arm or link (as shown) that allows it to be moved from near to far. (without).

The loudspeaker envelope (85) can be installed in the immediate vicinity of the main housing unit (81). The advantage of using a movable arm or linkage means that can be removed from the loudspeaker unit is that this means Accurately control the relative position of the knits (51-53) to ensure the best stereophonic sound effect. However, inexperienced users can nevertheless use a totally removable loudspeaker envelope (8 5b) may be installed in an undesirable location. movable arm or The link also connects the entire unit so that the loudspeaker envelope (85) is connected to said main housing unit. (81) while being removed from the immediate vicinity of the main housing unit. One carrying handle (86) or shoulder strap attached to the belt can be carried by a belt. do. Instead of having a three-loudspeaker system covering the main frequency ranges, The outer set is placed inside the movable loudspeaker envelope (85), and the inner set is placed inside the movable loudspeaker envelope (85). 4 x 2.4 x 3, or can be provided for use in conjunction with a 4×4 matrix decoding means.

installation of other or alternative stereo sources, removability of the central loudspeaker unit, and details such as replacement or supplementation of the control means (82) by means of a remote control unit. Minor variations will be apparent to those skilled in the art.

Generally, different loudspeaker systems ( 51゜53) have different frequency response characteristics and/or phase response characteristics in the main housing. equalizer compensation means, which may be included in the is incorporated into the device and used in conjunction with the matrix decoding means to determine the characteristics of the loudspeaker. Said difference can be compensated for. In particular, said matrix decoder means are limited to In order to minimize the amount of bass energy supplied to loudspeaker systems with high bass capabilities, frequency-dependent matrix parameters can be used. For example, center The loudspeaker system (52) has more bass power than the mobile loudspeaker system (51, 53). - output, and the 3x2 matrix decoder according to the invention has a low bass output. Using a decoder parameter φ that can be reduced to a value close to Oo in frequency be able to. Similarly, the 4x2 matrix decoder according to the invention The decoder parameter φ4 is reduced to a value close to Oo in number. can be used.

Use with related visual images A particular application of the invention is playback with associated visual images, in which case It is necessary to align the direction of the sound with that of the relevant visual images for listeners over a wide viewing range. There is a point. Visual images can be physical, such as in a theater or in a live musical performance. The invention is particularly applicable to situations in which objects present in Television broadcasting, video recording, film projection, or computer graphics digital signal storage or processing means, such as a game console or electronic game machine. It can be applied to reproduced video obtained from

For use with visual reproduction means in which the direction of the visual image and associated directional sound are substantially aligned. In a preferred form of the invention for use, a display screen or projection hand is provided within the main housing unit. Visual reproduction means such as steps are provided, said housing unit comprising at least Both encompass the primary audio frequency range of 400Hz to 5KHz, and each covering at least said frequency range and from two sides of said main housing unit; be movable so as to be placed apart from each other and be disposed on the two sides; at least one loudspeaker used with at least two loudspeaker systems capable of comprising or securely attached to a system, further comprising: , responsive to a stereophonic signal associated with the visual image and played by the loudspeaker system. matrix decoding means according to the foregoing description of the invention for generating a signal intended to be ing.

Said movable loudspeaker is attached to said main housing by means of mounting or fastening, if necessary. Can be attached to and removed from the unit and/or sliding , rotation, pantograph, or other action to move the movable loudspeaker system to the Any of said main housing units in close proximity to or at a distance from the main housing. A movable arm or link hand so that it can be placed and used on either side. A step may provide a physical connection to the main housing unit.

Figures 28 and 29 illustrate two examples of audiovisual devices according to this aspect of the invention. main The housing unit (81) has a display screen (87) or other display means or projector. one of which is installed in one of the main housings (81) and loudspeaker means spaced apart from each other, each covering at least said main frequency range. used with two loudspeaker envelopes (85) containing said main housing unit. Nih (81) is also one or more covering at least said main frequency ranges. It houses two loudspeaker systems (52). Main housing unit (81) responds to stereo signals and of such processing (mugi, said loudspeaker system (52 ) and (85). comprising or in conjunction with matrix decoding means (not shown) by or incorporate amplification means as necessary or desired. Figure 28 is 3 −2 central expansion in conjunction with ×2 or 3×3 matrix decoder means (not shown) The case where a loudspeaker (52) is used is shown, and the loudspeaker is placed at the center with respect to the visual image. In order to ensure correct localization of the sound image, preferably said display screen (87) or Placed centrally below or above the means.

Figure 29 shows two loudspeaker systems (52) installed in the main housing unit (81). Either integrated or directly mounted on either side of said display screen (87), a 4X2. For use with 4x3 or 4x4 matrix decoder means (not shown) It shows.

When used with the matrix decoder according to the invention, the quality of the stereophonic image is between the inner loudspeaker systems (52) of the spacing between the side loudspeaker systems (85); Depending on the spacing ratio, the value of this ratio varies widely over a range between approximately 2 and 5. Re In addition to affecting the overall width of the live sound stage, other loudspeaker envelopes (85) may be wider or or even narrower spacing may reduce the acceptability of stereophonic images over a wide range of installations. Affect. Matrix decoder means, if required, ) to obtain the required width of the stereophonic sound stage in a given installation. Stages can be incorporated.

The audiovisual equipment is based on the frequency response or frequency response between the inner (52) and outer (85) loudspeaker systems. or equalization means can be incorporated to compensate for differences in phase response, and the matrix The encoder means separately or alternatively comprises low frequency modified decoding matrix parameters. , redistributes bass energy between loudspeakers to compensate for differences in their bass reproduction capabilities. You can consider it.

high fidelity equipment The invention may also be used, for example, for music reproduction not necessarily associated with visual images. Well suited for use with high quality high fidelity sound reproduction systems. like this In high definition applications, loudspeaker units are generally separated from each other and the amplifier control means. the loudspeaker unit is equipped with matrix decoding means according to the invention. or associated with a physically separate matrix decoder means device. and can be used.

Referring to FIG. 30, in accordance with another aspect of the present invention, the previously described matrices according to the present invention Preamplifier control responsive to a stereo source signal (1) incorporating Rix decoder means A means device (91) is provided, which device can be integrally integrated into the device if necessary. a sector in front of the preferred listening position (4), after subsequent amplification means (92) which can consisting of at least three loudspeaker systems located over the area (3); generates an output signal to be supplied to a stereophonic loudspeaker arrangement (50).

In a preferred form of this embodiment of the invention, said preamplifier control means device (91) is Receives, selects, and/or selects relevant visual images to be matched to the oriented playback image. It also includes visual signal control means for correcting or modifying the signal.

Referring to FIG. 31, another form of the present invention is shown in FIG. in response to the signal output (20) within the fan-shaped area (3) in front of the preferred listening position (4). A three-dimensional sound loudspeaker consisting of at least three loudspeaker systems arranged across the Generates an output signal to be supplied to the arrangement (50).

In a preferred form of this embodiment of the invention, said preamplifier control means device (91) is Receives, selects, and/or selects relevant visual images to be matched to the oriented playback image. It also includes visual signal control means for correcting or modifying the signal.

Referring to FIG. 31, another form of the present invention is shown in FIG. in response to the signal output (20) within the fan-shaped area (3) in front of the preferred listening position (4). A stereophonic loudspeaker configuration consisting of at least three loudspeaker systems arranged across the The present invention generates an output (40) which is supplied to the amplification means (29) which is supplied to (50). A matrix decoder means device (2) is provided.

public address device The present invention also provides statue security for larger sized audiences typically found in home use. Along with public address (PA) equipment that attempts to reproduce stereophonic sound with improved clarity. Also suitable for use. The P^ device is suitable for use in movies or movies, among other uses. or film audiences, live amplified music, and audiovisual and theatrical applications. can be done.

In P^ applications, to increase power output capability or to cover a wide directional listening area. , instead of one loudspeaker system sharing one envelope. It is common to use clusters of loudspeakers that are physically close together. Such a bundle is a bundle according to the present invention. It is understood that it constitutes a single "loudspeaker" as far as the application is concerned. In this document Terms like "loudspeaker" or "loudspeaker system" include such bundles of loudspeakers. You can In many P^ systems, different loudspeakers in a given bundle are can handle a range of frequencies. Place a bunch of loudspeakers vertically on top of each other Such a bundle is often referred to as a "stack" of loudspeakers.

