MXPA00002235A - 5-2-5 matrix encoder and decoder system - Google Patents

Abstract

A sound reproduction system for converting stereo signals on two input channels (92, 94), at least one signal component being directionally encoded and correlated and at least one signal component that is not directionally encoded and uncorrelated in the two input channels, into signals for several ouput channels, including decoding apparatus (90) for enhancing the correlated component of the input signals in the desired direction and reducing the strength of such signals in channels not associated with the encoded direction, while preserving the separation between the respective left and right ouput channels (172, 176) and the total energy of the uncorelated component of the input channels in each output channel, such that instruments recorded on the right input channel stay on the right side of the output channels and the instruments recorded on the left stay on the left side, and the apparent loudness of all the intruments in all the output channels stays the sameregardless of the direction of the directionally encoded component of the input signals, and encoding means to encode five input channels so they will encode with correct direction and level in decoders according to the invention, and in decoders according to the current film standard.

Description

MATRIX ENCODER AND DECODER SYSTEM 5-2-5 Campt rfe the Invention This invention relates to sound reproduction systems that include the decoding of a stereo pair of input audio signals, within a multiplicity of output signals for reproduction after appropriate amplification through a similar plurality of loudspeakers that are configured to surround a listener. More particularly, the invention has to do with an improved set of design criteria and their solution to create a decoding matrix that has an optimal psychoacoustic performance, with high separation between the left and right components of the stereo signals, while keeping the components encoded in a non-directional manner at a constant acoustic level without considering the direction of the directionally encoded components of the input audio signals. Additionally, this invention relates to the coding of multi-channel sound over two channels for reproduction by the decoders, in accordance with the invention. In particular, it is related to the coefficients that are put into improved matrix for an encoder system 'and matrix decoder 5-2-5.

BACKGROUND OF THE INVENTION Commonly, reference is made to the apparatus for decoding a stereophonic pair of left and right input audio signals at a multiplicity of output signals such as a decoder or processor of the surrounding sound. Surround sound decoders work by combining left and right input audio signals in different proportions to produce the multiplicity of N output signals. Different combinations of input audio signals can be described mathematically in terms of a row N by 2 column matrices, in which there are 2N coefficients, each relating the proportion of the input audio signals either left or right contained in a particular output signal. The matrix coefficients can be fixed, in which case the matrix is called passive, or they can vary in time in a way that is defined by one or more control signals, in which case the matrix is described as active. The coefficients in a decoding matrix can be real or complex. Complex coefficients in practice include the use of precise phase quadrature networks, which are expensive, and therefore the most recent surrounding sound decoders do not include them, so that all matrix coefficients are real. In most of the work described in this patent application, the elements of the matrix are also real. The actual coefficients are inexpensive and will optimally decode a five-channel film that was encoded with the active encoder that is described in this patent. However, the actual coefficients are not optimal when decoding a film that was encoded from a five-channel original using a passive encoder, such as the one described in this application, and are also not optimal when decoding a movie that it was made with the standard four channel encoder of the prior art. A modification to the design of the decoder is also described, which will decode these films optimally. Although the description is of a phase corrector to the decoder inputs, correction could also be achieved by manufacturing the array of matrix elements. In a passive matrix, which is defined as a matrix in which the coefficients are constant, such as the Dolby Surround matrix, many ideal properties are achieved by proper selection of the coefficients. These properties include the following: Signals encoded with a standard encoder will be reproduced by a passive matrix decoder with equal sound intensity regardless of its encoded address. The signals where there is no specific coded address, such as the music that has been recorded so that the two inputs to the decoder have no correlation, that is, uncorrelated signals, will be reproduced with equal sound intensity on all output channels. When the input signals are a combination of a directionally coded component and an uncorrelated component, there is no change in either the sound intensity or the apparent separation of the uncorrelated component as the coded address of the coded component Directionally changes. A disadvantage of passive decoders is that the separation of both directional and uncorrelated components of the input signals is not optimal. For example, a signal that is wanted to come from the front center, is also reproduced in the left and right front output channels, usually with a level difference of only 3dB. In this way, most modern decoders employ some variation of the matrix coefficients with the apparent direction of the predominant sound source, that is, they are active rather than passive. In the original Dolby Surround decoder format, only one rear channel output is provided, which is typically played on more than one speaker, all loudspeakers being operated in parallel, so there is no left-right separation in the channels later. However, there is a high separation between the signals that are encoded in opposite directions.

The above patents have described many aspects of active matrix surround sound decoders for the conversion of a pair of stereophonic audio signals into multiple output signals. The prior art describes how the apparent direction of a signal component encoded in a directional manner can be determined from the logarithm of the proportion of the component amplitudes in the left and right channels of the stereophonic, together with the logarithm of the ratio between the sum of these amplitudes and the difference between them. This technique will be assumed in this patent application, along with a large amount of the technique concerning the equalization of the directional control signals that are derived in this way or in any other way. We assume that these two directional control signals exist in a form that can be used. For the purposes of this invention, these directional control signals can possibly be derived from the directional information that was recorded on a sub-channel of a digital audio signal. The invention has to do with the use for which these directional control signals are placed in the control of an active matrix, which takes the signals in the two inputs and distributes them to a number of output channels in variable proportions in a manner appropriate, which depend on directional control signals. A simple example of this matrix is given in U.S. Patent No. 3,959,590 to Scheiber. Another commonly used matrix is that of Mandell, which is described in U.S. Patent No. 5,046,098. In United States Patent No. 4,862,502 to Griesinger, a matrix with four outputs is described in detail and a mathematical description of this matrix, together with a mathematical description, is given in United States Patent No. 5,136,650 to Griesinger. of a matrix of six outputs. In U.S. Patent No. 5,307,415 to Fosgate, a different six-output array is described. All of these prior art arrays distribute the input audio signals between the different outputs under the control of the directional control signals, as described above. Each of these matrices is constructed differently in a certain way, but in each case each output is formed by a sum of the two input signals, each input signal being multiplied by a coefficient. In this way, each matrix in the prior art can be fully specified by knowing the value of the two coefficients for each output and how these coefficients vary as a function of the directional control signals, which provide the directional information as described. previously. These two coefficients are the matrix elements of an N by 2 matrix, where N is the number of output channels, which completely specifies the character of the decoder. In most of the above techniques, these matrix elements are not established specifically, but can be deduced from the descriptions given. In a superior mode, they can also be easily measured. Griesinger, in the patent of the United States No. ,136,650, issued August 4, 1992, gives the complete functional dependence of each matrix element on the directional control signals. From the Griesinger issued patent referred to above, the film industry has developed a discrete sound standard "five plus one". Many premieres of cinemas and some home premieres are made with sound tracks comprising five separate broadband audio channels, namely central, left front, right front, left rear, and rear right, with a sixth audio channel Reduced broadband that is proposed for very low frequency effects. The reproduction of these sound tracks requires special digital hardware to demultiplex and decompress the audio tracks within the 5 + 1 output channels. Nevertheless, there is a large selection of prints and videos of films that were previously shown, which employ a coded format of a two-channel sound track matrix, both analogue and digital. These sound tracks are encoded during the mixing process using a standardized four-channel to two-channel encoder. Although the earlier work by Griesinger and others, he has described the outputs of the decoder in terms of a complicated sum of different signals: the input signals, their sum and their difference, and the same four signals after passing through amplifiers. variable gain that are controlled by the directional control signals, it is possible to collect the terms of each output that relate to a particular input and to describe the matrix completely in closed form, so that decoding can be performed on the components of hardware, either analog or digital. In a standard movie decoder, an increase is applied to the front channels when a strongly directed signal is present, such as a dialogue. This upsets the balance between these signals and the background or music effects, in relation to the balance between those signals in the 5-channel discrete movie theater system. The improved active encoder described herein is needed to correct the balance between strongly directed front signals and music. There is also a need to improve the performance of both the encoder and the decoder, with respect to the left-side and right-side signals. An additional improvement in the decoder is to limit the effects of abrupt changes in directional control signals, to provide a better dynamic response to rapid changes in them. In addition, the present invention constitutes additional improvements to the decoder of the Griesinger US patent applications referenced above. SUMMARY OF THE INVENTION The present invention relates to the realization of the active matrix that has certain properties that optimize its psychoacoustic performance. The invention is a surround sound decoder having variable matrix values, which is constructed so as to reduce the audio components encoded in a directional manner in the outputs that are not directly involved to reproduce them in the desired direction; improve audio components encoded directionally in the outputs that are directly involved to reproduce them in the desired direction, such as to maintain a constant total power for those signals; while maintaining a high separation between the left and right channel components of the non-directional signals, regardless of the directional signals; and maintaining the sound intensity which is defined as the total audio power level of the non-directional signals effectively constant, whether or not the coded signals are present in a directional manner and without consideration for the desired direction, if there is any present . In a preferred embodiment, a surround sound decoder is provided to redistribute a pair of left and right input signals including directionally and non-directionally encoded components, in a plurality of output channels for playback through loudspeakers that surround a listening area, and that incorporates a system of circuits to determine the directional content of the left and right audio signals and that generate from them at least one signal of left-right direction and a directional signal of the surrounding sound central. The decoder includes a system of delay circuits for delaying each of the left and right audio input signals, to provide left and right audio signals; a plurality of multipliers equal to twice the number of output channels, organized in pairs, a first element of each pair receiving the delayed left audio signal and a second element receiving the delayed right audio signal, each of the multipliers that multiply their input audio signal by the variable matrix coefficient to provide an output signal; the variable matrix coefficient being controlled by one or both of the directional signals. A plurality of addition devices is provided, one for each of the plurality of output channels, each of the addends receiving the output signals of a pair of multipliers and producing at its output one of the plurality of output signals. The decoder has the variable matrix values constructed in such a manner as to reduce the audio components encoded directionally in outputs that are not directly involved in their reproduction in the desired direction; and constructed in such a way as to improve the audio components encoded directionally in the outputs that are directly involved in their reproduction in the desired direction, such as to maintain the constant total power for those signals; while maintaining a high separation between the left and right channel components of the non-directional signal, without regard to the directional signals; and constructed in such a way as to maintain the sound intensity which is defined as the total audio power level of the non-directional signals effectively constant, whether or not the coded signals are present in a directional manner and without consideration for the desired direction , if there is any present. This invention also includes improved active encoder modes, which corrects the balance between strongly directed front signals and correlated music signals due to the increase in front signals, which occurs in a standard movie decoder, and which also increases the separation between the encoder outputs when the left and right side inputs are presented uncorrelated to the encoder. It also covers modified performance in the specifications of the film decoder with respect to the coded signals on the left or right side. An additional improvement in the decoder is related to the effects of abrupt changes in the directional control signals and limits the slower change signal to provide a better dynamic response to the rapidly changing signal. Although the invention is described mainly in terms of analog modes, an advantage of the invention is that it can be implemented as a digital signal processor. An advantage of the present invention is that the design of the decoding matrix provides high separation from left to right on all output channels. A further advantage of the invention is that it maintains this high separation regardless of the direction of the dominant encoded signal. Another advantage of the invention is that the total output power level of any non-coded non-coded signal remains constant regardless of the direction of the dominant coded signal. Another advantage of the invention is that it can conventionally reproduce encoded sound tracks in a manner that closely matches the sound of a discrete sound track premiere of channel 5 + 1. Yet another advantage of the invention is that it provides a simple passive matrix encoding two channels of a five-channel sound track that will be decoded into five or more channels with very little subjective difference from the original five channels. Another advantage of the invention is that it provides an active encoder that performs better with respect to the inputs of the left and right surround sound than that which can be achieved with a passive five-channel encoder. Although the decoder of the invention operates optimally with the active five-channel encoder, another advantage of the invention is that, with an added phase correction network, it can also optimally reproduce cinema soundtracks encoded with either the four channel passive encoder of the prior art or with the five channel passive matrix encoder, which is an aspect of the present invention. An advantage of the active matrix encoder of the invention is that it provides dynamic control of the balance between the strongly directed front signals and the non-directional music, to compensate for the increase that was applied to these directed signals in the standard movie decoders. A further advantage of the encoder is that it provides improved separation of the simultaneous left-side and right-side signals when decoded with a standard movie decoder. An advantage of the decoder of the invention is that it provides more than one level change in the front loudspeakers, relative to the back, when a signal is rotated on either side of the listener, improving the apparent movement of those signal sources. . Another advantage of the decoder according to this invention is to limit the absolute value of one of the two directional signals when the other is changing rapidly, so that the dynamic effects are reproduced better. More particularly, the present invention relates to improvements to the derivation of the appropriate variable matrix coefficients, as described above in U.S. Patent Nos. 4,862,502 (1989), 5,136,650 (1992) to Griesinger, US Patent Application No. 08 / 684,948 to Griesinger, July 1996, and United States Patent Application No. 08 / 742,460 to Griesinger, November 1996, as described in the patent application provisionally presented in September 1997. The coefficients that were previously used were implemented in a decoder, which is referred to here as version 1.11. The present invention includes two main changes to the coefficients that are derived in the earlier patent application Number 08 / 684,948, of July 19, 1996. The first is a change to the correction of the Television Matrix "in the subsequent channels. There is no change when the direction is in the posterior direction, but it was found that the results were better if the reduction of 3dB in the posterior output level is maintained as if the direction were forward.The output level rises to the level original when the absolute value of the control signal | lr | rises from zero to 22.5 degrees, but it is independent of the value of the control signal | cs |. This change is described in the section on the television matrix correction, and is illustrated in the figures reviewed for LRL and LRR. An advantage that is claimed for this review, is that there is less variation in the relative sound level of the posterior channels when conduction occurs in the forward direction, providing a more natural and smooth decoding. The second change is in the treatment of the frontal channels and the center channel. Extensive attention to the decoder version 1.11 showed that it is necessary to consider the power in all the channels when the treatment of the center channel is being determined. Mathematics' have been modified in the section in the center channel in the previous patent application (1996) to reflect this change. The assumption is that the proportion of the power in the input channel of the center of the encoder should be preserved with the power in the other channels in the total acoustic power at the decoder output. Changes in the center channel require a different strategy to implement the LRL matrix element. The strategy given here is mathematically elegant, but it includes a division. An implementation of this division uses rather a dimensional lookup table. Specifically, this invention comprises the following improvements to the coefficient values that are derived in the three patents and prior applications. First, add interpolation to the LRL matrix element in the left back quadrant near the limit cs = 0. Second, correct a software error in the LRR matrix element in the left back quadrant along the limit lr = 0. Third, add complexity to the LFL and LFR matrix elements in the left back quadrant, so that an entry that was directed toward the left rear exit of the left front exit is eliminated. Fourth, repair the matrix elements LFL and LFR so that they follow the curve of the sine and cosine along the boundary from the left to the center, and from the right to the center, while retaining an increase along the axis lr = 0. Fifth, recalculate the mathematical analysis for the LFL and LFR elements in the physical left quadrant. The objective of redesigning (which goes beyond the 1991 patent), is to make the sum of the squares of the elements equal to 1 along the axis cs = l. An advantage of this invention is to reduce the unwanted variations of the total power as a result of the direction. Sixth, include a redesigned center-channel boost function, which is increased at a lower speed than the decoder version 1.11. In this aspect of the invention, an advantage is that the center augmentation function has been carefully selected on the basis of listening tests to give a minimum sense of vowel movement or dialogue between the left and right main speakers and the center speaker, while minimizing the left - right separation of the instruments that are present, along with the vowels. Seventh, add a special function CF that replaces the previous increase in LFR along the axis 2r = 0 with a cut, the cut that is designated to preserve the sum of the powers from the outputs of the decoder, the proportion of the power of the component from the center of the signals to the encoder for the total power of the signals to the encoder. An advantage of this aspect of the invention is that this procedure causes the vowels in the music, and the dialogue in the movies, to have the same balance in the decoded environment as they did in the material before encoding it. This procedure also preserves the balance in the recordings that were originally mixed for the reproduction of two channels. The new CF function remains close to zero - this is, there is no subtraction of the right input to the decoder from the left input of the decoder when it is forming the left front output, and this value is kept low until cs reaches approximately 30 degrees towards the front. As the cs of the control signal increases above this range, the level of the center channel rises rapidly, first to a value of about 3dB lower than the value for the Dolby Pro-Logic, and then keeps constant. As the cs rises beyond 30 degrees, the center level rises rapidly to the same maximum that was used for the Dolby Pro-Logic. The CF function also rapidly decreases over this range, increasing the subtraction and removing the center component from the left and right front outputs. The value of CF also falls quickly to the previous value when the absolute value of the control signal Ir approaches the limit. Eighth, add a panning correction to the elements of the new center matrix, which correct the level along the boundaries. This aspect of the invention imparts an advantage in reducing the level fluctuations during the n direction in these directions. The main advantage of this invention is a reduction in the variations of different directional signals in the presence of strong direction, especially in the later signals when the direction is towards the front, and in the signals of the center is when the direction is in others addresses. This is particularly true in the corrections to the decoding of the television matrix. A further advantage of this invention is that it provides a smoother and more transparent reproduction of the effects of the surrounding sound, without undesired variations of the total acoustic output of the front signals of the center, due to the steering activity. Another advantage of this invention is to more accurately balance the vowel levels in the music and the dialogue in the films with respect to the non-directional sounds, so that the balance is identical in the decoded environment than that in the material before the coding. Another advantage of the invention is to preserve the balance in the recordings that were originally mixed for the reproduction of two channels. BRIEF DESCRIPTION OF THE DRAWINGS The novel features that are considered characteristic 20 'of the present invention are set forth in the appended claims. The invention itself, as well as other features and advantages thereof, will be better understood by reference to the following detailed description of an illustrative embodiment when read in conjunction with the figures of the accompanying drawings, wherein: Figure 1 is a block graphic of a Dolby surround sound decoder of passive matrix, in accordance with the prior art; Figure 2 is a block graph of a standard Dolby matrix encoder, in accordance with the prior art; Figure 3 is a block graph of a five-channel encoder for producing compatible Dolby matrix encoding of discrete five-channel sound tracks, in accordance with the present invention. Figure 4 is a block graph of a five channel mode of the decoder according to the invention; Figures 5A and 5B show detailed graphs for a typical phase shifter that can be used in the circuit of Figure 4; Figures 6a-6e show the relationships between the different signals in the decoder of Figure 4; Figure 7 shows a block graphic of an active coder according to the invention; Figure 8 shows a phase sensitive detection circuit for the generation of an / rs signal for use with the phase correction circuit of Figure 9; Figure 9 shows an input phase correction circuit that will be applied later on of the decoder of Figure 4, for the optimal decoding of the sound track of the film that was passively encoded, which includes a graph showing the relationship between the control signal the / rs and the steering angle ß ^; Figure 10 shows a block graphic of a simplified active coder according to the invention, which also includes a graph of the steering angle TRS against the control signal rs / las; Figure 11 shows a block graph of an active matrix encoder having amplitude compensation for the strongly directed frontal signals and better separation for the simultaneous side entrances, in accordance with the invention; Figures 12a-12c graphically show the variation of the GL, GC and GR signals for the front quadrant direction and the left-left (LL) and left-right (LR) array elements as the direction goes from the left to left side in the encoder of Figure 11; and Figure 13 shows graphically the maximum permissible values of each of the control signals 1 / r and c / s as the other changes, for the signals that are directed between the left and the center, as applied to the decoder of the Figure 4 or the seven-channel variant thereof. Figure 14 is a perspective graphical view showing the value of the left-left left-hand (LRL) element in version 1.11 of the decoder of the general type shown in Figure 4, illustrating a discontinuity near the left vertex; Figure 15 similarly shows the right-rear right-hand array (LRR) of version 1.11 of the decoder, which illustrates a discontinuity in the rear along the axis lr = 0; Figure 16 shows in perspective graphical view the LRL matrix element as it was intended to be in the 1.11 version of the decoder, in contrast to the imperfect matrix element in Figure 14, which was actually implemented; Figure 17 similarly shows the Left Back Left (LFL) matrix element of U.S. Patent No. 4,862,502 and Dolby Pro-Logic, graded such that the maximum value is one; Figure 18 similarly shows the left front right (LFR) matrix element of U.S. Patent No. 4,862,502 and Dolby Pro-Logic, graded by .71 so that the minimum and maximum values are ± 0.5; Figure 19 graphically represents in perspective the square root of the sum of the squares of LFL and LFR of U.S. Patent No. 4,862,502 and Dolby Pro-Logic, graded such that the maximum value is one, which shows that the value is constant at 0.71 along the axis from no direction to the right, while from no direction to the left 3dB rises to value 1, and no direction towards the center or the back drops by 3dB to the value 0.5, in whose graph the profile of posterior sense is identical to that of the central sense; Figure 20 similarly represents the square root of the sum of matrix elements LFL and LFR from U.S. Patent Application No. 08 / 742,460, graded such that the maximum value is 1, illustrating the constant value of .71 in the full right half of the plane, and the moderate elevation to one towards the left vertex; Figure 21 graphically shows in perspective the left front left (LFL) matrix element in decoder version 1.1, which illustrates that the increase as the direction moves toward the center, applies both along the axis lr = 0, as along the left border to the center, and also the reduction in level as the direction moves towards the back; Figure 22 similarly shows the left front left (LFL) matrix element having the correction width along the left to center boundary, as well as along the center to right boundary; Figure 23 is a graph showing the behavior of LFL and LFR along the posterior limit between the left and full back, where the slight flaw is due to the absence of a point at 22.5 degrees; Figure 24 shows a graphic perspective of the left front left (LFL) matrix element as seen from the left rear, which illustrates the large correction along the left-rear limit, which causes the left front output to go to zero when the direction goes from left to left back, while the output remains at zero as the direction progresses to the full back, but along the axis lr = 0 and in the right back quadrant the function is identical to that of the Dolby matrix; Figure 25 similarly shows the left front right (LFR) array element, where the large peak at the left to back limit works in conjunction with the LFL array element to reduce the front output to zero as length of this limit as the direction goes from the left rear to the full rear, and once again in the posterior direction along the axis lr = 0 and in the rear right quadrant the element is identical to the matrix Dolby; Figure 26 illustrates the average square sum of the root of LFL and LFR, in accordance with the present invention; Figure 27 shows the square root of the sum of the squares of LFL and LFR, which includes the correction at the posterior level, seen from the left rear, which shows that the axis with no direction (middle) towards the right axis has the value one, the vertex of the center has the value 0.71, the posterior vertex has the value 0.5, and the left vertex has the value 1.41, and which shows the peak along the middle part to the center axis; Figure 28 is a graph showing as a solid curve the value of the center matrix as a function of CS in dB, assuming sound power proportions identical to stereo, and using elements of the Dolby matrix with 3dB less power in the back that is typically used, and as a dotted curve the actual value of the elements of the center matrix in Pro-Logic, which illustrates that while the actual values give reasonable results for a signal without direction and a signal completely directed, these are approximately 1.5dB too low in the middle part; Figure 29 shows in a similar way as a solid curve the value of the elements of the center matrix, assuming equal power proportions to the stereo, given the elements of the matrix and the calibration that was actually used in the Dolby Pro-Logic , and as a dotted curve the actual values of the elements of the center matrix in the Pro-Logic, which illustrate that the real values are more than 3dB too low for all directions; Figure 30 shows the square root of the sum of the squares of LRL and LRR, which use the elements that were implemented in version 1.11 of the decoder that illustrates that in the left front quadrant there is a slope of 3dB along the line from the middle to the left vertex, and almost an increase of 3dB at the level along the boundary between the left and the center, which also shows the "mountain range" in the back quadrant and which includes the tilt of the "television matrix" of 3dB in the center of the plane, which it's hard to see in this projection; Figure 31 graphically illustrates the numerical solution for GS and GR for the constant power level along the cs = 0 axis, and the zero output along the boundary between the left and the center; Figure 32 graphically shows in perspective the square root of the sum of the squares of LRL and LRR, using the values for GR and GS in accordance with the present invention, which illustrates that, except for the valley that is created by the correction of the "television matrix", the sum of the squares is close to one and continuous; Figure 33 similarly shows the left-center matrix element (CL) of the four-channel decoder (and the Pro-Logic decoder), which is also the graph of the center-right matrix element (CR) if they are interchanged the left and the right, which shows that the middle part of the graph and the right and posterior vertices have the value 1, the vertex of the center has the value 1.41, but in practice this element is graduated so that the maximum value is one; Figure 34 shows by comparison the left matrix element of the center in the decoder '1.11 version, in which the mean value and the right and rear vertices have been reduced by 4.5dB, so that as cs increases, the center rises to the value of 1.41 in two angular coefficients; Figure 35 graphically shows as a solid curve the attenuation of the center that is needed for the LFLR and LFR values, in accordance with the present invention, if the power of the component of the center of the input signal in the three front channels as the direction increases towards the front, and also shows as a dotted curve the values of the center for a standard decoder; Figure 36 graphically shows as a solid curve the value of GR that is needed for the constant power ratios with center attenuation GC, according to the invention, and as a dotted curve the value of sin (cs) corrl ( the previous LFR element), while the dotted curve shows sin (cs), which illustrates that the GF is close to zero until cs reaches 30 degrees, and then increases sharply; Figure 37 shows the right front left (LFR) matrix element with the correction for the center level along the axis lr = 0, which indicates that the value is zero in the middle part of the plane (without direction) and remains at zero as cs increases to 22.5 degrees along the axis lr = 0, falling to coincide with the previous value along the limit of left to center and right to center; Figure 38 shows in perspective the left center array element (CL) with the increased function of the aggregate center according to the invention, which also shows the correction for panning along the boundary between the left and the center; Figure 39 graphically illustrates the levels of the center exit and the left exit as a signal pans from the center to the left, showing that with the correction, the center panning, although not perfect, is reasonably close the inverse of the left output (the values are inverted in the cs axis); and Figure 40 shows a block graphic of an active encoder, in accordance with the present invention. Detailed Description of the Invention Preferred embodiments of the invention include a five-channel and seven-channel decoder with maximum lateral separation, although reference will be made to general design principles that could be applied to decoders with other channel numbers as well. When designing a passive matrix, it will be assumed that the coding will follow the standard Dolby Surround matrix, and the decoder has four outputs so that the left output signal of the decoder comprises the left input once; the center is 0.7 times the left entry (strictly or .5 or 0.7071) plus 0.7 times the right entry; the right output signal is once the right input signal; and the posterior exit is the sum of 0.7 times the left entry, and -0.7 times the right entry. With reference to Figure 1, there is a simplified graph of a Dolby Surround 1 matrix decoder, in accordance with the prior art, in which the relationships of these signals are maintained. The audio signals A (LEFT) and B (RIGHT) are applied respectively to the input terminals 2, 4, and are damped by the gain buffers of units 6 and 8, respectively. They are also combined in the proportions that were previously specified by the signal combiners 10 and 12. The outputs of the buffers 6, 8 appear as the RIGHT (R) and LEFT (L) 14, 16 output terminals, respectively, and the outputs of the signal combiners 10, 12, appear at the output terminals of the CENTER (C) and CIRCUNDANTE (S) 18, 20. As previously stated, this matrix has constant gain in all directions, and all outputs are equal in amplitude when the entries do not correlate. It is possible to extend the design of the passive matrix to more than four channels. If we want to have a left rear speaker, the appropriate signal can be made by using the appropriate matrix elements, but additional conditions are required to form a unique solution; the sound intensity of the uncorrelated component of the signal must be equal in all outputs, and the separation must be high in the opposite directions. The elements of the matrix are given by the sines and cosines of the direction angle of the output. For example, if the angle a is defined so that a = 0 for a full left output and is 90 ° for an output in the front center, then the matrix elements in the front center are: Left matrix element = eos (a / 2) .. (1) Matrix element right = sin (a / 2) .. (2) Thus, for a = 90 °, the two matrix elements are 0.71, as specified by the Dolby Surround matrix standard. The matrix elements as defined by equations (1) and (2) are valid for a = 0 (full left) aa = 180 ° (full right), where the sign of the matrix element for the left, change . For the left rear quadrant, a goes from 0o to -90 °, so that the sign of the right component is negative. For the right rear quadrant, however, the sign of the left matrix element is negative. In the posterior part of the center, a = 270 ° or -90 °, and the two components are equal and opposite in the sign; Conventionally, the coefficient of the right signal is negative in this case. This could be specified by setting the range of a in equations (1) and (2) as [-90 °, 270 °], where a bracket implies the inclusion of the adjacent limit value and a parenthesis implies that the limit It is not included in the range. The separation between the two outputs is defined as the difference between the levels of a signal in one output and the signal in the other, which is expressed in decibels (dB). In this way, if there is a complete left signal, the component of the right input is zero, and the components at the left and center outputs are 1 and 0.71 times, respectively, at the left input signal. The separation is a level ratio of 0.71 or -3dB (the minus sign is normally suppressed). The separation between any of the two senses that have an angle difference of 90 °, is always 3dB for this matrix. For the senses separated by less than 90 °, the separation will be less than 3dB. For example, the outputs in the back full (a = -90 °) and the left back (a = -45 °), will have a separation given by: Separation = eos (45 °) * L / (eos) 22.5 °) * L) = 0.77 = 2.3 dB .. (3) This situation can be improved with an active matrix.