Traditional stereophonic music composition or theater P^ equipment is usually a stage or performer. Use a pair of stacks or bundles on either side of the control area, sometimes with a third central bundle. located above or behind the center of the performance area. A bunch like this or The stack is fed by an amplification device, which contains pre-recorded sound, micro from various performers or their instruments picked up by telephone or electrical means. sounds, and sounds obtained from effect devices such as synthetic reverberation or reverberation units, such as 3D sound that allows you to control the level and 3D position of many separate sound sources A stereophonic signal obtained from the mixing water or the mixing device is provided.

Referring to FIG. 32 as an example, such a three-dimensional sound mixing device (1) is similar to that of comprising or supplying matrix decoder means (2) according to the description of The matrix decoding means (2) can perform the main listening performance or visual display area (87) encompassing the extent of the sector in front of the capture area; A system or bundle of three or more loudspeakers arranged above or around a stereophonic sound configuration. , or a stack (50).

332 has two loudspeaker stacks (5Z) and (53) a performance area (8 7) installed on the left and right sides respectively of the central loudspeaker system or loudspeaker bundle; (52) in order to avoid visual disturbance of the performance area ( 87) is suspended above the front.

Broadly the same kind of matrix decoding device (2) can be used as in other applications of the invention. However, certain features are desirable for such P^ applications. what kind The input and output sockets or connection means may also be of the XLR type or ten inches ( 6. By using an axle load jack connector, all products meet professional standards for heavy loads. Adjustments should be provided to address typical operating problems. For example, the matrix decoding means may be used in a central or internal loudspeaker system or (A delay compensating means for compensating the positioning of the distance of the distance is preferably incorporated, and should be used in connection with Also commonly used is a suspended central loudspeaker system. Stems or bundles ensure that large bass units do not visually obstruct the performance range. If the outside stack is too heavy or large to be hung in the A bunch can have a more limited bass capability.

The matrix decoding means (2) are therefore preferably supplied to such a central loudspeaker. For example, at low frequencies, φ42 should be very close to 90°. means for adjusting the low frequency decoder parameters to minimize the Should.

Preferably, such parameter modification has an effect on a bass transition frequency different from 4. It should be adjustable to suit your bass deficit. A matrix decoder like this (like Allows the user to adjust the values of the matrix decoding parameters over one or more frequency ranges Therefore, the decoding effect can be optimized for each P^ device hζ can.

It is also possible to use different loudspeaker systems or bundles with distinct loudspeaker types. t) for each frequency using a separate decoder for each frequency You can throw it away and use it. For example, put mouth, = 5, in a high-pitched loudspeaker system. 02 = 3 or 4 for the interfrequency loudspeaker unit and 0 for the bass loudspeaker unit 2-2, direct feed from two channels for bass loudspeaker, intermediate frequency 3X2 or 4X2 decoder for each loudspeaker and 5 tl for each treble loudspeaker. can be given using an X2 decoder. The input of the decoder on each side is the partial Typically used to generate the feed signal to a P^loudspeaker unit covering a range of frequencies can be obtained using an electronic crossover filter circuit of the type described.

in-car stereo The present invention relates to a stereo system used in a vehicle, particularly a car, such as an automobile. solve specific problems related to the system. In a car like this, you can play stereo. , due to the inevitable limitations on the location of both the loudspeaker and the listener) Gives particularly poor directed positive hallucinations. For example, drivers may Stereo loudspeakers are generally placed on each side of the front of the interior of the vehicle. Do not install it inside the doors on either side of the front seat area or on either side of the front seat area. Nara l, N. This arrangement is far from the ideal arrangement for a good stereoscopic image. l).

The present invention can provide much better stereo image quality fJt. See Figure 33. In contrast, a third central loudspeaker (52) is provided to replace the traditionally installed standard loudspeaker (52). Complementing the conventional left and right loudspeakers (51) and (53), said central loudspeaker The voice box is typically mounted in the center, above or below the vehicle instrument panel. typical The left (51) and right (53) loudspeakers are located on both sides of the instrument panel or on the vehicle. One size can be installed inside each front door.

When such a loudspeaker is used with a 3X2 matrix decoder according to the invention, The stability of the central image is maintained even when the driver is very close to the loudspeaker configuration and inside it. , especially for frequency-dependent decoders (5k) for frequencies above Iz and low 1 above 400Hz ．． When used with larger decoder parameters φ or φ4□ at N frequencies It has also been found that there is a significant improvement in

The present invention also uses an n2xnl matrix decoder according to the present invention to It is also possible to use two or three further loudspeakers between the pair of loudspeakers.

different loudspeakers by incorporating or adding equalization means associated with each of the loudspeaker systems; The system's different frequency response or the sound from each loudspeaker differs in its path to the listener. Both the typical absorption or refraction properties experienced can be compensated for.

Generally, the invention is installed in the front and possibly on both sides of the front seat area of the vehicle; a stereophonic sound configuration consisting of three or more loudspeakers responsive to two or more stereo source signals; The delay compensation means can be used together with the delay compensation means of the stereophonic loudspeaker supply means. It can also be used in conjunction with each or several. In front of or behind the rear seating area According to the present invention, a second three-dimensional sound configuration installed in either of the above is additionally provided, It can serve listeners in the rear seat area.

In in-vehicle systems, listeners are in close proximity to the loudspeaker configuration, so for optimal results It is possible to adjust some of the machine parameters such as φ, φ4□, φ5, φD empirically. A good value is 15° of the "storage decoder" value described earlier, although this may be necessary. or within 25°.

Still other aspects Various modifications within the scope of the invention will be apparent to those skilled in the art from the above description. parable For example, all the matrix means already explained can be rearranged, combined, divided and separated, Recombined element means can be provided. Incorporates gain and polarity inversion , adding means with subtracting means at different points while preserving the performance of the entire matrix means. with full path means that can be replaced and affect all signal paths identically be able to. MS matrix means responsive to left/right format signals; Make it respond to MS-form signals by adding or subtracting, as applicable, and vice versa. The means for responding to a signal in the form of an MS may also be used. Similarly, left-right morphology or The means for generating one of the signals in the MS form includes the MS matrix means as appropriate. Other forms of signals can be generated by adding or subtracting from the . Already All means satisfying the matrix equations of knowledge can be designed by methods known to those skilled in the art. Substitution by other means that yields a result that satisfies the same matrix equation Can be done. In particular, a matrix consisting of two cascade-connected matrix means The means is a matrix of products of matrices that describes the left/right behavior of the element matrix means. can be replaced by a single matrix means described by the xx coefficients .

Aspects and examples of the invention described in terms of electrical analog signal processing means are substantially equivalently using digital processing means such as superior and conversely in a manner obvious to a person skilled in the art. Good〈Can be implemented.

When a loudspeaker or loudspeaker system is referred to, it is used to effectively act as a single loudspeaker. A bundle or system of loudspeaker units installed relatively close to each other. Can be used equally.

Use separate loudspeaker systems to encompass different parts of the audio frequency range provided by a suitable decoder according to the invention for that frequency range. Different complex loudspeakers can be used to reproduce each component frequency range.

Although the present invention has been described nominally in terms of front, left, and right directions, the present invention For example, in a sector behind the listener, on one side of the listener, or above or below the listener. , or similar to stereophonic loudspeakers encompassing the range of other sectors, such as vertical sectors. It can be applied to etc.

The invention also provides a rear loudspeaker such as a rear loudspeaker which also encompasses the rear fan area according to the invention. A loudspeaker encompassing a sector area used in conjunction with other loudspeakers in other directions , or a loudspeaker supplied with a delay variant or a reverberation variant fed into a front loudspeaker, can be applied to three-dimensional sound configurations. Some stereophonic signals are provided according to the present invention. The elements of the larger loudspeaker configuration are further processed to provide loudspeaker signals for stereophonic sound configurations. any other signal from another loudspeaker or any other source fed to the loudspeaker. This does not affect the scope of the invention. For example, in IDTV or movie applications: Transmitting and reproducing a separate "surround" signal to create the front stereo signal created by the present invention. can supplement digital stereo effects.