The objective of an active matrix is to improve the separation between the adjacent outputs when there is a directionally encoded signal at the decoder inputs. We can also raise the question of how this decoder behaves when the inputs consist entirely of uncorrelated "music", and how the decoder behaves when there is a mix of a directional signal and music. In this context, we should use the word "music" to denote any uncorrelated signal of such complexity, that the two directional control signals that were previously referenced and that are supposed to be derived from the input signals of stereophonic audio, they are effectively zero. The following design criteria can be applied to any active matrix, noting that different degrees of success are met by decoders in the present art. A. When there is no uncorrelated signal, there should be a minimum output from those channels not related to those included in the directional signal reproduction. For example, a signal that is attempted to be played at a location halfway between the right and the center should not produce any output on the left and the back channels. Similarly, a signal that is tried for the center, should not have any output on the left or on the right. (This is the principle of mixing in pairs, as it extends to the reproduction of surrounding sound). B. The decoder output for the directional signals should have the same sound intensity regardless of the encoded address. That is, the sum of the squares of the different outputs should be constant if a constant-level directional component moves through all the senses. Most decoders of the current technique do not achieve this criterion perfectly. There are volume intensity errors in all, but these errors are not significant in practice. This is the criterion of constant sound intensity. C. The sound intensity of a music component (ie, uncorrelated) of an input signal shall be constant on all output channels, regardless of how the directional component of the input moves, and regardless of the relative levels of the directional component and music. This requirement means that the sum of the squares of the elements of the matrix for each output must be constant as the elements of the matrix change with the direction. Decoders in the current art do not obey this criterion in ways that are usually noticeable. This could be called the constant power criterion. D. The transition between the reproduction of an uncorrelated music component only, and 1 reproduction of a directional signal only, as their relative levels change, should occur smoothly and does not include changes in the apparent sense of the sound. This criterion is also violated by the decoders in different significant ways in the current technique. It could be called the constant address criterion.

In a film decoder that must obey the specification for the Dolby Pro-Logic, a surrounding sound reproduction system in common use, does not apply criterion D above, and instead the following criterion must be satisfied E: E. you must increase in level the signal that you want to reach from any direction in the front of the room, from the left through the center to the right, by 3dB in relation to the level such that you would have the signal in a passive Dolby Surround matrix when there is little or no uncorrelated component of the input signals (that is, there is no music present). When music is the dominant input signal (no correlated components present), the level is not increased. In this way, as the decoder makes the transition from the only music signal to a purely directional encoded signal, the level of the directional signal in the frontal hemisphere must be increased. The optimal design of a decoder that matches the Dolby Pro-Logic specification, should have the uncorrelated music constant in all channels, except for the outputs where there is a strong directionally encoded signal, and the music in these channels can be raised in a maximum level of 3dB, proportional to the strength of the directional signal in relation to the music. The level of the music should never decrease in any output where there is no directionally coded signal. This could be called the minimum profit driving criterion. In all current active matrix decoders, an implied principle of operation is that, in the absence of a directionally coded signal, the matrix should be reverted to the passive matrix described above, as implemented for the desired number of channels of exit. This assumption seems reasonable at first glance; however, it is neither necessary nor desirable from the point of view of psychoacoustic perception. The decoders in accordance with this invention, replace the above assumption with a requirement: F. An active decoder array shall have a maximum lateral separation at all times, both during the reproduction of uncorrelated music signals, and for music signals. in the presence of a directionally coded signal. For example, if the music signal has only violins on the left and cellos only on the right, these locations should be maintained regardless of the strength or direction of a concurrently present directional signal. You can only relax this requirement when you are removing a strong directionally coded signal from an exit, which should not be reproduced. Under these conditions, the music will lower in level unless the elements of the matrix are altered to add more power to the affected channel from the direction opposite to the directed sense. This will reduce the separation, but this reduction in separation is difficult to hear in the presence of a strong directionally coded signal. The need for high separation (especially when there is no directionally coded signal) comes from psychoacoustics. The prior art has conceived the matrix as inherently symmetric, all senses being treated as equally important. However, this is not the case in practice. Humans have two ears, and when they are watching a movie or listening to music, they are usually looking forward. In this way, frontal and lateral sounds are perceived differently. There is a dramatic difference between a sound field that has up to 4dB of separation and one that has more. (This fact was recognized in the CBS SQ matrix, which had a lateral separation that exceeded 8dB in the passive decoder, while sacrificing separation from the front to the back). In the opinion of the inventor, the difference between a discrete five-channel movie reproduction and a conventional matrix reproduction is due to the low lateral separation between the channels of the surrounding sound. Griesinger, in U.S. Patent No. 5,136,650, recognizes the value of this requirement (F) and describes a six-channel decoder, where two additional channels are designed to be placed on the sides of the listener. These outputs have the desired properties for a left and rear right after exit channel, provided that the directional component of the output is directed to the frontal hemisphere. That is, they reduce the level of the directed component, regardless of its direction, and have a left separation. - Full right when there is no directionally coded signal. The outputs described in the patent referred to above, do not have a constant level for non-directionally encoded music, in the presence of a directed signal, and that defect is corrected in the present invention. The design of the encoder in the patent referred to above was used with some modification to manufacture a number of commercially available decoders. The design of the matrix in the posterior hemisphere for these decoders was developed in a heuristic manner, but generally meets well the requirements that were established previously. There is, however, more "pulsation" with the music than would be optimal, and the loss of signals directed between the left and right rear outputs is more than the desired level. In this context, the "pulsation" is the audible variation of the signal of the music that is due to the variation of the directional control signals that are responding to the sense of the signal encoded in a directional manner.

For both reasons, it was necessary to improve the design of the decoder, and this invention was the result of this design effort. It turns out that the requirements A to F above uniquely specify a matrix, which will be described mathematically below. For mathematical simplification, the encoder that is assumed in the design of the decoder, is a simple pan left-right pan. When going from left to center to the right, a standard sine-cosine curve is used, as described in equations (1) and (2) above. These can be exposed again in the form: L = eos t .. (4) R = sin t .. (5) where t = a I 2 .. (6) In the previous frontal direction mode, the angle t It varies from 0 to 90 °. For the direction in the back half of the room, from left to back (surrounding) to right, the polarity of exit of the panning vessel of the right channel is inverted. This can be described by the pair of equations: L = eos t .. (7) R = -sen t .. (8) The complete posterior direction occurs when t = 45 °, and the direction to the surrounding left sound, a position intermediate between the left and the back, occurs when t = 22.5 °. Note the similarity of this coding with the matrix elements of the passive matrix described above. Here, however, the direction angle is divided by two and the sign change for the posterior direction is included explicitly. When designing the decoder, it must first be decided what outputs will be provided, and how the amplitude of the targeted component of the input will vary as the address angle of the input coding varies. In the mathematical description below, this function can be arbitrary. However, in order to satisfy requirement B, the constant volume intensity criterion, so that the sound intensity is preserved as a signal paging between two outputs, there are some obvious selections for these amplitude functions. Assuming there will be left, right, and left left outputs, it is assumed that the amplitude function for each of these outputs is the sine or the cosine of twice the angle t.

For example, as t varies from the left, t = 0, to the center, t = 45 °, the output amplitudes should be: Left output = eos 2t •• (9) Central output = sin 2t .. (10 ) Right output = 0 .. (11) As t goes from the center to the right, t = 45 ° to 90 'Left output = 0 .. (12) Central output = sin (2t - 90 °) = -eos 2t .. (13) Right output = cos (2t - 90 °) = sin 2t .. (14) These functions they result in the optimal placement of sources between the left and the center, and between the right and the center. These functions also result in very simple solutions to the matrix problem. In any of the above cases, any output signals that are desired for reproduction in the back of the room must be zero in an identical manner. When designing the five-channel version of the enhanced decoder, a signal directed in the posterior hemisphere between the left and the surrounding left sound, t = 0 at = 22.5 °, should have: Left rear output = sin 4t .. (15 ) Right rear exit = 0 .. (16) and when it goes between the surrounding left sound and the complete rear sound, the total rear exit should remain the same. The matrix coefficients that are used to achieve this are not constant, but vary so that in the complete posterior direction the matrix element for the right input within the left posterior output goes to zero.