Transmission channel signals can be sent and received in either left/right form or MS form. Wear. This is also 2-1/2 (T211-1 +721.) and 2-”2 (T Transmission signal T2°-1 in the form of 211-t T211) and left/right form of T211 may include possible uses of.

Having described a specific embodiment of the invention that is bilaterally symmetrical, the invention has been described above. The decoder design method can also be applied to playback using asymmetric loudspeaker configurations. can be done.

The present invention also provides that the different loudspeakers in the configuration are at different heights or at different angles of elevation or depression. The volume of a loudspeaker encompassing the extent of a sector in front of the listener, which may be present at the corners It can also be applied to acoustic configurations.

Examples of matrix regeneration decoders according to the invention mainly pass them through Preserves the total energy of a signal accurately (into a proportionality constant that can be frequency dependent) The limited and only However, deviations to a lesser extent can be tolerated. Psychoacoustic advantages of the present invention The tolerable degree of deviation that substantially holds The gain of the two stereophonic sound signal components passing through the reproducing decoder does not exceed 3 dB. preferably differ by less than 2 dB, very preferably by less than 1 dB. , expressed in terms of their influence on the direct loudspeaker group supply signal, are some of the decoders The matrix coefficients remain dominant over several octaves of the audio frequency range. or of the largest matrix coefficient of substantially opposite polarity and magnitude less than 275. It is like being a thing.

Such small deviations from exact energy conservation up to the interior of the proportionality constant are typical Specifically, matrix A means (33g) and matrix 38 means (33g) in FIG. 3h). Matrix A means and matrix 38 means at each frequency, for example for the purpose of electronic width control or other desired effects. Pass through Matrix A means (33g) or Matrix 38 means (33h) A different signal of signal (28) or (29) giving rise to signal (48) or (49) component does not exceed 3 dB, preferably less than 2 dB, I# always preferably less than 1 dB can be adjusted to have a relative total energy gain difference of .

For example, matrix A means (33g) can be energy conserving; Trick 38 means (33h) has an additional overall gain of between -3dB and +3dB. or the signals (28) and (29) may be are given different gains within 3 dB of each other, or the signals (48) and (4 9), matrix A means (33g) and matrix 38 means (33h) When energy is conserved, these gain modifications can be applied to the entire matrix decoder in Substantially opposite of some matrix coefficients to the dominant matrix coefficients of the field such that the polarity is maintained and the coefficients of substantially opposite polarity are the predominant coefficients. If the matrix coefficients have magnitudes less than 215, they are probably more than 3 dB apart from each other. Different gains are given within. In such a case, the decoder according to the invention That will happen.

For much larger deviations from exact energy conservation, the desired subjective effect of the present invention and has been found to substantially deteriorate the psycho-acoustic effects.

The above explanation is based on the MS matrix convention Mp=2-1/2(Lp+Rp) and 5p =2-172(Lp-Rp), but Mp=2"2(Lp+Rp) or and Sp=2-”2(Lp-Rp) or Mp=k(Lp+Rp) and 5p =ks(Lp-Rp) may be used in the practice of the present invention. can be used. However, k and ks are constants other than O.

The invention also provides, for example, 3 This also applies to speaker stereo equipment. All ampicnic methods are NRDC Patents GB2073556, GB1550627, GB1494 assigned to 752, GB 1494751 and is claimed in the present invention. 's paper "^l1biosonics in Multichannel Br oadcasting and VideoJ, pp, 851-871, J, ^ audio Eng, Soc, Vol 33 no, 1P (198 5 Nov, ). This aspect is not limited to equations, but also other equations. It can also be applied to picnic style.

In the above example with reference to Figure 8, the authors assume that a two-channel stereo signal in MS form Decoders for M2 and S2 have been described. where the signals M, S, and T is equal to about 35.26' below 5kHz and about 54.70' above about 5kHz. M=M2cos(φ−45°) for parameter φ equal to ° S=82 T=H2sin(φ-45°) is obtained as (21), where three speakers are provided. Salary is the equation % formula % given by.

The matrix in equation (21) is energy conserving as φ changes, and in Fig. 8 Equation (22), shown as (9), is orthogonal, so it also represents the total energy save.

The same method applies to each sound from the azimuth angle θ counterclockwise from the front. B-style signal W with gains of 1.21/2 cos θ and 21/2 sin θ, It can be extended to obtain an optimal three-speaker decoder for X, y. rear If we impose the constraint that the sound in the azimuth direction produces a sound that is no stronger than the sound in the forward azimuth direction, we get − From the equation, the best Signal M% supplied to matrix (9) of equation (22) giving calculation local quality S, and T are essentially M=0.41421 (W+X) S = 0.58579Y T=0.41421(W-X) (23)

This has a uniform energy gain as the azimuth changes (equation (23) Since the matrix is 0.58579 times the orthogonal matrix, the properties of ) and -) Optimum 3 The property of reproducing sound at the same azimuth angle O゛ as the teleo decoder, and 5k) lz The sound at the same azimuth angle of 60° as the left and right sounds played by the optimal 3×2 decoder on the It also has the property of being alive. The optimal 3x2 decoder is It handles center sounds best, and above 5kHz it handles stage edge sounds best. Since the B-style 3 speaker decoder in equation (23) is designed as Frequency dependent, but over the entire frequency range across the forward stage azimuth. Raku is said to be able to produce a stereo effect that produces unpleasant sounds in front of the rear azimuth. Manage to achieve these two optimal behaviors at a cost.

However, by introducing a frequency dependence similar to that in equation (21), 5k Optimal center image below Hz, optimal stage edge image above 5kHz It is possible to further improve the subjective results of 3-speaker B-style reproduction by Ru. This is a flea, frequency dependent on the M and T signals in equation (23) by an angle φ-45゜. Perform rotation, Mdec'=0.41421[(W+X)cos(φ-45' ) - (W-X) sin (φ-45°)]5dec' = 0.58579Y Tdec’=0.41421 [(W-X)cos(φ-45°)+(W+X) sin(φ−45°)] Modified M, S, and T signals given by (24) What to do by giving Mdec', 5dec', and Tdec' Can be done. where, as before, φ typically ranges from about 35° below 5kHz to 5 It varies up to about 55' above kHz. Precise changes in φ with frequency affect image quality The selection can be made through a subjective test.

If we use frequency dependent rotation as in Fig. 34a and equation (24), then 0 5k of values of γ9, γ8, θ9, and θ2 for different encoded azimuths close to ° As can be shown by calculations for matrices below Hz, the step The effect of providing a sharp and stable image near the center of the rheostage is typically achieved. On the other hand, above 5 kHz, the sense of stage width and the sharpness of stage edges improved. It will be done. Since the rotation matrix conserves energy, the resulting three-speaker supply is Maintain a constant regeneration energy gain with changes in azimuth.

Sounds encoded at an azimuthal angle close to 180° can be decoded using equations (23) and (24). Depending on the device, the sound may be reproduced with unpleasant stereo quality, so if you It is desirable to reduce by 3 or 6 dB. This attenuates the W-X signal (this (somewhat degrades the quality of the stereo localization across the front stage), or As shown in the figure, a forward dominant B format converter is installed in front of the 3-speaker decoder to reduce the This is done by reducing the contribution of the rear stage sound, as explained in equation (25). Can be done.

Obviously, the composite algorithm shown in the figure has its matrix coefficients given by the figure. placed by any frequency-dependent matrix algorithm equal to that It can be replaced.

The multiplexer described above for three-speaker stereo has n It can be generalized to speaker decoders. In the limit, the attenuator reduces Tdec to 0 This is similar to the nX2 matrix converter described above as the become equivalent. The final matrix Dn,3 is again an nX3 transform of the form explained above. This is the number matrix. Here, from the input matrix (Equation (23)) The signal is Mdec, 5dec.

It is called Tdec and the signal entering the output matrix is called M%S%T. Mato Rix is realized by band division in a manner similar to that of FIG.

The rotation matrix is implemented in a manner similar to that described with respect to FIG. 8 above.

The functions of the filter means 38 and the entire bus 38, the matrix of which is shown in Figure 34c. a and gain 38b are the same as in FIG. 8, and the element marked 38 is 38e. 38c is the same as 38a, and 38d is the same as 38b. this is the ideal rotation map for values of φ close to 45°, e.g. 35° or 55°. It can be shown that it is a very close approximation to the trix.

Besides applying to modalities with different numbers of speakers, this aspect also applies to omnidirectional signals W.

A signal X with a gain proportional to the cosine of the direction, and a gain proportional to the limit of the direction directionally encoded 360° surround that is a linear combination of signals Y with It can also be used with sound signals. Furthermore, these elements may be e.g. It may have been converted by orientation-dominant conversion. The inventor says ``forward-looking superiority'' One particular Lorentz transformation, named the ``transformation'' and specified in detail below, is the forward It has the effect of increasing the sound gain by λ times and changing the backward sound gain to l/λ. ing.

W’=+(λ+λ”) W+8-”2(λ-λ-1)xX’=+(λ+　λ-’) X+γI/2(λ-;A-')WY'=Y (25) This produces the Lorentzian transform signal components W′, Xo, Y′ that satisfy the above equations. where λ is a real parameter with a predetermined positive value. The conversion component is It still satisfies the characteristic relationship between the B-mode signals w, x, y, but the gain and azimuth The orientation is different from that of the untransformed component.

From the above relationship, the genuine ones with w, x, y gains of 1.2I/2 and 0, respectively. The surface B mode before is converted to one with a gain that is λ times larger, but the original gain is A true rear sound with gains 1, -2-1/2, and O is a rear sound with gain λ-1. It is derived that it is converted into Therefore, this forward dominant conversion reduces the forward sound gain. It increases by λ times, but by changing the rear sound gain to 1/λ, the front sound's relative to the rear sound increases. The relative gain changes by a factor of λ2, reducing the relative gain of the rear sound reproduction and reducing its relative contribution. It can be modified to increase (or increase) The master specified in the related claim 34a. and the matrix shown in Figure 34b. It can also be realized by a combination of matrices.

FIG. 16 shows the design process for the transmission layer. In this process, the last Columns can be freely selected to be linearly independent from other columns, and this selection The corresponding coefficients of the encoding matrix are determined. Rather than using immutable values , the value of the encoder can perform the inverse of the encoding function at a given moment when the decoder becomes If the selection is conveyed to the decoder as a side chain signal, it can change from moment to moment. Ru.

Error noise artifacts, such as those introduced by data compression, change the encoding equation Adaptively modified to match the instantaneous distribution of signal energy between speaker-supplied signals However, it can be subjectively minimized using the inverse equation for the Z1 unit. good A suitable strategy is to use a signal correlation matrix that more closely approximates the transmission channel than a fixed coding function. Adjust the coefficients to diagonalize the squares.

The process in Figure 16 utilizes transmission channels, but these channels are simply converted can be used as an aid to obtain appropriate values for the channel itself. need not exist explicitly as a predetermined configuration of the hierarchy.

The appendix below is φ, φ5, φ. , φ4, φ6, and vectors (a, b, c) Designed according to the flowchart in Figure 16 using the very suitable values listed before each parameter. In addition, the orthogonal transmission system whose transmission encoding and decoding matrices are shown above is The result of encoding from n8 speakers and decoding to 02 speakers using Based on the multi-speaker system, the number nJInspeaker is between 1 and 5 for stereo. 1 of the transformation matrix between the speaker-supplied signal and the 02 speaker-supplied signals. It lists two possible hierarchies. of the smaller number in the hierarchy listed in the appendix. The conversion matrix from the loudspeakers to the larger number of loudspeakers is as explained above. The preferred matrix regenerative decoder is The conversion matrix for the number of loudspeakers eliminates all frequency-dependent total path elements. Matrix of matrices from the smaller number to the larger number of divided states It is noted that it has a matrix that is a transpose. The following pages are further A general hierarchy constitutes a conversion between modalities with different numbers n of channels. Explain how it can be done.

In a cascade-connectable hierarchical transmission system configured according to the G16 method, the prescribed conversion When the input to a stage has fewer channels than the output of the stage, An up-conversion matrix as described earlier for n-speaker stereo is used. be done. If a small number of channels are output, the down conversion matrix is used. Transmission decoding and encoding matrix 70 songs and In the special and favorable case where and Enn are orthogonal, the down-transform matrix is It can be shown that it is the matrix transpose of the transformation matrix.

Generalization to other systems The above cascaded hierarchical method for reproducing stereophonic sound across a fan area is more general. It can also be applied to multi-channel directional sound encoding and playback systems. Ru. In the previous explanatory chapter, the hierarchical transmission and reception system of the directional sound reproduction system was explained. An example of such a system that describes the requirements for and applies to a stereophonic sound system A method for configuring the is shown in connection with FIGS. 14-16.

Not only does it feature multi-speaker stereo encoding and playback modes, but also sound reproduction, which also has various proposed surround sound modes. We describe a more general hierarchical method that can be applied to raw systems. This more general It is convenient to explain this method using mathematical notation, but this explanation The system described above, and other examples that will be presented later, are examples of this general method.

The directional sound encoding represented by the letter ^i for i=1.2,...,N With N required modes, the system has a number of nl audio channels. Furthermore, for each i%J from 1 to N, the system Ai has a code suitable for reproducing encoded oral audio signals from encoding system Aj. There exists a suitable njXni transformation matrix Rjl for converting into one audio signal. Let us call Rji an “up-conversion” matrix, and let Rji The linearly independent signals in system Aj are taken as the linearly independent signals in system Aj (this If ni≦nj), and only in that case, write ＾i≦＾j. Ku.

A collection of directional encoding systems ^i with 1==1 to N and each set of systems The following mathematical conditions (1) to (5) are satisfied for the collection of transformation matrices Rji between When added together, they are said to form a ``cascade-connectable hierarchy'' of the system.

(1) For i = 1 to N, Rli is n1Xni identity matrix 1i i, i.e. changing the system to itself does not change the signal.

(2) i and j are Rji invertible matrices (this includes n1=nj ), then it is Rlj and RlJ is Rjl is the matrix inversion of

(3) From l to N! , j, and k such that ^i≦^J and ^j≦Ak , then the associated up-transformation matrix satisfies the equation Pki = PkjPji. Add. , (this is called "cascade connection possibility of up-conversion matrices", and two This means that the cascade of up-transformation matrices of is also an up-transformation matrix. Taste. ) (4) Regarding the two systems Ai and ^J consisting of (i) Ak≦^ i and Ak≦＾j, and (i i) system Ah is ＾h≦Ai and ＾h If ≦Ai and Ah≦Aj, then Ah≦8. There are one or more systems 8k such that . (This condition is based on the third The existence of an ``up-conversion'' that relates two systems made up of ``small'' systems. , and there are one or more "maximum systems" of which they are both up-conversions It says that. ) (5) For the three systems Ai, Aj, and Ak, the system A consists of When h is Aha;Ai and Ah≦8, ^h≦Aj, and this When Pki=PkjPji.

This cascade connection possibility condition not only applies to the up-conversion matrix, but also It also applies to any three systems, with the intermediate system being connected to the two outer systems. is now an up-transformation of the ``maximum'' system, which is an up-transformation.

Conditions (1) to (5) are all applicable to up-conversion between n-speaker stereophonic signals. and downconversion for the systems described above, but also applies in other cases. be done.

A cascading hierarchy is a matrix of encoded sounds for one system ^i consisting of converted with satisfactory results from other systems ^J in the hierarchy by box means Rj1. Not only can it be broadcast, but it can also be done in long broadcast chains. The intended material for one system is transformed and then reaches the final user. The result of repeated conversions between different systems, such as when the Ensures satisfaction and all systems in the cascade chain are up-converted A single transformation down to the ``i&large'' system of Never produces a worse sound than the result obtained with a single upconversion up to the stem Therefore, it is desirable. A cascading non-connectable hierarchy as proposed in the prior art is an iterative transformation. The playback directivity effect is continuously degraded as the process is performed.