In the seven-channel mode, as t goes from 0 to 22.5 °, the output in both the left and left rear output must be equal and rise smoothly, proportional to the sin A t. As t goes from 22.5 ° to 45 °, the output on the left side goes down 6dB and the output on the left back goes up 2dB, maintaining the total sound intensity, the sum of the squares of each output, constant. As mentioned above, in the improved decoder even when the directed signal is completely towards the rear, the left rear and right rear outputs have the maximum separation for the uncorrelated music, because the elements of the matrix for the right input towards the left rear exit (and for the left entrance inside the right rear exit) are zero, resulting in complete separation. Although the right rear has the zero output for a directed signal as the steering angle t goes from 0 to 22.5 °, the elements of the array that are used to achieve this signal cancellation are adjusted so that the output of music is constant and has a minimal correlation with the music signal on the left back. To further decrease the correlation in the surrounding sound field, the seven-channel mode includes a time delay of approximately 15 ms in the side channels, and in the two versions the subsequent channels are delayed by approximately 25 ms. Once the functions of the sound intensity for the different outputs are selected under the conditions of direction, these functions having symmetry from left to right, the functional dependence of the elements of the matrix on the steering angle can be calculated. A standard Dolby surround sound installation has all the surround sound speakers wired in phase, and the Dolby projection rooms are similarly equipped. However, the standard passive matrix, which is described above with reference to Figure 1, has a problem with the left rear and right rear outputs. A left panning to the surrounding sound results in a transition between L and L-R, and a panning from the left to the surrounding sound ranges from R to R-L. In this way, the two rear exits are out of phase when they are completely directed to the rear. The Fosgate 6-axis decoder, which is described in U.S. Patent No. 5,307,415 among others, has this phase anomaly. When listening to these decoders, it was felt that this reversal of the phase was unacceptable, since a later directed sound, such as a passing plane, became both thin and in phase at the rear. The decoder of the present invention includes a phase shifter for changing the sign of the right rear exit under the full posterior direction. The phase shift becomes a function of the logarithmic ratio of the center over the surrounding sound, and is inactive when there is a forward direction. The typical phase shifters for this purpose are described below with reference to Figures 5a and 5b. The real world encoders are not as simple as the panning container mentioned above. However, by careful selection of the method to detect the direction angle of the inputs, problems with a standard four channel encoder can be largely avoided. In this way, even a standard film that is made with a four-channel encoder will decode with a substantial amount of directional guidance in the posterior hemisphere. With reference to Figure 2, which represents a standard encoder 21, in accordance with the prior art, as shown in Figure 1 of the prior United States Patent No. 5,136,650 to Griesinger, there are four input signals L, R, C, and S (for its acronyms in English, for left, right, central and surrounding, respectively), which are applied to the corresponding terminals 22, 24, 26 and 28 and the combinators of signals and elements of displacement of phase as shown. The left signal (L) 23 of terminal 22 and the signal of the center (C) 25 of terminal 24, are applied to a signal combiner 30 in proportions of 1 and 0.707 respectively; the right (R) signal 27 of the terminal 26 and the center signal (C) 25 are similarly applied with the same proportions to the signal combiner 32. The output 31 of the signal combiner 31 is applied to a phase shifter 34, and the output 33 of the signal combiner 32 is applied to a second identical phase shifter 38. The signal of the surrounding sound (S) 29 of the terminal 28 is applied to a third phase shifter 36, which has a phase delay of 90 ° relative to the phase shifters 34, 38. The output 35 of the shifter phase 34 is applied to the signal combiner 40, along with 0.707 times the output 37 of the phase shifter 36. Similarly, the output 39 of the phase shifter 39 is combined with -0.707 times the output 37 of the phase shifter 36 in the signal combiner 42. The outputs A and B of the encoder are the output signals 41 and 43 of the signal combiners 40 and 42, respectively. Mathematically, these encoder inputs can be described by the equations: Left output (A) = L + 0.707C - 0.707JS .. (17) Right output (B) = R + 0.707C + 0.707JS .. (18) Although a standard four-channel encoder will not work with the discrete five-channel film, it is possible to design a five-channel encoder that will work very well with the improved decoder, in accordance with the present invention. This encoder is described with reference to Figure 3. The additional elements of the new encoder 48 are applied later than the standard encoder 21 of Figure 2, which was described above. The left, middle and right signals 51, 53 and 55 are applied to terminals 50, 52 and 54, respectively, of Figure 3. In each of the left, middle and right channels, a phase shifter is inserted. all step, 56, 58 and 60, respectively, which has a phase shift function f (f) (shown as f) in the signal path. The signal from the left surround sound 63 is applied to the input terminal 62 and then through the phase shifter of all step 66 with the phase shift function f-90 °. The signal from the right surround sound 65 of the input terminal 64 is applied to a phase shifter 68 of f-90 °. The signal combiner 70 combines the output signal 57 of the left phase shifter of the phase shifter 56 with 0.83 times the output signal changed by the surround phase 67 left of the phase shifter 66, to produce the output signal 71 which label L, which is applied via terminal 76 to the left input terminal 22 of standard encoder 21. Similarly, signal combiner 72 combines the output signal 61 of the right phase shifter of phase shifter 60. with -0.83 times the output signal changed by phase 69 of surround sound right of phase shifter 68, to produce output signal 73 which is labeled R, which is applied by means of terminal 82 to the right input terminal 26 of the standard encoder 21. Similarly, the signal combiner 74 combines -0.53 times the output signal 67 of the phase shifter of the left surround sound of the fader displacer. 66 with 0.53 times the output signal 69 changed by phase of the surround sound right of the phase shifter 68, to produce the output signal 75 which is labeled S, which is applied via the terminal 80 to the input terminal of surrounding sound 28 of the standard encoder 21. The output signal 59 of the central phase shifter58, which is labeled C, is applied via terminal 78 to the central input terminal 24 of standard encoder 21. The encoder of FIG. 3 has the property that a signal in any of the discrete inputs LAS, L, C, R and RS will produce a coded signal that will be reproduced correctly by the decoder of the present invention. A signal that is in phase at the two inputs LAS, LR of the surrounding sound, will produce a fully directed rear input, and a signal that is out of phase at the two inputs of the surrounding sound will produce a signal without direction, because the outputs A and B of the standard encoder will be in quadrature. The mathematical description of the encoder of Figure 3 which is used in conjunction with the standard encoder of Figure 2, can be given in the form of: A = (L - J0.83LS) + 0.71C + 0.38 (RS - LAS) .. (19) B = (R + J0.83RS) + 0.71C - 0.38 (RS - LAS) .. (20) All current surround decoders, which use active matrices, control the matrix coefficients that are based on the information that is supplied from the input signals. All current decoders, including that of the present invention, derive this information by finding the logarithms of the rectified and smoothed left and right input signals A and B, their sum A + B and their difference A-B. Then these four logarithms are subtracted to obtain the logarithmic proportion of the left and right signals, 1 / r, and the logarithmic ratio of the sum and difference signals, which will be identified as c / s, for the center on the surrounding sound . In this description, we assume that 1 / r and c / s are expressed in decibels, so that 1 / r is positive if the left channel is higher than the right one, and c / s is positive if the signal is directed forward, it is say, the sum signal is larger than the difference signal. The attenuation values in the passive encoder of five previous channels are chosen to produce the same value of 1 / r when only the LAS input is addressed, it being understood that the simplified encoder is used to design the decoder when the angle t has been set at 22.5 ° (later). In this case, 1 / r is 2.41 or approximately 8dB. For a monophonic signal that has been distributed with the simplified encoder between the two input channels such as A = cos and B = ± sent, 1 / r and c / s are not independent. To find the direction angle t, we only need to find the arctangent of the left level divided by the right level, or if we define the complete left as t = 0, then: t = 90 ° - arctan (10"((1 / r) / 20)) .. (21) degrees if 1 / r is in dB as stated above, however, because the two levels are only compared in magnitude, to determine if the direction is frontal or posterior, we need to know the sign of c / s, which is positive for the forward direction and negative for the backward direction In the real world, the input signals to the decoder are not derived from a panning vessel, but from In addition, there is almost always uncorrelated "music" present along with the directed signals.In the following description, the problem of specifying the elements of the speech is shown in Figure 2, which uses quadrature phase shifters. matrix is divided into four sections, depending on which quadrant of the encoded space is being used, that is, left front, left back, right front or back right. We will assume a seven-channel decoder with front left, center, front right, left side, right side, left rear and right rear outputs. You must specify two matrix elements for each output, and these will be different depending on the quadrant for the address. The right frontal and right posterior quadrant coefficients can be found by reflecting around the front-back axis, since the matrix has a left-right symmetry, so that only the effects of the left front and left back direction will be derived here. For the front quadrant, we will assume that the previous D requirement is used, rather than the E requirement for the surrounding Dolby sound, and we will add the correction later. The front direction is similar to that of Griesinger (U.S. Patent No. 5,136,650) but the functions that describe the direction in the present invention are different, and unique. To find them, we must consider each output separately. The left output should decrease to zero as the angle t varies from 0 to 45 °, since we do not want any central directional signal to appear in the left front channel. If t = 0 is full left, we define an angle ts = arctan (10"((c / s / 20)) - 45 ° .. (22) The left output is the element of matrix LL times the left input plus the element matrix Lr times the right input A completely directed signal from the simplified encoder results in the left input A = eos ts and the right input B = sin ts over this range We want the level in the left output to decrease smoothly as t increases, following the function FL (ts), which in our example decoder is assumed to be equal to cos (2 s) .This way, the left output is described by: Left output = LL eos ts + LR sin ts = FL (ts) = cos (2ts) .. (23) If the output for the uncorrelated music must be constant, the sum of the squares of the coefficients of the matrix must be one, ie LL2 + LR2 = 1 .. (24) These equations, which are basically in the same form for all outputs, gives As a result, the quadratic equation for LFR, which has two solutions. In all cases, one of these solutions is preferred over the other. For the left output, LR = sin ts cos (2fcs) +/- cos ts (2ts) .. (25) LL = cos ts (2 ts) - / + sin ts sin (2ts) .. (26) By choosing the preferred sign, which is less in equation (25) and more in equation (26), and by applying mathematical identities, this is further simplified to: LL = eos ts .. (27) LR = - sen ts. . (28) The right output must be zero above the same range of the angle ts, ie Right output = RL eos ts + RR sin ts = 0 .. (29) Again, the uncorrelated music should be constant, so that RL2 + RR2 = 1 .. (30) and these lead by the same reasoning to the result RL = -sen ts .. (31) RR = eos ts .. (32) The central output should decrease smoothly as the direction moves either to the left or to the right, and this decrease should be controlled by the magnitude of 1 / r, not the magnitude of c / s. This will result in very different values for the CL element of the left center matrix and the center right CR element, which will be changed when the direction changes from right to left. The steering angle based on 1 / r, will be called tl here. It is assumed to go from 0 on the full left to 45 ° when the address is full central or when there is no directed signal. tl = 90 ° - arctan (10"((1 / r) / 20)) .. (33) where 1 / r is expressed in dB The central output should increase smoothly as tl varies from 0 (full left) at 45 ° (center) The function for this increase will be called FC (t?), which is equal to sin (2tl) in this modality.With the above method, Central output = CL eos tl + CR sin tl = FC (tl) = sin (2tl) .. (34) Again, for the constant sound intensity of the music, CL2 + CR2 = 1 .. (35) which produces the solutions CR = sin tl sin (2 tl ) - / + eos tl cos (2 tl) .. (36) CL = eos tl sin (2 tl) +/- sin tl cos (2 tl) .. (37) The preferred sign is more in the equation (36) ) and less in equation (37) The elements of the matrix for the posterior outputs during the frontal direction are not as simple to derive as those for the frontal outputs.To derive them, we use the argument and the formulas presented in the U.S. Patent No. 5,136,650 to Griesinger. oakma is that we want the matrix element The LRL on the left is 1 when there is no direction, and yet we do not want any directional output from this channel during the left or central direction. If we follow the method that was used previously, we obtain the elements of the matrix that do not output when the signal goes to the left or the center, but when there is no direction, the output will be the sum of the input signals. This is a conventional solution, where there is a poor separation when the direction is stopped. We want the complete separation, which means that LRL must be one and LRR must be zero without direction. To solve this problem, the matrix must be made to be dependent on both the value of 1 / r and that of c / s. In U.S. Patent No. 5,136,650, Griesinger gives a solution, in which the left and right outputs are the "complementary outputs". The solution that is derived there resolves the problem of canceling the directional component at all angles in the left lateral output, but the music component of the output decreases by 3dB as the direction goes to the full center. We can correct the coefficients to avoid this defect by multiplying them by the factor (eos ts + sin ts), where ts is an angle that is zero when c / s is one, and which increases to 45 ° when c / s It is big and positive. In the following equations, the angles ts and tl are derived from c / s and 1 / r, 'respectively: ts = arctan (c / s) - 45 ° .. (38) tl = arctan (1 / r) - 45 ° .. (39) Note that the tl here is different from the angle previously defined for the exit from the center.

In the terminology of the previous patent, the control signals that were devel in the inputs for different variable gain amplifiers (VGAs), are called GL, GC, GR and GS (for its acronym in English) for left, central, right and surrounding, respectively, and two complementary GSL and GSR signals are derived from these for the VGAs of the surrounding left and right sound.

The coefficients described here use a linear combination of the G values to provide the left and right coefficients as a function of the two angles ts, which are derived from c / s, and tl, which is derived from 1 / r, respectively . By the definitions in the present, GL = ((eos tl - sen tl) / eos ti) = 1- tan tl .. (40) GC = 2 (sin ts / (eos ts + sin ts)) .. (41 ) (there is a factor of two that was omitted in the printing of the previous patent), GS = (costs + sents) '1- (l-sentl). (costl - felt) -0.5 'eos tl eos tl 2sents = 0 (costs + sents) .. (42) (since this is a frontal quadrant), and GSL = GL ((1 - sin tl) / eos tl) = GL (sec tl - tan tl) = (1 - tan tl) (sec tl - tan tl) = (cos ts + sents) ' - (costl - felt) - 0.5 '2sents costl (costs + sents) .. (43) and the complementary left and right signals are given by: LAS = A (l-GSD-0.5 (A + B) GC - 0.5 ( AB) GS - B x GL .. (44) RS = B (l-GSR) -0.5 (A + B) GC + 0.5 (AB) GS - A x GR .. (45) In this way, the LSL coefficients and LRL are given by: LSL = LRL = (eos ts + sin ts) (1 - GSL - 0.5GC) .. (46) which becomes, after some manipulation, LSL = LRL = (eos ts + sin ts) (sec tl - 1) x (sec tl - tan tl) - sin ts .. (47) The coefficients LSR and LRR are also equal, given by: LSR = LRR = (eos ts + sin ts) (-0.5) GC - GL) .. (48) which becomes, after some manipulation, LSR = LRR = (eos ts + sin ts) (tan tl - 1) - sin ts .. (49) The outputs on the right side and later when the entry between the left and the center is directed, they can be found with the previous method, but the direction angle that is used should be ts, which is derived from c / s, so that it will be reversed. to the right entrance when there is no address. We just need to remove the signals that are directed towards the center. The equations to be solved are: Right posterior output = RRL eos ts - RRR sin ts = 0 .. (50) and RRL2 + RRR2 = 1 .. (51) which produces the solution: RRR = RSR = eos ts RRL = RSL = sin ts .. (52) The above equations completely specify the elements of the matrix for the frontal direction. For later direction, when c / s is negative, the following is true: The left and right main elements are the same as for the front direction, except that the angle ts is determined from the absolute value of log (c / s ) which produces: ts = arctan (10"(s / c) / 20)) - 45 ° .. (53) and the sign of the transverse matrix element is reversed, producing: LL = eos ts .. (54) LR = sin ts .. (55) and RL = sin ts .. (56) RR = ts ts. (57) The elements of the central matrix are identical in the posterior direction since they depend only on the angles that are derived of 1 / r, and are not dependent on the sign of c / s.The left lateral and right lateral exits should have a full separation when the direction is low or zero.However, the signal in the left lateral and right lateral outputs they must be removed when there is a strong left direction, we use the previous definition for tl for the direction of the center, tl = 90 ° - arctan (10"((1 / r) / 20)) .. (58) as tl varies from 0 to 22.5 °. Under a strong direction, the left-lateral and left-rear outputs are zero and when tl = 0o, but increase with tl in accordance with the value sin 4 ti. In the presence of uncorrelated music, which is represented by the signals A = eos t, B = -sen t, the coefficients LSL, LRL, LSR, and RSR must satisfy: LSL = LRL .. (59) LSR = LRR. (60) to have equal outputs on the sides and on the back, and the amplitude during the direction follows FS (tl) = sin 4tl, so that LSL eos tl - LSR sin tl = FS (t?) .. (61) So that the music has a constant level, LSL2 + LSR2 = 1 .. (62) Solving as before, -LSR = sin tl FS (tl) +/- eos tl 7 (1 - FS (tl) 2) .. (63) LSL = eos tl FS (tl) - / + sin tl (1 - FS (tl) 2) .. (64) Simplifying and using the preferred sign, as before, -LSR = sin tl sin 4 ti + eos tl eos 4 l .. (65) LSL = eos tl sin 4tl - sen tl eos 4tl .. (66) which can be further reduced to: -LSR = eos 3tl .. (67) LSL = sin 3 tl .. (68) The right lateral and right rear exits are inherently free from the left entrance when there is an address in the left rear quadrant, eg We must remove the signals directed from the center or the back, so that terms that are sensitive to C / S should be included. The right side and right rear outputs are the same, except for the different delays, and we have to solve: Right lateral / rear output = RSL eos ts + RSR sin ts = 0 .. (69) RSL2 + RSR2 = 1 .. (70 ) which produces the solution: RSL = sin ts .. (71) 'RSR = eos ts .. (72) So far, the design of the decoder complies with all the requirements that were established at the beginning. The signals are removed from the exits where they do not belong, the complete separation is maintained when there is no direction, and the music has a constant level in all exits, regardless of the direction. Unfortunately, we can not meet all of these requirements for the later exit in the subsequent quadrant. It should break with one of the hypotheses, and the least problematic to break, is the assumption of constant music level as the direction goes to the full back. The standard movie decoder does not increase the level for the rear speaker, and in this way a standard movie decoder does not increase the level of the music as a sound effect moves to the back. The standard film decoder has no separation in the rear channels. We can obtain the posterior separation that we desire only by means of allowing the level of the music to be increased by 3dB during the subsequent strong direction. This is more acceptable in practice. Some increase in the level of music under these conditions is not audible - it may even be desirable. We have been finding the elements of the matrix towards the back based on a steering angle tl that is derived from the ratio of level 1 / r. As we move from tl = 22.5 ° to tl = 45 °, this ratio that is expressed in dB decreases to zero, while the logarithmic ratio of the center to the surrounding sound (c / s) becomes a large negative value. Consider what happens when a directional signal at tl = 22.5 ° weakens to non-directional music. In this case, again, the logarithm of 1 / r decreases to zero as non-directional music becomes predominant. We need to distinguish this case from the previous one, where the direction goes strongly to the back. The best solution is to make the elements of the matrix relax at a high separation when 1 / r goes to zero, while keeping the music level constant. It is easy to derive the result: tl = 90 ° - arctan (1 / r) .. (73) LRL = eos (45 ° - tl) .. (74) LRR = -sen (45 ° - tl) .. (75 ) where tl goes from 22.5 ° to 45 °. These elements of the matrix maintain the level of the music constant, but cause the output of a directed signal to decrease by 3dB when the signal goes to the back. We can fix this by adding a dependency in c / s, by increasing the LRL value by means of an amount proportional to the increase in the logarithmic proportion of c / s. When solving for the value of the increase that is needed to maintain the constant subsequent output level, we can express the results in a table: c / s in dB RBOOST -32 0.41 -23 0.29 -18 0.19 -15 0.12 -13 0.06 -11 0.03 -9 0.01 -8 0.00 TABLE 1. Variation of RBOOST with c / s In terms of these results, the coefficients of the left rear output matrix in the five channel version are: LSL = eos (45 ° - tl) + RBOOST (log c / s) .. (76) LSR = -sen (45 ° - tl) .. (77) and similarly for the right channel, RSL = sin (45 ° - ti) .. (78) RSR = eos (45 ° - tl) + RBOOST (log c / s) .. (79) For the seven mode channels of the invention, we add an additional dependence on c / s, to take into account the desired reduction of the output in the left lateral and right lateral channels, as the direction goes to the full posterior part, remembering that the lateral coefficients left and rear left were the same in the case of the direction from the full left to the left rear. The reduction of the lateral output is achieved by an increase in the corresponding posterior output to keep the power in the directed signal constant. It may also be desirable to increase the transverse term, which reduces the separation a little, but apparently this is not audible. We define a posterior lateral increase function RSBOOS (ts), using an angle ts derived from the value of c / s: ts = 90 ° - arctan (s / c) where ts varies from 22.5 ° to 45 °, so that the RSBOOST function rises from zero at ts = 22.5 ° to 0.5 at ts = 45 °.

Then, RSBOOST = 0.5 sin (2 (ts - 22.5 °)) .. (80) and for the lateral outputs, LSL = eos (45 ° - tl) + RBOOST (log c / s) - RSBOOST (ts) .. (81) LSR = -sen (45 ° - tl) .. (82) RSL = sin (45 ° - tl) .. (83) RSR = eos (45 ° - tl) - RBOOST (log c / s) - RSBOOST (ts) .. (84) and for subsequent outputs, LRL = eos (45 ° - tl) + RBOOST (log c / s) + 0.5 RSBOOST (ts) .. (85) LRR = -sen (45 ° - tl) .. (86) RRL = sin (45 ° - tl) .. (87) RRR = eos (45 ° - tl) + RBOOST (log c / s) + 0.5 RSBOOST (ts) .. (88) For the movie decoder mode, we have to replace the previous criterion D by the criterion E, which causes an increase of the levels in the front channels by 3dB in all the frontal senses. The matrix can be made to perform in this manner by adding augmented terms similarly to the front elements during the front direction. For example, during the left direction the element of matrix LL, which is called here LFL, must be increased by an increase function that depends on 1 / r, where we define two angles: tlr = 90 ° - arctan (l / r ) .. (89) tlr = 90 ° - arctan (r / l) .. (90) Then (see eq. (27) above), LFL = eos ts + LFBOOST (tlr) .. (91) and for the direction to the right, RFR = eos ts + LFBOOST (trl) .. (92) The two central matrix elements are also increased during the direction of the center: CL = sin tl + 0.71 LFBOOST (ts) .. (93) CR = eos tl + 0.71 LFBOOST (ts) .. (94) These equations completely specify the additional requirements for a film decoder. When there is no central channel loudspeaker, the Dolby specification suggests that the center channel output should be added to the left front and right front outputs with a gain of -3dB or 0.707. While this reproduces the dialogue of the central channel at the appropriate level, it reduces the separation between the left and the right. For example, when there is no address, the central output is 0.71L + 0.71R. Adding this to the left and right produces a left output of 1.5L + 0.5R and a right output of 1.5R + 0.5L, so that the separation is reduced to 0.5 / 1.5 = 9.5dB. To avoid this effect, it would be better to modify the left and right matrix elements when there is a central direction, using the angle ts that is derived from c / s, so that: LFL = 1 + LFBOOST (ts) .. (95) RFR = 1 + LFBOOST (ts) .. (96) LFR = RFL = 0 .. (97) Unlike previously derived matrix coefficients, these do not remove the dialog from the left and right channels, and also maintain it at the appropriate sound intensity in the room, while maintaining the full left-right separation for the music, as long as the direction is in the frontal hemisphere. In a preferred five-channel mode shown in Figure 4, five of the seven channels described above are implemented, and the decoder provides the left, middle, right, left rear, and right rear outputs, omitting the side outputs left and right side. It is understood from the above mathematical description that the circuit system for the left rear and right rear outputs of the seven channel decoder can be obtained by the circuit system similar to that for the left and right surround sound outputs that are they show, with an additional lOms delay, similar to blocks 96 and 118, which implement delays of 15ms. The addition of the RBOOST, RSBOOST and LFBOOST functions as described for the seven-channel decoder, the movie decoder mode and the missing center channel mode in the last section will be apparent simple modifications for those skilled in the art. In the digital implementation, these consist merely of the addition of the appropriate augmentation expressions that are derived from the angles ts and tl with the appropriate definitions that are based on the direction directed towards the corresponding matrix coefficients, before performing the multiplications and additions that are required to generate the output signals placed on matrix. In the decoder 90 of Figure 4, the input terminals 92 and 94, respectively, receive the left and right stereo audio input signals which are labeled A and B, which can typically be outputs of the encoders of Figures 2 , 3, or 7, directly or after transmission / recording and reception / playback through typical audio playback media. The signal A at terminal 92 goes through a short delay (typically 15ms) before application to other circuit elements that will be described later, so as to allow processing of the signal, which results in completion the signals 1 / r and c / s in a similar period of time, causing by the same thing that the control signals act on the audio signals delayed at the precise moment to direct them to the appropriate loudspeakers. The signal A from terminal 92 is damped by a unit gain buffer 98 and passed to a rectifier circuit 100 and a logarithmic amplifier 102. Similarly, signal B from terminal 94 is passed through a buffer 104 , a rectifier 106 and a logarithmic amplifier 108. The outputs of the logarithmic amplifiers 102 and 108, which are labeled "A" and "B" respectively, are combined by the subtracter 110 to produce the directional control signal 1 / r, the which is passed through the switch 112 to the matrix circuitry described below. In the alternative position of the switch 112, a resistor 114 comprising a time constant and a capacitor 116 are interposed in this path to delay the output transitions of the 1 / r signal. The signal B from terminal 94 is also passed through a delay of 15ms for the reason stated above. The signals A and B of the terminals 92 and 94 are combined in an analog adder 120, rectified by the rectifier 122 and passed through the logarithmic amplifier 124. Similarly, the signals A and B are subtracted in the subtracter 126 , then passed through the rectifier 128 and the logarithmic amplifier 130. The signals of the logarithmic amplifiers 124 and 130 are combined in the subtracter 132 to produce the signal c / s, which is passed through switch 134. In the alternative position of switch 134, the signal passes through the same constant that is formed by resistor 136 and capacitor 138, which have identical values to components 114 and 116 corresponding. Thus far, the control voltage generation circuit has been described. As is typical of these circuits, the signals 1 / r and c / s vary in proportion to the logarithms of the proportions between the amplitudes of the left A and right B, and of the center (sum) and the surrounding sound (difference) of these signs The matrix elements are represented by the circuit blocks 140-158, which are each labeled in accordance with the coefficient they exhibit, in accordance with the preceding equations. In this way, for example, the block 140 labeled LL, performs the function described by equation (27), (54), (91) or (95), as appropriate. In each case, this function depends on the output of c / s, which is shown as an input to this block with an arrow, to designate it as a control input, rather than an audio signal input. The audio input is the delayed version of the left input signal A after passing through the delay block 96, and is multiplied by the coefficient LL in block 140, to produce the output signal from this block. The outputs of the different array elements are summed in the adders 160-168, thereby providing the five outputs L, C, R, LAS and RS at terminals 172, 174, 176, 178, and 180, respectively. As mentioned above, the signal RS is passed through a variable phase shifter 170, before being applied to the output terminal 180. The phase shifter 170 is controlled by the signal c / s to provide a phase shift, which changes from 0 to 180 ° as the c / s signal is directed from the front to the rear. In the seven-channel version of the decoder, the circuit elements 152-158, 166, 168 and 170 are duplicated, which are being fed from the same points as their corresponding elements shown in Figure 4, but with the LRL coefficients , LRR, RRL and RRR in blocks corresponding to 152 - 158, respectively, and with additional lOms delays, similar to blocks 96 and 118, which can be inserted either ahead of these blocks or after the elements adders corresponding to blocks 166 and 168. Although an analog implementation is shown in Figure 4, it is equally possible, and could be more physically simple, to implement the decoder functions completely in the digital domain, using a digital signal processor chip (DSP, for its acronym in English). These chips will be familiar to those skilled in the art, and the block chart of Figure 4 will be easily implemented as a program operating in that DSP to perform the various delays, multiplications and additions of the signal, as well as to derive the signals 1 / r and c / s and the angles tl and ts of these signals, which will be used in the equations that were previously described, in order to provide the complete functionality of the decoder, in accordance with the present invention. Paying attention to Figure 5a, an analog version of the phase shifter 170 is shown. In this phase shifter circuit, the input signal RS 'is damped by an operational amplifier 184 with the input resistor 186 and the feedback resistor 188 just as it defines unity gain. The outputs of the amplifiers 182 and 184 are respectively applied through the variable resistor 190 and the capacitor 192 to the third operational amplifier 196, which dampens the voltage at the junction of the variable resistor 190 and the capacitor 192, to provide the RS output signal to terminal 180 of Figure 4. This circuit is a phase shifter in a single conventional pole having all-pass characteristics. The variable resistance 190 is controlled by the c / s signal in such a way that the shift frequency of the phase shifter is high when the signal is directed towards the front., so that the subsequent output signals are out of phase (due to the matrix coefficients) but is reduced as the signal is directed towards the rear, so that the subsequent output signals become in phase due to the investment of the right rear exit RS.

Although the phase shift is not the same at all frequencies, the psychoacoustic effect of this phase shifter is acceptable and substantially reduces the phase adjustment of the subsequent signals. As will be apparent to those skilled in the art, more complex multi-pole phase shifters could be used, but would require an additional circuitry on all output channels, so an effective way to reverse cost would not be provided. Gently phase the one channel back where it is desired. Figure 5b shows a conventional variable digital delay element that can be used in the implementation of a digital mode of the delay block 170 of the circuit of Figure 4. In this circuit, the gain value g is controlled by the value of the control signal c / s, so as to perform the same function as for the analog phase shifter of Figure 5a. In this circuit, the signals that are applied to the adder 200 are added and delayed by the delay block 202, the output of which is fed back through a multiplier 204 of the gain g, to one of the inputs of the adder 200. The signal RS 'is applied to the other input of the adder 204 and also to the multiplier 206, where it is multiplied by a coefficient -g. The output signal from the delay block 202 is multiplied by (1 -g2) in the multiplier 208, and added to the multiplier signal 206 in the adder 210 to provide the signal RS at the output of the adder 210.