Using a cascading hierarchy of systems in this way means that any user directionally encoded sounds, whatever their history and initial source within the sound chain. Doing so may unduly degrade the results to other directional acoustic encoding modes in the hierarchy. It means being able to change knowing that you will not be changed.

Although it is desirable to clarify the hierarchy that can be connected in cascade, the conventional technology does not allow for designing it. It was not clear how to do so. In general, changes that meet requirements (1) to (4) above A sophisticated encoding system virtually preserves the originally intended directional effect. All you need to know is the transformation matrix. As in the case of stereophonic sound mentioned earlier By generalizing the method described in connection with FIGS. 14 to 16, A cascading hierarchy of directional coding systems starting from only knowledge of the transformation matrix It is possible to design.

The design method consists of n1 encoded system signals Ai for each i from 1 to n, With the reversible n1Xni transmission coding matrix (7) Eii, nl transmission signals are is encoded into a collection Zi, and nl codes are encoded from the transmission signal (6o) on the mouth side in Zi. The system ^i is decoded into the inverse n1Xni transmission decoding matrix Dii (9), and Fig. As shown in XI, it is based on decoding so that Eii=Dth-1.

Such transmission signals are related to each other for different i and j as shown in Figure x2. Must be. In this case Ai≦^ with up-conversion matrix RJi For J, EjjRjj from n1 transmission signals Zi to nl transmission signals Zj =　IjiEi　i A matrix mapping ■ji such as One of the transmission channel signals has a shape of nl. this is easier The transmission channels for more sophisticated systems are a subset of those for more sophisticated systems. This is an extension of the ideas expressed in connection with FIGS. 14 and 15.

If we use Dii=Eii-', Rj iD i i = Djj lj i is obtained, and from this it exists in Zi The 11 columns of the transmission decoding matrix Dii corresponding to the transmission signal are njXni matrices. It is derived that it has the form of Rix DjiDii.

The remaining nl-ni columns of Djj are linearly independent of each other in order for Djj to be reversible. and must be in the DjiDii column.

Therefore, by analogy with the stereophonic hierarchical flowchart in Fig. 16, for the system The transmission decoding matrix DJj is, for all systems ^i≦^J, Djj corresponding to the transmission channel in Zi which is in the transmission channel in Zi The ni columns of are chosen to be equal to the njXni matrix RjiDii. and the remaining columns must be linearly independent of each other and of the other 01 columns. This is done and the sign The matrix Ejj is equal to Di i-', and then the Zi in Zj is Taking the internal transmission channel to itself, the transmission in Zi that is not in Zj Using the matrix Iji that takes the channel to 0, we can get the system from system ^1. Transformation matrix Rji to stem ^j = DjjljiEii has all such transformation matrices Rji that all ^j are columns It becomes possible to show that a connectable hierarchy is formed.

Therefore, in strict similarity to the configuration of FIG. 16, EjjRji=1jiEii A transmission channel that satisfies or equivalently satisfies the above conditions for the sequence of Djj The system that encodes and decodes the automatically encodes Ai and retains only the transmission channels in Zk, Decoding to Ak yields a matrix formed by the transformation matrix P A cascading possible hierarchy of oriented phonetic coding systems with ki is automatically defined.

Therefore, regardless of whether the transmitted signals in the collection zl are actually used or not, , such a signal for Ai≦＾j satisfying the above condition regarding the column of Dii It can be used to construct a cascade connectable hierarchy from the up-conversion matrix Rjl. I can do that.

As already mentioned, in the configuration related to FIG. 16, Ai is the i speaker stereo speaker. This is done in the special case of a front stage stereo signal, which is the signal you want to feed into the front stage stereo signal. It provides a hierarchy that can be connected in cascade.

However, other types of sound reproduction systems may be added to the above front stage stereo hierarchy. allows flexible conversion between all playback modes or directional encoding modes. However, it still allows for various forms of surround sound and unphonic sound reproduction. A more flexible cascading hierarchy can be created that allows for more flexible cascading.

These are written as HA, Sr, bSBT, and FT, which will be explained next with reference to Figure x3. An example using five transmission channels is shown below.・Next directional encoding mode Think about de.

Carrying a single tone signal C1, 2 speaker stereo signal and carries R2.

3-speaker stereo All signal to the front stage first, 9, C5, and Carrying R3, etc. 2:1 stereo front stage 2 channel stereo signal L2F and R2F1 and carries the rear stage single tone signal C1B=B.

3:1 stereo front stage 3 channel stereo signals L3F, C3F, and and R2F, and rear stage single tone signals C, ,=B.

3:2 stereo Front stage 3 channel stereo signal, 2', C5,', R52'', and rear stage 2-channel stereo signal L2. and R2l1 carry.

These systems encode front and rear stage stereo directivity for HDTV and has been widely proposed for use with movie sound.

B-style ampic encoding 360” Three signals carrying horizontal direction self W%X, and Y, respectively − gain 1.2”2cosφ, and 2””sinφ encode the sound from the azimuth angle φ with BEF style enhanced ambiosonic encoding 5 signals carrying 360° horizontal azimuthal sound Carry the numbers W, X, Y, E, F, and each gain W:1 X: 2””cosφ Y: 2”sinφ E: l(, [l-to, (1-cosφ)] 0 if 1φ1≦φ8 1φ1 〉For φ8 F: 21'2krsinφ 1φ1≦φ8-2I/2, sinφ118 0 if 0°-φ1≦φ, otherwise 0 encode the sound from an azimuth angle with . Here φ8 is typically 60° and 7 is the predetermined forward encoding stage half-width between 0° and φ, typically 60 and 70°, and ka is 3 and 3172. is an invariant gain chosen from a value between (the preferred value is 3.25), and the gain is 2, and ks are selected by the user to be greater than or equal to 0 and less than or equal to 1. and typically equals 1 for a tone of azimuth O゛, typically km and kr can have values approximately equal to about ten.

BE style unpacking This is the four signals W specified above for the BEF style. , X, Y, and E are used.

BE style unpacking This is the four signals W specified above for the BEF style. , X, Y, F.

The BEF style signal has improved stability of the forward stage image and Another aspect that allows for acoustic reproduction in which the separation is improved compared to reproduction from mode B. provide information. The BE-style signal only improves the stability of the forward image, and the BF''''-signal Modifies only the front/fourth stage separation.

For illustrative purposes related to the above description of cascading hierarchies and associated transmission systems, Single sound (^,), 2-speaker stereo (^2), 3-speaker stereo (A , ), 2:1 stereo (^, ), 3:1 stereo (^5), 3;2 stereo (A 6), B style (A7), BE style (^8), BF style (^,) and BEF style From ^1 to ^, to the above ten directional coding systems for (^,.). For up to Can be labeled.

The cascading hierarchy consists of encoding matrices whose matrix coefficients are the same as shown below. When a transmission system using the Rix Eii is configured, the Out of 10 directional encoding systems, one encoding is reproduced by the reproduction of the other. The five transmission channels HA, ST-Tr, B give satisfactory subjective results when The authors have discovered that it can be formed using T and FT.

2 speaker stereo E22 3 speaker stereo E33 B style Ann Picnic E77 3:1 stereo E55 3:2 stereo E66 BEF style enpinic EIO10 BE style enpicnic E,.

BF Style Anpicnic E99 For i=1 to 3, the three subsections Eil explained in connection with FIGS. 6 and 7. ・What is initially given as the preferred encoding matrix for the castereo layer is attracting attention. Also, the front for 2:1, 3:1, and 3:2 stereo The stereo stage signal also has My, S, in the same way as the front-only stereo signal. T, and one transmission channel, but the rear stage stereo signal is , if the rear stage sound is played back approximately 3 to 6 dB, it is called [surround sound]. Since forward stereo playback in front of the material is known to be best, these three transmissions are Note also that the transmitted channel is encoded with a reduced gain.