While the performance of this phase shifter is not completely identical to that of its analog counterpart in Figure 5a, it is sufficiently similar to provide the desired effect. Figures 6a to 6e show graphically the variations of the different matrix coefficients of the decoder of Figure 4 and their improvements that are described by the equations in the preceding section for the description of Figure 4, for further clarification of the operation of this decoder. In Figure 6a, curves A and B represent the variation of the coefficients LL (LFL) and -LR (LFR), respectively, as the value of c / s fluctuates from OdB to approximately 33dB. These curves follow the sine-cosine law as derived in equations (27) and (28). The variation of RR (RFR) and RL (RFL) is similar in shape to the direction in the right frontal quadrant. The curves C and D respectively show the corresponding values of LFL and LFR for the decoder in accordance with the prior US Pat. No. 5,136,650 of Griesinger, for comparison. In these curves, which approximate the value of 0.5 under strong central direction, the component of the music is 3dB too low, therefore the curves A and B of the new decoder, which are at 0.71, provide constant music level , while the old curves do not. Figure 6b shows the curves E and F that represent the central coefficients CL and CR under the direction of 1 / r from the center (OdB) to the left (33dB). The left coefficient CL is increased by 3dB while the right coefficient CR decreases to zero as the direction moves to the left. Similar considerations apply, but in the opposite direction, when the direction is to the right. The curves G and H represent Cl and CR, respectively, in the decoder of the previous Griesinger patent referred to above, which show that, again, the level of the music does not remain constant, since the curve G is not increased by the 3dB that are required. Paying attention to Figure 6c, the curves J and K represent the values of the LSL and LSR coefficients respectively, as the ratio 1 / r goes from OdB (without direction or central direction) to 33dB, which represents the complete left direction . The J curve of LSL is reduced to zero, as it removes the left signal from the left surround sound channel, while the LSR signal increases so that the level of the music remains constant in the room. It is clear from the curves that there is a bankruptcy point at 8dB, which corresponds to a steering angle of 22.5 ° towards the rear. Here the array elements must totalize (in a form r.m.s) 1 when the input has only one directional signal. This is achieved if they have the values of eos 22.5 ° or 0.92 and of sen 22.5 ° or 0.38, as can be seen from the curves. In this context, note that 1 / r can be zero dB either when the signal is completely routed to the back, or when there is no directional component of the signal. In any case, the matrix relaxes towards the complete left-right separation that is desired. In Figure 6d, curve L represents the RBOOST value that was tabulated above in Table 1 and used in equations (76) and (79), and subsequently. The value of LSL is too small when it is directed to the full back, so that the RBOOST value is added to keep the level of the music constant. Only increase LSL, so that the complete separation is maintained. The value of RBOOST depends on c / s, as c / s varies from -8dB to -33dB (full posterior) that is, the x-axis of the graph is -c / s, with c / s in dB. The curve M, which represents the value of RBOOST, is also shown in Figure 6d. In the seven-channel version of the decoder, this value is subtracted from the left lateral coefficient and half of it is added to the left rear component, when it is going from the left back (-8dB) to the full back (-33dB) .

Again, the axis is - (c / s in dB), and this curve goes from zero to 0.5, as expressed in equation (80) above. Finally, Figure 6e shows the curve N, which represents the variation of the correction factor (sin ts + eos ts) with the control signal c / s applied to the surrounding and lateral surround sound channels, to keep constant the level of music, as described above subsequent to equation (39). With attention to Figure 7, an active coder suitable for use in encoding sound tracks of films generally is shown, and in particular with reference to the decoder modes that were presented above. In Figure 7, the same five LAS, L, C, R, and RS signals are applied to the correspondingly numbered terminals 62, 50, 52, 54, 64, respectively, as in the encoder of Figure 3. For each of these signals there is a level detector and a corresponding logarithmic amplifier, to provide signals proportional to the logarithms of the amplitudes of each of these signals. These elements are numbered 212-230. The logarithmic signals are labeled respectively lsl, 11, cl, rl and rsl, which correspond to the inputs LAS, L, C, R and RS. Then, these signal levels are compared in a comparator block (not shown), whose action is described below. The attenuators 254 and 256 attenuate the LAS signal by the factors of 0.53 and 0.83 respectively, and the attenuators 258 and 260 attenuate the RS signal by the factors of 0.83 and 0.53, respectively. Each of the five input signals passes through a phase shift network of all pass, the blocks that are labeled 232, 234, providing the phase shift functions f and f-90 ° respectively for the attenuated signal LAS of the attenuators 254 and 256 respectively, the blocks 236, 238, and 240 providing the phase shift function fa for each of the signals L, C and R, respectively. A signal combiner 242 adds 0.38LS with -0.38RS to provide a central surround sound to the phase shifter block 244, which has a phase shift function f. The phase shifting blocks 246 and 248 provide the phase shifting functions f-90 ° and f respectively, in the RS channel of the attenuators 258 and 260, respectively. A signal combination matrix 250 adds the signal LAS (f) attenuated by attenuator 254, with gain signal, signal LAS (f-90 °) attenuated by attenuator 256, with gain eos? ^, Signal L (f), signal C ( f) with 0.707 gain, and the surrounding sound signal S = (0.38LS-0.38RS) with phase f, which is labeled S (f), to produce the left output signal A in terminal 44. A similar matrix 252 adds the signal RS (f) with the gain sen ®IAS, the signal RS (f-90 °) with the gain TRS eos, the signal R (f), the signal C (f) with 0.707 gain, and the signal of S (f), to produce the right output signal B at terminal 46. The direction angles 0 ^ and TRS are made dependent on the logarithmic amplitude signals lsl, 11, cl, rl and rsl in the following manner in this embodiment of the invention: Whenever lsl is larger than any of the remaining signals, then the © ^ approaches 90 °, otherwise, the 0 ^ approaches 0. These values can be the extre a smooth curve. In a similar way, if rsl is larger than any of the other signals, the TRS approaches 90 °, otherwise, the TRS approaches 0. The particular advantage of this mode of operation is that when a signal is applied to the LAS or RS input only, the encoder output is real, and produces a 1 / r ratio in the decoder of 2.41: 1 (8dB), which is the same value that was produced by the simplified encoder and the passive encoder. With attention to Figure 8, which shows a part of a decoder according to the invention having complex rather than real coefficients in the matrix, the Figure illustrates a method for generating a third control signal the / rs (in addition to the signals 1 / r and c / s generated by the decoder in Figure 4), which is used to vary the additional phase shift network of Figure 9 which is placed in front of the decoder of Figure 4, in order to effect the generation of complex coefficients in the matrix. It will be seen that signals A and B are now applied to terminals 300 and 302, respectively, instead of terminals 92 and 94 of Figure 4. A phase shift network of all step 304 having the phase function f (f) of frequency f, and a second phase shift network of all step 306 having the phase function f (f) -90 °, receive signal A from terminal 300. The phase shifted signal from 304 it is attenuated by a factor of -0.42, in the attenuator 308, and the shifted quadrature phase signal delayed from 306 is attenuated by a factor of 0.91, in the attenuator 310. The outputs of the attenuators 308 and 310 are summed in the adder 312. The signal B at terminal 302 is passed through a phase shift network of all of step 304, such that the output of adder 312 is signal A displaced by 65 °, relative to the signal B at the output of phase shifter 314. The output of adder 312 is passed through the attenuator 316 with an attenuation factor of 0.46, and an input of an adder 318, where this is added to the signal B displaced in phase of the displacer 314. Similarly, the output of the phase shifter 314 is attenuated by means of the attenuator 320, with the same factor of 0.46, and it is passed to the adder 322, where it is added to the output of the adder 312, the signal A shifted in phase. The particular options of coefficients in the attenuators 308, 310, 316 and 320 are made in such a way that the signals applied only to the input LS of the passive encoder, do not produce any output in the adder 308, and a signal applied only to the input RS does not produce any output in adder 322. The goal, therefore, is to design a circuit that recognizes the case as the signal is applied to the left side or the right side of the encoder as the decoder input. It does so by means of a cancellation technique, in such a way that one or the other of the two signals goes to zero when the condition exists. The output of the adder 318 is passed into a level detection circuit 324, and the logarithmic amplifier 326, while the output of the adder 322 is passed through the level detector 328 and the logarithmic amplifier 330. The outputs of the logarithmic amplifiers 326 and 330 are passed to the subtractor 332, which produces an output proportional to its logarithmic proportion. This output can be selected by means of switch 334, or the output of time constant RC, formed by resistor 336 and capacitor 338, which have identical values to the corresponding components shown in Figure 4, can alternatively be selected. , by means of switch 334, and passed to terminal 340 as the address signal ls / rs. In this way the signal ls / rs will be either a maximum positive value when a signal is applied to the input LS of the passive encoder, or a maximum negative value when a signal is applied to the input RS. The purpose of the signal ls / rs is to control the input phases applied to the decoder of Figure 4. For this reason, the network of Figure 9 is interposed between signals A and B applied to terminals 92 and 94 of Figure 4. The circuit shown in Figure 9 includes a phase shifter 342, of phase function f, which it may be the same displacer as 304 in Figure 8, followed by an attenuator 344 having the attenuation value eos TRS, while the phase displacer 346, which may be the same displacer as 306 in Figure 8, of function of phase f-90 °, is passed through the attenuator 348 with the attenuation factor sin TRS. The outputs of the attenuators 344 and 348 are summed by the adder 350, to provide a modified signal A in the terminal 352, which is to be connected directly to the terminal 92 of the Figure. In the lower part of Figure 9, the signal is applied B to terminal 302, as in Figure 8, and in one branch passes through phase shifter 354 of phase function f, and attenuation factor attenuation 356 eos TLS, while in the other branch it passes to through phase shifter 358, phase function f-90 °, and attenuation factor attenuator 360 sin TLS. The signals from the attenuators 356 and 360 are combined in the subtracter 362, to provide a modified signal B in the terminal 364, which is to be connected directly to the terminal 94 in Figure 4. The result in the phase shift is produce a better separation between the LS and RS outputs of the decoder (as well as the LR and RR outputs in a 7-channel version), when only the LS or RS inputs of the passive encoder are being activated with the signals. The relationship between the control signal ls / rs and the TLS steering angle is shown in the inserted graph of Figure 9. As ls / rs reaches 3 dB, the TLS angle begins to change from 0o, rising to 65 ° at high values of ls / rs. The other steering angle TRS applies an exactly complementary relationship, which is controlled by the inverse of ls / rs, which we call rs / ls, in such a way that when rs / ls exceeds 3 dB, the value of TRS begins to increase from 0o, moving towards an asymptote at -65 °, when rs / ls is at its maximum value. Since TLS and TRS vary, the coefficients of the matrix become effectively complex, due to the phase changes in the inputs to the main part of the decoder shown in Figure 4. Figure 10 illustrates an alternative mode of an encoder which differs from that of Figure 7, by simplifying the phase shift networks. The number of phase shift networks can be reduced by combining the actual signals, before sending them through the phase shifter F, thus resulting in only two phase shift networks f and two f-90 ° . The description of TLS and TRS is also simplified. TLS approaches 90 ° when lsl / rsl is greater than 3 dB, and otherwise it is zero (just as in the decoder design). In the same way, TRS approaches 90 ° when rsl / lsl is greater than 3 dB, and otherwise it is zero. In Figure 10, corresponding elements have been numbered corresponding to those in the right half of Figure 7, namely the attenuators 254-260, and the phase shifters f-90 ° 234 and 246. In order to provide a more detailed description of the difference between this encoder and that of Figure 7, the elements of Figure 10 that do not correspond have also been numbered. It will be seen that the coefficients shown in the signal combining elements 242, 250 and 252 of Figure 7 have been extracted from the signal combiners and applied separately to each of the relevant signals in the attenuating elements 262 -274, and that these modified signals are combined into simple summing devices 276-284, while the five phase shifters f 232, 236-240 and 248 have been replaced with two phase shifters 286-288.

In Figure 10, the path of the signal for the signal LS from the terminal 62 of Figure 7, passes as before through the attenuator element 256 and the phase shifter f-90 ° 234, then passing through the attenuator actively controlled 270, which has the attenuation factor eos TLS, this coefficient being the one shown above in the signal combiner 250 of Figure 7. This signal is added in the adder 276 as a component of the signal output labeled A, at terminal 44 of Figure 7. The signal path for signal RS at terminal 64 in Figure 7, similarly passes through attenuator 258 and phase shifter 246, then through active attenuator 274 having the attenuation coefficient eos TRS, formerly part of the signal combiner 252 of Figure 7, to adder 280, where this is a component of the signal labeled B, at terminal 46 of Figure 7. The signal path for the central signal C from terminal 52 of Figure 7, passes first through of the attenuator 266 with the attenuation coefficient 0.71, after which it is applied to the adders 278 and 282. The signal L from the terminal 50 of Figure 7 is applied directly to the adder 278. The signal R from the terminal 54 of Figure 7 is applied directly to the adder 282. The signal LS is also applied through the attenuator 254, and through the active attenuator 268 with the attenuation coefficient sin TLS, to the adder 278. The signal RS is also passed through of the attenuator 260 and of the active attenuator 272 with the attenuation coefficient TRS to the adder 282. Finally, the signal LS passes through the attenuator 262 of coefficient 0.38, and the signal RS passes through the attenuator 264 of coefficient -0.38, both attenuated signs as being added in the adder 284, before the result is applied to the adder 278 with a positive sign, and to the adder 282 with a negative sign. The output of the adder 278 is passed through the phase shifter f 286 to the adder 276, and the output of the adder 282 is passed through the phase shifter f 288 to the adder 280, the adders 276 and 280 providing respectively the signals A and B to terminals 44 and 46 of Figure 7. The examination of the attenuation and the sum of each of the signals LS, L, C, R, and RS in each of the outputs A and B, it will show that these output signals are identical in content to those in Figure 7, but with three less than the expensive phase shifters related to Figure 7. Figure 10 also shows graphically the relationship between the TRS angle and the value of rs / ls (or -ls / rs) for the signals directed in the quadrant of the right side. This angle affects the circuit elements 272 and 274, as indicated by the arrows. There is an exactly similar relationship between the TLS steering angle and the value of ls / rs, this angle affecting the circuit elements 268 and 270.

Turning to Figure 11, an encoder is shown, which is very similar in construction to the encoder of Figure 10. Those elements that can be compared in function are, therefore, correspondingly numbered. There are many new elements, the four gain control elements, the variable attenuators 290-293, and two control signal generating elements 294, 295. The input and output terminals have been numbered in correspondence with Figure 7. 'The The purpose of the added gain control elements is both to correct the balance between strongly directed frontal signals and music, and to reduce the separation in response to the simultaneous signals on the left and right sides. When strongly directed signals occur on the left, center or right channels, the type of decoder compatible with Dolby Pro-Logic, that is, in this case one that meets criterion E, rather than with criterion D, applies a 3 dB increase in the front channels.

^^ This increase is very audible as a shift in the balance between dialogue and music, for example. Typically, when mixing a sound track for playback with a decoder compatible with Dolby, the dialogue recording levels and other front channels strongly directed by means of the sound mixer, which listens to the sound track through a , "Decoder that applies this increase. However, the movies of five channels coded through either a passive encoder or the type of active encoder previously described with reference to Figures 7 and 10, will not be compensated in this way. In the new encoder, the three front signals L, C and R are passed through three variable attenuators 290-292, which respectively have the gain coefficients GL, GC and GR. These coefficients are controlled by means of the direction control signals derived from the outputs of the encoder. To do this, the output signals A and B are fed into the inputs of an address signal voltage generator 294, which comprises a circuit system identical to that of the decoder of Figure 4. Thus the two are derived. 1 / ryc / s address voltages, and they will be identical to those generated in an active decoder. These two steering voltages affect the gain coefficients in a manner shown in Figures 12a and 12b. The signals 1 / r and the inverse r / 1 control the control gains GL and GR, respectively of the elements 290 and 292, while the gain GC of the element 291 is controlled by means of c / s. When 1 / r is positive (left direction), the value of GL is reduced from 1, in accordance with the curve shown in Figure 12a, while the value of GR remains at 1. Similarly, when 1 / r is negative, the value of GR is reduced in accordance with the same curve (in relation to 11 / r |) while the value of GL is constant in 1. Equally, when the front / rear direction c / s is positive (directed to the front), the gain GC varies with c / s, according to the curve of Figure 12a, but GC remains at 1 when the signal c / s is negative. The curve in Figure 12a is the inverse of the curve N shown in Figure 6e. Since signals 1 / r and c / s are generated within a feedback loop, because the change of a gain also affects the steering voltage, the correction applied to each of the front signals will exactly match the increase applied to them in the movie decoder. The result of this is that the effects of dialogue, music and loud sound left or right, maintain the balance of the original discrete mix, when the five original channels are encoded to two, and then decoded back to five or seven channels . There is in fact very little loss of subjective quality when comparing the two-channel version with the original five-channel version. Most of the time, there is no apparent difference at all. A further improvement in the encoder of FIG. 11 is the addition of the gain coefficient GS of the variable attenuator 293, which is controlled by the control voltage generation circuit 295. The gain coefficient GS acts on the signal from the adder 284, which is the difference signal between the input signals on the left and right sides (multiplied by 0.38). The purpose of this side difference signal is to provide the appropriate negative value of the c / s signal when there is a left side or right side input strongly directed to the encoder. However, this lateral difference signal reduces the separation between the left side and right side inputs when both are present at the same time. This reduction in separation is particularly important in the case when the LS and RS inputs are almost equal but are not correlated, such as during music, applause, or surrounding effects such as rain. During these non-targeted effects, we would like to disable the difference signal, and this can be achieved by reducing the GS value as long as there is no strong correlation between the left-side and right-side signals applied to the encoder. The presence of correlation can be determined by the steering voltages derived from the left-side and right-hand inputs to the encoder, using a control voltage generation circuit 295, similar to that in element 294, which thus produces the signals control ls / rs and cs / ss. The address voltage ls / rs was also derived in the original version of the active encoder shown in Figure 7, to control the TLS and TRS values. Although this feature is maintained in the encoder of Figure 11, the additional circuitry determines the front-rear components of the side signals. Both the signals ls / rs and cs / ss control the gain GS of the attenuator element 293. The signal ls / rs also controls the steering angle 1 / s in the attenuators 270 and 272, and its inverse, rs / ls controls the angle of direction trs in the attenuators 272 and 274. Then the value of GS is determined by taking the largest of the absolute values of the signals ls / rs and cs / ss, limiting this value to 7 dB, dividing by 7, then subtracting the result of 1. In this way, any signal with correlation of 7 dB or more, will result in GS = 1, in such a way that the encoder works as before, but when the signals are not correlated to LS and RS , the value of GS will decrease in accordance with the above, and the decoder will revert to the high separation between these inputs. In the process of comparing encoded / decoded signals against uncoded multi-channel sound, it becomes evident that the output from the left front or right front channels was not reduced sufficiently during the lateral direction. In accordance with the Dolby Pro-Logic specification, which does not include the left and right side channels, the left front output of the decoder is reduced in amplitude by only 2.5 dB. This behavior of the front channels is intentional, in order to follow the Dolby specification, but in the standard Dolby specification there are no side channels to decode, and there is only a single rear output. There is, therefore, a need to modify the Dolby specification for the left and right front outputs, during the rear direction when there are side speakers. In the modified specification, the left front and right front outputs are reduced by an additional 3 dB, when there is a rear direction on the same side. In this way the left front signal is reduced by this amount, as a signal pans from the left to the left side, and the right front signal is reduced in a similar manner as a signal pans from the right to the right side. With the side speakers installed, this clearly improves the apparent movement of a signal that moves from the front to either side, and then to the rear; however, this is not a very large deviation that makes much audible difference with a standard Pro-Logic encoded movie. In Figures 12b and 12c, respectively, the variation in gain for the elements of the matrix LL and LR for the lateral direction from left to left is shown. Similar curves apply to the right lateral direction. Another aspect of the decoder improvements is a special limiting correction that can be applied digitally to the directional control signals 1 / r and c / s. This has the advantage of improving the speed and accuracy of the address. During panning from the left to the center of a strongly directed signal, the 1 / r and c / s signals are not independent, but follow a complementary path, shown in Figure 13. If the logarithmic detectors act quickly, it is The curve will be followed dynamically, but when a time constant is included, the value of the signal that is rising can be increased rapidly, but the signal that is falling is usually changing at a slower speed. The result is that the signal that falls is higher than it should be, reducing the dynamic separation. To correct this problem, the signal that is changing most rapidly is used to limit the other signal, to follow the curve of Figure 13. Although some prior art decoders include circuits to limit the control voltage excursions, during changes fast, these circuits were not based on the speed of change of the control signals, but rather on their absolute values. A particular advantage of the rate of change method is that the increasing signal is enabled to rise rapidly, while the falling signal, which represents the direction of the matrix in a previous direction, is forced to yield to the signal that changes more quickly Remembering the definitions of these signals, it is very easy to find the relationships between the control signals, which can only occur as a maximum limit, which will not be achieved in the presence of decorrelated music. If we consider a panning from the left to the center, L = eos and R = sin t, then the control signals are 1 / r = 20 * log10 (cos t / sin t) ... (98) c / s = 20 * log10 ((cos t + sin t) / (cos t - sin t)) ... (99) These relationships are plotted against each other in Figure 13. Figure 14 shows for reference the form of the coefficient of the left-rear-left matrix element that is used in the matrix of Figure 4, as implemented in the decoder, in accordance with the above patent application No. 08 / 684,948 . The value of this coefficient is plotted in three-dimensional format as the height of the coordinates with respect to the control signals cs and Ir, which are derived from the usual logarithmic ratio detectors in the decoder of Figure 4. The signal cs represents the ratio of the amplitudes of the signals surrounding the front to the back, and the signal Ir represents the proportion of the amplitudes of the signals left to right. Each of these signals is coded as an angle that varies from zero to 90 degrees. As you can see in the illustration, there is a discontinuity in the value of this element near the left vertex, forming a small "crag" as the signal moves towards the back. In this representation there is also a central "valley". Similarly, Figure 15 shows the value of the left-rear-right matrix element, which has a similar discontinuity along the axis lr = 0 towards the rear. These discontinuities are due to an error in the theory described in the previous patent application. The problem is that there is a correction applied to the matrix element in the left rear quadrant, which was made by means of a table indexed through a variable called Circumscribed Go. It turns out that this type of correction only works when the error that is being corrected is symmetrical. The correction along the cs axis is symmetric, but the correction along the limit lr = 0 is not. A better way to do this correction is through an interpolation, as was done in the theory in the previous document for the LRR matrix element, along the Ir axis. In practice, interpolation can be done through two additional search tables, at little computational cost. Subsequently, the equations for this interpolation will be shown. The matrix element LRR is correctly interpolated along the Ir axis. However, as can be seen in Figure 15, there is a discontinuity along the cs axis. This discontinuity is due to a programming error in the decoder of version 1.11. The theory presented in the previous patent application produces the matrix element shown in Figure 16. Figure 16 shows the left-rear-right matrix element as it should have been implemented in the decoder of version 1.11, of conformity with the previous patent application, in which the discontinuity has been removed. With the correct interpolation and implementation, there are no discrepancies in the limit lr = 0. Turning now to the TV sound matrix, Figure 17 shows the LFL matrix element, as implemented in U.S. Patent No. 4,862,502, and in Dolby Pro-Logic, in scale, such that the value maximum is 1. Assuming that cs and Ir are the addresses of the direction in degrees, in the central / surrounding and left / right axes respectively, in the 1989 patent the equations for the front matrix elements were given as: In the quadrant left front, LFL = 1 - 0.5 * G (cs) + 0.41 * G (lr) ... (100) LFR = -0.5 * G (cs) ... (101) In the right front quadrant, LFL = 1 - 0.5 * G (cs) ... (102) LFR = - 0.5 * G (cs) ... (103) In the left rear quadrant, LFL = 1 - 0.5 * G (-cs) + 0.41 * G ( lr) ... (104) LFR = 0.5 * G (-cs) ... (105) In the right rear quadrant, LFL = 1 - 0.5 * G (-cs) ... (106) LFR = 0.5 * G (-cs) ... (107) The G (x) function is described in U.S. Patent No. 4,862,502, and is specified in the State's patent. United States No. 5,307,415. This varies from 0 to one, as x varies from 0 to 45 degrees. In the previous patent application it is shown that this is equal to 1-tan (| r / l ¡1¡), where | r / and \ l are the right and left input amplitudes. In Figure 18, the LFR matrix element of U.S. Patent No. 4,862,502 and Dolby Pro-Logic, in a scale of 0.71, is shown in such a way that the minimum value and the maximum values are ± 0.5. In the above patent application No. 08 / 684,948 these elements were improved by adding the requirement that the sound intensity of the non-directed material should be constant, regardless of the direction of direction. Mathematically this means that the mean square root of the matrix elements LFL and LFR must be a constant. In the request it was noted that this goal should be relaxed in the direction direction - that is, when the direction is completely to the left, the sum of the squares of the matrix elements should be raised by 3 dB. We can see that the previous matrix elements do not meet this requirement. In Figure 19, the square root of the sum of the squares of LFL and LFR of U.S. Patent No. 4,862,502, in scale, is shown in such a way that the maximum value is one. Note that the value is constant at 0.71 along the axis of the non-directed to the right. The unguided to the left rises 3 dB to the value one, and the unguided to the center or to the rear falls by 3 dB to the value 0.5. The rear direction is identical to the central direction, but it is not easily seen due to the perspective in this view. The earlier patent application No. 08 / 684,948 corrected this amplitude error by replacing the function G (x) in the equations of the matrix with sines and cosines: For the left front quadrant LFL = cos (cs) + 0.41 * G (Go) ... (108) LFR = -sen (cs) ... (109) For the right frontal quadrant LFL = eos (cs). . . (110) LFR = - sin (cs) ... (111) For the left rear quadrant LFL = cos (-cs) + 0.41 * G (Go) (112) LFR = sin (-cs) (113) For the right rear quadrant, LFL = eos (-cs) (114) LFR = sin (-cs) (115) Figure 20 shows the square root of the sum of the matrix elements LFL and LFR of the previous patent application No. 08 / 684,948, in scale, such that the maximum value is 1. Notice the constant value of 0.71 in the entire right half of the plane, and the smooth elevation to one toward the left vertex. The decoder of version 1.11 made many changes to these matrix elements. Maintaining basic functional dependence, an additional increase was added along the cs axis at the front, and a cut was added along the cs axis at the rear. The reason for the increase was to improve the operation with stereo music that was panned forward. The purpose of the cut in the back was to increase the separation between the front channels and the rear channels when the music was played in stereo to the rear. For the left front quadrant, LFL = (eos (cs) + 0.41 * G (lr)) * boostl (cs) ... (116) LFR = (-sen (cs)) * boostl (cs) ... ( 117) For the right front quadrant, LFL = (eos (cs)) * boostl (cs) ... (118) LFR = (-sen (cs)) * boostl (cs) ... (119) For the quadrant left back, LFL = (cos (-cs) + 0.41 * G (lr) / boost (cs) ... (120) LFR = (sin (cs)) / boost (cs) ... (121) For the right rear quadrant, LFL = (eos (cs)) / boost (cs) ... (122) LFR = (sin (cs)) / boost (cs) ... (123) In the previous patent application it was defined the function G (x) .This is equal to G (x) = l-tan (45-x), and is identical to the function that is used in the Dolby matrix.The boostl (cs) function as used in the The decoder of version 1.11 was a linear gain of 3 dB total, applied over the first 22.5 degrees of direction, decreasing back to 0 dB in the following 22.5 degrees. (See run in the pseudocode later.) Boost (cs) is given by corr (x) in the subsequent code where the comments are preceded by the symbol%:% calculate a function of increase from +3 dB to 22.5 degrees% corr (x) goes up to 3 dB and stays up, ran (x) goes upwards and then towards down again for x = 1:24; % x has values from 1 to 24 corr (x) = 10? (3 * (x-1) / (23 * 20)); % goes up to 3 dB over this range I ran (x) = corr (x); end for x = 25:46; % goes back again to run on this range corr (x) = 1.41; I ran (x) = corr (48-x); end These equations produce the surface shown in Figure 21 for the matrix element LFL in the decoder of version 1.11. Note that the increase as the direction moves toward the center, applies both along the axis lr = 0, and along the limit from left to center. Note also the reduction in level as the direction moves to the rear. The increase along the limit creates a panning error. In addition, the cut in the rear direction is not optimal. There are two areas where you can improve the operation. The first is in the behavior of the direction along the boundaries between the left and the center, and between the right and the center. As a strong single signal panneates from left to center, in Figure 21 it can be seen that the value of the matrix element LFL is increased to a maximum mean path between the left and the center. This increase in value is an unintended consequence of the deliberate increase in the level of the left and right main outputs, as the central signal is added to the music in stereo. As explained in a previous description, when forwarding a stereo signal, it is desirable that the left and right front outputs are raised in level, to compensate for the removal by the matrix of the correlated component of these outputs. However, the method used to increase the level under these conditions should only occur when the Ir component of the inputs is minimal-that is, when there is no net left or right direction. However, the method selected to implement this increment in the decoder of version 1.11 was independent of the value of Go, and resulted in an increase in level, when a strong signal was panned across the boundary. The problem is that the increase is only needed along the axis 2r = 0. When Ir is non-zero, the array element must not be increased. This problem can be solved by using an additive term to the matrix elements, instead of a multiplication. We define a new address index, the value cs limited by the limit, with the following code: We assume that both Go and cs > 0 for a signal in the left front quadrant, also assuming that cs and Ir follow the Matlab conventions of variation from 1 to 46; % find the c / s limited if (cs < 24) ibes = cs- (lr-1); if (£> cs < l)% this limits the maximum value is = 1; end also ¿cs = 47-cs- (lr-1); yes (bes < 1) ibes = 1; end order If cs < 22.5 and Ir = 0, (in the Matlab convention cs <24 and Ir = 1) bes equals cs. However, as Ir increases, bes decreases to zero. If cs > 22.5, as Ir increases, bes also decreases. Now, to find the correction function that is needed, we find the difference between the increased and non-augmented matrix elements, along the axis lr = 0. We call this difference cos_tbl_plus and sin_tbl_plus. (This code is written in a modified Matlab, where the variables are multivalued vectors, the comments are preceded by the% symbol). a = 0:45% define a vector in steps of a degree, a has the values of 0 to 45 degrees al = 2 * pi * a / 360; % convert to radians% now define the sine and cosine tables, as well as the increase tables for the front sin_tbl = sin (al); cos_tbl = eos (al); cos_tbl_plus = eos (al). * corrl (a + 1); cos_tbl_plus = cos_tbl_plus-cos_tbl; % This is the one we use cos_tbl_minus = eos (al). / corr (a + 1); sin_tbl_plus = sin (al). * corrl (a + 1); sin_tbl_plus = sin_tbl_plus-sin__tbl; % This is the one we use sin_tbl_minus = sin (al). / corr (a + 1); sin_tbl_plus and cos_tbl_plus are the difference between a simple sine and cosine, and the sine and cosine increased. Now we define LFL = eos (cs) + 0.41 * G (lr) + cos_tbl_plus (bes) ... (124) LFR = -sin_tbl_plus (bcs) + sin (cs) ... (125) LFL and LFR in the frontal right quadrant are similar, but without the term + 0.41 * G. These new definitions lead to the following matrix elements. The direction in the back quad is not optimal either.