The BT transmission channel is intended to carry the dominant aft stage material and is is the difference signal across the front stage minus the difference signal across the rear stage. respond to things. These transmitted signals, including rear stage sound, are explained in connection with FIG. It does not exist in the front stage stereo hierarchy.

The decoding matrix Dii of this transmission layer is related to the computer or calculator. can be calculated from the above matrix using a matrix inverse program. This is easily given by the inverse matrix of Eil.

For the above 10 systems, the transformation matrix Rji from Ai to Aj is as follows. The signal of Ai is encoded into a transmission signal by Eii shown above, and the signal of Ai or Aj is encoded by Eii shown above. Put all transmitted signals not used by any equal to 0, then this It can be calculated by decoding the transmission signal from Aj to Aj by Djj=Ejj”. can. The transformation matrix obtained for 10 systems is based on condition (1) A cascade connectable matrix is created that satisfies (5), but in this case Rjx is always the transmission signal of Ai is also the transmission signal of Aj, this is the test in Fig. 3. This can be determined by the up-conversion matrix.

Moreover, the transformation matrix obtained in this way to satisfactorily reproduce the signal intended for or transformed by the system. - to convert between different directional encoding modes, e.g. is configured to play in one mode when it receives a signal that is Can be used directly for sound reproduction. Such a direct conversion is Available programs for recording, playback or transmission connected in other modes. Concerning the conversion of programs in a professional or studio environment, it is also possible to can be used without fear of possible undue degradation of directional quality due to

Instead, such a conversion can be applied to the signals M, Sr, T, 8, and F as described above. encoding and decoding matrices, which can, but need not be, This can be achieved using an intermediate transmission channel signal by Rix. and For example, the transmitted signal can be encoded with another non-zero gain, and the inverse of said gain has a sign the gain possibly being different for each transmitted signal; Or the required independent linear combination of Ma, St, Ta, Ba, and FT as intermediate transmission. and can be used.

Additionally, other directional acoustic encoding systems can be added to the cascaded hierarchy above if desired. It will be recognized that For example, a 4-speaker and 5-speaker front stage Teleo system ^1. and ^, 2 in the previous description in conjunction with FIGS. 14-16. Add and encode using another transmission signal T4 and T, as described in Chap. and these signals have the same matrix coefficients 4:1.4:2.5: 1, and 5; Incorporating front stage transmission signals for 2 stereo systems. Can also be done.

Alternatively or in addition, a 2:2 stereo system can be used to connect the two front stage stereos. Signal L2. Using ゛ and R3,°, the encoding matrix equation %formula% system A using two rear stage stereo signals, and R211. It can be added as 13.

The cascading layers are encoded by Eii and decoded by Djj as before. By constructing Djj=Eii-' and Rji by The system can be expanded.

A useful and subjectively acceptable hierarchy containing these directional coding systems is It is not limited to the exact values of the coefficients of the encoding matrix Eii, but slight changes in the coefficients are It is also recognized as acceptable or preferred. Transmission channel gain and/or modify the gain when incorporating rear sounds into H A and ST. correct and/or the coefficient may be below 0.05, 0.1, or 0.2. The change that maintains left-right symmetry by only a small amount of correction is the transformation matrix Rj It is possible for i to give a cascaded hierarchy a subjective effect that is still acceptable. Wear.

The above ampinic directional coding system ^i with i=7 to 10 is described below. Although used for convenience, linearly independent combinations of the signals W, X, Y, E, and F The signal composed of , and the encoding matrix is It will be recognized that (Etl), law=(E++)otaci+ can be modified. . Here matrix C1i represents the new linear combination of BEF style signals in BEF style (or B style or BE depending on whether 1=10 or 7 or 8 or 8) (Ei i). 1d is above is the encoding matrix shown in , where (Eil)new is used with modification^i. This is the encoding matrix.

The following describes special modifications to the BEF and BF forms that are often desirable for professional applications. Explain. signal W'=W-E X'=X-2-I/2E Y Defines the reduced BEF format that constitutes the . where W, X, Y, E, and F are BE This is specified for Form F. Professional signal processing for reduced BEF mode Compared to the application BEF format, for a value l of 2 for the gain of E, the reduced signal W and X' are equal to 0 for the forward center of φ=o° It has the above advantages. Therefore, the sound that tries to sharpen the front center localization decreases like BEF. must be accurately mixed with the E signal of the equation and all other signals of the reduced BEF signal. The sign is equal to 0 at the position of the note.

The reduced BE mode is such that the four signals w', X', Y', and E from the reduced BEF mode are It is similarly described as consisting of.

For recording or mixing purposes, recorded single notes can be converted into BE, BF, BEF, You may want to place it at the azimuth angle φ in each mode of decreasing BE or decreasing BEF, This is a single tone signal, with four or five signal gain arrangements each of 1 for W. ．． Regarding X 21/2c.

2”2 sinφ for Sφ, Y, and for signal E, if 1φ1≦φ8. For ke[I-, (1-cosφ)], for 0 signal F when 1φ1>φ8. If 1φ1≦φ8, 21/2 krsinφ, l180'−φ1≦φ, -21/2ks sin φ in the case of , 0 in the case of other φ, and for the signal W', When 1φ1≦φ8, 1-kz [1-ko (1-cosφ)], 1φ1 × φ8 1 in the case, for x', if 1φ1≦φ8 2I72 [cosφ-ke (1-ko (1-cosφ))], 1φ1>φ8 This can be done by setting it to 2"2 cosφ in the case of Widths φ8 and φb and! l is as before, kg, Kr, and km are Positive user gain≦1, which can be adjusted at will.

Such a gain arrangement, in which the value of φ is manipulated by a user-adjustable control means, ``Panpot'' device or positioning device for these ampinic styles. Configure. B-style amplifiers containing significant signals that traverse only a limited sound stage. BE style, BEF style, OF style, decreased BB style, and decreased style from the picnic signal. It is also possible to create each signal in a low BEF manner by mixing. For example, B If the modal signals w2, XF, and Y2 are restricted to 1φ1×φ that azimuth Then, the signal W=death, X=XF, Y=YF, E=kt(Wr　ka(Wr　2 -"2XF), F=krYr, W'=ilr kg (6ka (6-2-1/2 XF)), X' = Xr 2172kg (Wrko(Np 2-"2XF)) is coded as a signal for 4-channel and 5-channel amplifier modes. of the B-mode signal, which is Regarding sound, W=Wa, X=b, Y=ym, E=O, F=-keYs, w'=w ,,X'=Xate. The B-style signal confined to the rear stage here is II IB, XI, and Y.

yields BE mode, BF mode, BEF mode, reduced BE mode, and reduced BEF mode. Due to the above description of the panpot means and mixing means, any output signal can be used. suitable for use with capable signal channels, or recording or transmission channels. Possibly polarity inversion to obtain an output signal with a level and/or polarity that It will be appreciated that predetermined non-O gains can be accommodated, including.

3, including all systems with prior art UMX layers and B-style encoding. Prior art surround sound system with directional encoding of sound at 60° azimuth angle n of directional encodings for each angle of rotation across the 360° sound stage. There exists a corresponding nXn matrix for the channel signals, and this matrix the coded signal for the same coding system. All encoded sound sources are rotated by the rotation angle within the 360° stage. It has mathematical rotational symmetry in the sense that it appears to be in a rotated position.

A hierarchical system for converting between such encoding systems with mathematical rotational symmetry Designing systems is known, for example in connection with 1MX technology, but this Until now, there was no known method for designing cascade-connectable hierarchies in the prior art; Some directional coding systems in technology, especially those that use more than two channels. , lacks rotational symmetry. The B style has mathematical rotational symmetry, but Fig. 3 All other systems of cascading hierarchies described in connection with 7 are mathematical rotations. There is no symmetry.

Annex Orthogonal transformation hierarchy The up-conversion matrix below is subjectively a very good performer, with a large number of To provide subjectively optimal preservation of the stereo effect originally intended by the speakers.

3×2 up conversion matrix R32 In this case, for the best subjective results, a frequency-dependent transformation matrix such as This includes the use of

Here, A is 5kl (with gain -1 below z and +1 gain above 5kHz). is the total bus circuit gain. If we set A=O, a reasonable frequency-independent up-conversion A matrix is obtained, but it is not very good in the case of frequency dependence.

4×3 up conversion matrix R43 5X4 up conversion matrix R54 Other up conversion matrices The other up-conversion matrices preferably connect the above three matrices in cascade. It is formed by This yields the “composite” up-conversion matrix below. It will be done.

4×2 up conversion matrix R42 Here, as before, A preferably has a gain of -1 below 5kHz and a gain of -1 below 5kHz. Above is the total path with gain +1, or in the frequency independent case A=O It is.

5X2 up conversion matrix R52 1-R6　(-0,1525+　0.0676^　0.6325　+　0.0 676A) Here, as before, A is 5k) with a gain of -1 below lz, and a gain of -1 below 5kHz. full bus with gain +1 on It is.

5×3 up conversion matrix R5S Down conversion matrix The down-conversion matrix for this case is set A=0 above and the matrix by performing a matrix transposition (i.e., turning rows into columns and vice versa). This “transposability” is special in the case of orthogonal hierarchies and is not generalized. Be careful. Therefore, Kanha et al. obtained the following down conversion matrix.

2×3 down conversion matrix R23 3x4 down conversion matrix R54 4x5 down conversion matrix RAS 2x4 down conversion matrix Ih4 2x5 down conversion matrix R15 3×5 down conversion matrix R3゜ Single note down conversion R1n (n = 2 to 5) Cs”0.7071 L2 + 0. 7071 R2C1=0.5000 L3 + 0.7071 Cs + 0. 5000 R3C, = 0.3998 L4 + 0.5832 L3 + 0.5 832 R5 + 0.3998 R.

C+=0.3394 L6+ 0.4786 L7+0.55790S+ 0.4786 R7 + 0.3394 Rb selection down/up conversion matrix Kusu 3→2→3 conversion R32R25 4→3→4 conversion R43R34 The above transformation matrix has compound device parameters φ, φ′, φ3, φ0, φ4, φ5, It is optimized according to specific values of (a, b, c). Regarding different speaker placements Slightly different values associated will give a different margin equation than above, but all For the case, the coefficients differ only slightly from those shown here.

I -FIG, 1c- -FIG, 1e- C3 -FIG, 1f- -FIG, 1g- Target on ζ♂ Retoto L/)≦ -FIG, 16- □Mouth G, 20□ LJ5 30 15 0 -15 -30 -1t5 Panpot angle θ -FIG, 2F-- Panpot angle θ -FIG-2'l- Panpot angle θ -F 1.3, 23- 45 30 75 0 -75 -30 -tJ5 Panpot angle θ -FIG, 2L+- 4530750-F5-30-45 Panpot angle θ -FIG, 2S- C r-111ko -FIG, 29- -FIG, 30- -　F'lG, 31　-□ treetop -FIG 33- Signal source/playback mode transmission options international search report

Claims (1)

[Claims] 1. When n1 and n2 are integers greater than 1 and n2>n1, n1 speeds a first audio signal (20) stereophonically encoded for reproduction by a speaker; and a second audio signal (40) stereophonically encoded for reproduction by n2 loudspeakers. In the matrix transformer Rn2,n1 for converting to A matrix converter characterized by: 2. The matrix converter is further adapted to substantially preserve within a second proportionality constant the reproduction angular configuration of the velocity vector, and further to store within a third proportionality constant the reproduction angular configuration of the sound intensity vector. The matrix transformer of claim 1, wherein the matrix transformer is adapted to substantially preserve playback angular configuration. 3. 3. The matrix converter of claim 2, wherein the ratio of the second proportionality constant to the third proportionality constant is between 1/2 and 2. 4. the loudspeaker supply signal represented by the first audio signal to the second audio signal; The matrix coefficients represented by the connection matrix relationship to the loudspeaker supply signal represented by the audio signal are applied over several octaves of the audio frequency range. some of the matrix coefficients are substantially of opposite polarity to the dominant or largest coefficient and have a magnitude of 2/5 of the dominant or largest coefficient; The two stereophonic sound signal components to be reproduced by the loudspeaker are The second number n2 loudspeakers differ by less than 3 dB by the xx converter. 4. The matrix transformation according to claim 1, 2 or 3, which is regenerated with an energy gain of exchanger. 5. 5. The matrix transformer of claim 4, wherein the two stereophonic signal components are reproduced with energy gains that differ by less than 2 dB. 8. 6. The matrix transformer of claim 5, wherein the two stereophonic signal components are reproduced with energy gains that differ by less than 1 decibel. 7. A matrix according to any of claims 2 to 6, wherein the eighth proportionality constant is configured to be greater in an audio frequency band above 5 kHz than in an audio frequency band of frequencies between 700 Hz and 3 kHz. converter. 8. A matrix converter according to any one of claims 2 to 7, wherein n1 is equal to 2 and the third proportionality constant varies with frequency. 9. Claims 1 to 8, wherein n1 is equal to 2 and means are provided for modifying the playback width having the effect of changing the gain of the signal component representing the difference component of the first audio signal. The matrix converter according to any of the above. 10. The first and second audio signals are stereoacoustically encoded to a loudspeaker configuration that is symmetrical on the left/right with respect to reflections around a fictitious forward direction, and the matrix transducer means that all of the left input and output signals are by the said matrix means when replaced with the symmetrically arranged right counterpart of Claim 1 is left/right symmetric in the sense that the result given by The matrix converter described in 9. 11. The matrix is configured such that the angular arrangement of the reproduction velocity vector of the second audio signal is substantially equal to the angular arrangement of the reproduction sound intensity vector of that signal at frequencies spanning several octaves of the audio frequency range. The matrix converter according to any one of Items 1 to 10. 12. a second symmetrical acoustic configuration in response to a first plurality of signals proportional to signals to be reproduced by a first plurality of loudspeakers arranged in a first symmetrical stereophonic configuration; matrix reproduction decoding means for generating a second plurality of signals proportional to a symbol to be reproduced by a second plurality of larger loudspeakers arranged in a matrix, the matrix decoding means The input sum and input difference maps for each pair of target signals are calculated for each pair of symmetrically arranged loudspeakers that make up the configuration. trix means and a central loudspeaker supply to all of said sum signals and said first arrangement; a first linear means or matrix means of a first output signal for providing a first number of signals not less than the number of signals in response to all of said first plurality of signals proportional to said first output signal; 1 linear means or matrix means and corresponding to all said difference signals. means for generating a second number of output difference signals not less than the number of difference signals, wherein the first number and the second number are added up to the second plurality; associated with each symmetrical loudspeaker pair, one of which comprises one of said second configuration, each of said first output signal and one of said output difference signal; intended for said associated loudspeaker pair responsive to one and forming said second configuration; output sum and output difference matrix means for generating a signal from the second signal representing the second plurality of signals proportional to the central loudspeaker supply signal for the second configuration. Matrix regeneration decoding means for obtaining either from the output of said first linear means or one of said matrix means. 13. The decoding means comprises a matrix converter according to claim 10 or 11. 13. The matrix reproduction decoding means according to claim 12, wherein the matrix reproduction decoding means comprises: 14. Let each left, center, and right loudspeaker of a three-speaker stereo configuration respond to signals L2 and R2 intended for each left and right loudspeaker of a two-speaker stereophonic configuration. In a matrix exchanger generating signals L3, C3, and R3, Hp=2-1/ for p=2 and 3. 2(Lp+Rp), and Sp=2-1/2(Lp-Rp), and φ is the frequency. Therefore, let w be a parameter between 15° and 75° that may change, and w is also the frequency. When we take a width gain greater than sinφ which depends on the number, and φ can take values close to 0° or 90° at low bass frequencies, in effect we have ▲mathematical formulas, chemical formulas, tables, etc.▼ and S3=wS2 as the frequency Claims 1 to 13 are stored in the entire gain proportional constant that can vary with The matrix converter according to any of the above. 15. Each outer left, inner left, inner right, and outer right of a four-speaker stereo configuration in response to signals L2 and R2 intended for each left and right loudspeaker of a two-speaker stereo configuration. In the matrix converter generating the signals L4, L5, R5 and R4 to be reproduced by the loudspeaker, Hp = 2-1/2 (Lp + Rp) and Sp = 2- for p = 2, 4 and 5. 1/2 (L p - Rp), and φ42, φD, and w are parameters that can change with frequency. and φ is within 25° of its “save decoder” value of 39.79° = 50.36° − 10.57°, and φD is within 25° of its “save decoder” value of 28.64°. If we assume that φ42 and φD can be between 0° and 90° at low bass frequencies, then in effect ▲There are mathematical formulas, chemical formulas, tables, etc.▼ and ▲There are mathematical formulas, chemical formulas, tables, etc. Claims 1 to 13, in which ▼ is stored in the entire gain proportional constant that can vary with frequency A matrix exchanger as described in any of the above. 16. In response to signals L3, C3, and R3 intended for the respective left, center, and right loudspeakers of a three-speaker stereo configuration, the respective outer left, inner left, and inner loudspeakers of a four-speaker stereo configuration In the matrix converter which generates the signals L4, L5, R5 and R4 to be reproduced by the right and outer right loudspeakers. Let Hp = 2-1/2 (Lp + Rp) and Sp = 2-1/2 (Lp - Rp) for p = 8, 4, and 5, and φo and φD are given angular parameters that can vary with frequency. Claims 1 to 7 in which substantially ▲There are mathematical formulas, chemical formulas, tables, etc.▼ and ▲There are mathematical formulas, chemical formulas, tables, etc.▼ are stored in the whole gain proportional constant that can vary with frequency, or Matrix converter according to any one of claims 10 to 13. 17.4-speaker stereo configuration for each outside left, inside left, inside right, and outside The respective outer left, inner left, inner right, and outer right loudspeakers of a five-speaker stereo configuration respond to signals L4, L5, R5, and E4 that are attempted to be reproduced by the side right loudspeaker. 14. Generating signals L6, L7, C5, R7 and R6 and comprising a 5x4 energy conservation matrix as described herein. Matrix converter. 18. The playback decoder is configured to transmit the first audio signal or to receive the recording medium. 2. A matrix converter according to any one of the preceding claims, comprising: an input signal having an output signal of n2 loudspeaker supply signals; and means for outputting a signal corresponding to the n2 loudspeaker supply signals. Trix regeneration decoder. 19. A first stereoacoustically encoded audio signal having components W, a conversion matrix for converting into a second stereoacoustically encoded signal for reproduction; In the box, the transformation matrix receives at one input a first signal Mdec formed from the sum of the omnidirectional component W and the first velocity component, and at the other input a first signal Mdec formed from the sum of the omnidirectional component W and the first velocity component. Any of the preceding claims arranged to receive a signal Sdec formed from a degree component Y. A transformation matrix comprising the n2×2 transformation matrix means according to any one of the above, and means for outputting still another signal component obtained from the difference Tdec between the components W and X. vinegar. 20. Furthermore, it is provided with a rotation matrix configured to apply a rotation to the sum and difference components Mdec and Tdec by an angle (φ-45°) that may be frequency dependent. 20. The exchange matrix of claim 19, wherein the exchange matrix comprises: 21. 21. The transformation matrix of claim 20, wherein φ varies from a low value in the range of substantially 25° to 45° below substantially 5 kHz to a high value in the range of substantially 45° to substantially 65° above 5 kHz. . 22. Claim 19 further comprising means for adding variable attenuation to the difference component Tdec. 22. The transformation matrix according to any one of the above. 23. 19. An audiovisual system comprising one or more loudspeakers centrally placed relative to a screen and left and right loudspeakers, the system comprising a matrix reproduction decoder according to claim 18. system. 24. A portable audio system comprising a matrix regeneration decoder according to claim 18. 25. Installed in a vehicle equipped with the matrix regenerative decoder according to claim 18. audio playback system for 26. A public address system comprising a matrix regeneration decoder according to claim 18. stem. 27. Transmission encoder according to any one of the preceding claims, characterized in that the transmission encoder comprises an input configured to receive a first audio signal and an output for outputting a second audio signal to a transmission or recording medium. A transmission matrix with a matrix decoder decoder. 28. The matrix converter further comprises means responsive to a transmission side chain signal carrying time-varying transmission matrix coefficients that change the matrix coefficients of the transmission matrix decoder. The coefficients are changed to minimize perceptual noise errors in the transmitted signal. The transmission decoder according to claim 18. 29. Let n1 and n2 be integers greater than 1, and n2<n1, then n1 speeds a first audio signal stereophonically encoded for reproduction by a speaker; a second audio signal stereophonically encoded for reproduction by n2 loudspeakers; matrix transformer Rn2,n1 for converting into a digital signal, characterized in that the matrix is a matrix transposition of the coefficients of the matrix transformer Rn1,n2 according to any one of claims 1 to 17. Matrix converter. 30. Let j be the input audio signal encoded for reproduction by i loudspeakers, where i, j are integers and at least one of i, j is ≧3 for one or more of the transformation matrices. The output is encoded to be played by multiple loudspeakers. In an audio transmission/reproduction system, which is equipped with a plurality of conversion matrices Rji in series, the conversion matrix Rji converts the output into an audio signal. For the trix Rn3n2 and Rn2n1, n2≧min(n1, n2). If then Rn3n2 Rn2n1=Rn3n1 If n2≧n1, then Rn1n2 Rn2n1=In1n1, where Inn is an n×n identity matrix, and if j>i, the transformation matrix Rji is frequency dependent. The entire playback effect of the encoded audio signal is inside the entire possible proportionality constant. an energy conservation matrix configured to substantially conserve energy; 31. Matrix transformer Rn3n1032. with coefficients determined by cascade matrix transformers Rn2n1 and Rn3n2 according to any one of claims 1 to 17. Furthermore, the distance between different loudspeakers and a given position inside the listening area introducing a delay in one or more of said signals corresponding to the supplied signal to compensate for the difference in separation; 19. A matrix regenerative decoder as claimed in claim 18, comprising delay compensation means for maintaining the required stereophonic sound effect over the listening area. 33. 18. A matrix exchanger according to claim 1, wherein the first audio signal input to the converter is a transmission signal. 34. The matrix coefficients of the matrix transformer are substantially orthogonal, unitary, or 19. A transmission matrix according to claim 18, wherein: forms an energy conservation matrix. encoder. 35. The matrix coefficients of the swapped matrices are substantially orthogonal, unitary, or energy conserving, or connect energy conserving matrices. 19. A matrix recovery decoder as claimed in claim 18, forming a combined Hermitian matrix. 36. The transducer can be an ideal quadrature, unit, energy conserving, or Hermitian coupled matrix. 3dB, preferably 2dB, or Encoder or encoder according to claim 34 or 35, further preferably offset by no more than 1 dB. 37. directionally encoded for transmission/reproduction by a first number ni channels; audio transmission/reproduction comprising a plurality of transformation matrices Rji in series for converting a first signal encoded in the audio signal into a second signal directionally encoded for reproduction by a second number nj channels; In the system, at least one of ni and nj is ≧3, and the matrix is the element of the hierarchy that can be connected in cascade as long as it is specified here. at least one of the directionally encoded signals has a directionally encoded signal whose directional encoding is Audio transmission/playback system for playback modalities that do not provide mechanical rotational symmetry. 38. A complex constructed according to the algorithms of FIGS. 14, 15, and 16. An audio transmission/playback system comprising a number conversion matrix. 39. The B-style signal is a signal in which a stereo sum signal M is formed from a front-facing combination of a directional component W and a velocity component A front-stage and rear-stage stereo hierarchy in which the rear single-tone signal B is formed from a rear-opposed combination of W and X. 40. formed as the inverse of the matrix according to claim 39, and stereo channels ambisonic or surround sound or front/back stages A decoder for use in front and rear stage stereo transmission/reproduction layers comprising: a transformation matrix configured to obtain a B-style signal for reproduction of a teleo system.