In the previous curve, when the direction is towards the back, the matrix elements are given by LFL = cos_tbl_minus (-es) + 0.41 * G (-cs) ... (126) LFR = sin_tbl-minus (-cs) ... (127) These array elements are very closely identical to the Dolby array elements. Consider the case when a strong signal pans from the left to the back. The Dolby elements were designed in such a way that there is a complete cancellation of output from the left front exit, only when this signal is completely to the rear. However, in a decoder according to the present invention, it is desirable that the output from the left front output is zero when the encoded signal reaches the left rear sense (cs = -22.5 and Ir = 22.5). The left front output must remain at zero as the signal pans in addition to the full rear. The matrix elements used in the decoder of version 1.11 - the previous ones - result in the output in the left front channel being about -9 dB, when paging a signal to the left rear position. This difference in level is sufficient for the proper functioning of the matrix, but it is not as good as it could be. This operation can be improved by altering the matrix elements LFL and LFR in the left rear quadrant. Notice here that we are concerned about how the matrix elements vary along the boundary between the left and the back. You can use the mathematical method given in the AES document ("Multichannel Matrix Surround Decoders for Two-Eared Listeners", David Griesinger, pre-print AES No. 4402, October, 1996), to find the behavior of the elements to along the limit. Let us assume that the amplitude of the left front output should decrease with the function F (t) as t varies from 0 (left) to -22.5 degrees (left back). The method gives the matrix elements: LFL = cos (t) * F (t) - / + sin (t) * (sqrt (lF (t)? 2)) ... (128) LFR = - (sin ( t) * F (t) +/- eos () * (sqrt (1-F (t) ~ 2))) ... (129) If we choose F (t) = cos (4 * t) and choose the correct sign, these are simplified to: LFL = cos (t) * cos (4 * t) + sin (t) * sin (4 * t) ... (130) LFR = - (sin (t) * cos ( 4 * t) -eos (t) * sin (4 * t) ... (131) These elements work well-the left front output is gently reduced to zero, as t varies from 0 to -22.5 degrees. Figure 22 shows that the left-front-left matrix element has the correct amplitude along the left to center boundary, as well as along the center boundary to the right.We want the output to remain at zero as the direction continues from 22.5 degrees to 45 degrees (full right) Throughout this part of the limit, LFL = -sen (t) ... (1'32) LFR = cos (t) ... (133) Note that these matrix elements are a far cry from the matrix elements at length of the limit lr = 0, where in the previous patent application the values were LFL = eos (cs) ... (134) LFR = sin (cs) ... (135) Figure 23 shows the behavior of LFL and LFR along the rear limit between the left and full back. The slight damage is due to the absence of a point at 22.5 degrees. We need a method to substantially transform the previous equations to equations along the boundary as Ir and cs approach the boundary. A linear interpolation can be used. In the processor that is typically used in these decoders, where multiplications are expensive, a better strategy is to define a new variable - the minimum of Ir and cs: new% - find the limit parameter bp = x; yes (bp > y) bp = y; end and a new correction function that depends on bp: for x = 1:24 ax = 2 * pi * (46-x) / 360; front_boundary_tbl (x) = (eos (ax) -sen (ax)) / (eos (ax) + sin (ax)); end for x = 25:46 ax = 2 * pi * (x-l) / 360; front_boundary_tbl (x) = (eos (ax) -sen (ax)) / (eos (ax) + sin (ax)); Then we define LFL and LFR in this quadrant as: LFL = eos (cs) / (eos (cs) + sin (cs)) - front-boundary-table (bp) + 0.41 * G (lr) ... (136 ) LFR = sin (cs) / (eos (cs) + sin (cs)) + front_boundary_table (bp) ... (137) Note the correction of eos (cs) + sin (cs). When we divide eos (cs) between this factor, we obtain the function 1 -0.5 * G (cs.), Which is the same as the Dolby matrix in this quadrant. In the right rear quadrant, LFL = eos (cs) I (eos (cs) + sin (cs)) ... (138) LFR = sin (cs) / eos (cs) + sin (cs)) ... (139) Figure 24 shows the left-front-left matrix element, as seen from the left rear. Note the great correction along the left-back limit. This causes the front left output to go to zero when the direction goes from the left to the left rear. The output remains at zero as the direction progresses to the full rear. However, along the r = 0 axis, and in the right rear quadrant, the function is identical to the Dolby matrix. Figure 25 shows the left-front-right matrix element. Note the large peak in the limit from the left to the back. This works in conjunction with the LFL array element, to reduce the front output to zero, along this limit, as the direction goes from the left rear to the full rear. Again, in the rear direction along the axis 2r = 0, and in the right rear quadrant, the element is identical to the Dolby matrix. One of the major design goals of the improved matrix design of the present invention is that the intensity of the sound at any given output of the unmanaged material presented to the decoder inputs should be constant, regardless of the direction of a signal directed that is present at the same time. As explained above, this means that the sum of the squares of the array elements for each output must be one, regardless of the direction of the address. As explained above, this requirement should be relaxed when there is a strong direction in the direction of the exit in question, that is, if we are looking at the left front exit, the sum of the squares of the matrix elements should be increased by 3. dB when the direction goes completely to the left.

We can prove the success of our design by plotting the square root of the sum of the squares of the matrix elements. Figure 26 shows the sum of the mean square root of LFL and LFR, using the new design. (For this graph we delete the correction 1 / (sin (cs) + cos (cs)) in the back quadrant, so that we can see how exactly the resulting sum arrives at the unit). Note that the 3 dB peak in the left direction, and the somewhat smaller peak as a signal, goes from unguided to 22.5 degrees in the central direction. This peak is the result of the deliberate increase of the left and right outputs during the middle frontal direction. Note that in the other quadrants the sum r s is very close to one, as was the design attempt. The value in the left rear quadrant is not very similar to one, since the method that was used to produce the elements is an approximation, but the correspondence is very good. Figure 27 shows the square root of the sum of the squares of LFL and LFR, including the correction of the back level, seen from the left rear. The unguided axis (half) to the right has the value one, the central vertex has the value 0.71, the back vertex has the value 0.5, and the left vertex has the value 1.41. Note the peak along the axis from the middle to the center. The next concern to which attention is given in the present invention is to correct the values of the back matrix elements, during the frontal direction. The rear matrix elements in the United States Patent No. 4,862,502 of Griesinger 1989 are given by: For the left front quadrant: LRL = .71-.71 * G (Ir) +. 41 * .71 * G (cs) ... (140) LRR = -.71 * G (lr) +. 41 * .71 * G (cs) ... (141) For the left rear: LRL = .71-.71 * G (lr) + .41 * .71 * G (-cs) ... (142) LRR = .71 * G (lr) +. 41 * .71 * G (cs) ... (143) (The right half of the plane is identical, but it switches LRL and LRR). The back matrix elements in the Dolby Pro-Logic are For the left front quadrant: LRL = .71-.71 * G (lr) + .41 * .71 * G (cs) ... (144) LRR = - .71 * G (lr) +. 41 * .71 * G (cs) ... (145) For the left rear: LRL = .71 - .71 * G (lr) ... (146) LRR = .71 * G (Go) ... (147) (the right half of the plane is identical but switches LRL and LRR). A brief digression about the surrounding level in Dolby Pro-Logic - note that the Dolby elements are identical in the front, but do not include the dependent increase in the cs in the back. This difference is in fact very important. The above equations somehow disguise the way in which these decoders are actually used. We de all matrix elements with a relatively arbitrary scale. In most cases the elements are presented as if they had a maximum value of 1.41. In fact, for technical reasons, the matrix elements are all scaled eventually, in such a way that they have a maximum value of one. In addition, when the decoder is finally put to use, the gain of each output is adjusted to the speaker, in such a way that the sound power is the same in the listening position when a signal is presented to the decoder, which has been coded from the four main directions -left, central, right, and back. In practice, this means that the actual level of the array elements is scaled, such that the four decoder outputs are equal under full address conditions. The lack of an increase of the level in the rear direction in the Dolby decoder means that during the calibration procedure the gain of the rear outputs will rise by 3 dB in relation to the other outputs. In fact, for the Dolby decoder in practice: LRL = 1 - G (lr) ... (148) LRR = -1 ... (149) The difference is not trivial. When the frontal elements are scaled such that they have a maximum value of one, when there is complete direction in one direction, we find that during non-directed conditions, the elements of the 1989 patent have the value of 0.71, and the sum of the squares of the elements has the value of one. This is not true of the Dolby rear elements when calibrated. LRL has the unaddressed value of one, and the sum of the squares is 2, or 3 dB higher than the outputs of the 1989 patent. Note that the calibration procedure results in a matrix that does not correspond to a passive matrix of "Dolby Environment" when the matrix is not directed. The passive matrix of Dolby Environment specifies that the rear output must have the value of .71 * (Ain + Bin), and the Pro-Logic matrix does not meet this specification. If there are two speakers sharing this output, each one will be 3 dB softer, which will cause the five speakers to have approximately equal sound power when the decoder inputs are not correlated. When the matrix elements of the 1989 patent are used, the same calibration procedure results in a sound power of 3 dB less from the rear, when the decoder inputs are not correlated. The elements for the rear outlets in the new design include a form of level increase, when the outputs are fully directed - either to the left or right sides - or completely to the rear. In this way, they follow the patent of 1989 in terms of their surrounding level when they are not addressed. To see the importance of this issue, consider what -HO-what happens when we have an input to the decoder, which consists of three components, a left and right uncorrelated component, and a separate and uncorrelated central component. Ain = Lin + .71 * Cin ... (150) Bin = Rin + .71 * Cin ... (151) When playing Ain and Bin in stereo, the sound power in the room will be proportional to Lin? 2 + Rin? 2 + Cin? 2. If the three components have approximately equal amplitudes, the ratio of the central component to the left component plus the right component will be 1: 2. We would like our encoder to reproduce the sound power in the room with approximately the same power ratio as the stereo, regardless of the power ratio of Cin to Lin and Rin. We can express this mathematically. Essentially, the equal power ratio requirement will specify the functional form of the central matrix elements along the cs axis, if all the other matrix elements are taken as given. If we assume the elements of Dolby matrix, calibrated in such a way that the power of the rear sound is 3 dB lower than the other three outputs, when the matrix is completely directed. that is, that the non-directed condition of the matrix is identical to the Dolby Environment, then the central matrix elements must have the forms described by means of Figures 28 and 29.

In Figure 28, the solid curve shows the value of the central matrix as a function of cs + 1 in dB, assuming proportions of sound power identical to the stereo, and using Dolby matrix elements with power of 3 dB less in the part back of what is typically used. The dotted curve is the real value of the central matrix elements in Pro-Logic. Note that although the actual values give reasonable results for a non-directed signal and a fully-directed signal, these are approximately 1.5 dB too low in the middle. Similarly, in Figure 29 the solid curve shows the value of the central matrix elements, assuming equal power proportions to the stereo, given the matrix and calibration elements currently used in Dolby Pro-Logic. The dotted curve shows the actual values of the central matrix elements in Pro-Logic. Note that the actual values are more than 3 dB too low for all directions. These two figures show something that mixing engineers are often aware of - namely, that a mix prepared for playback in a Dolby Pro-Logic system may often need more central sound intensity, than a mix prepared for stereo playback. Conversely, a mix prepared for stereo will lose vocal clarity when played over a Pro-Logic decoder. Ironically, this is not true of a Dolby Environment decoder, which is identical to the non-directed condition of the previous figure. The present invention also includes the creation of two independent rear outputs, as described below. The biggest problem with both the elements of the 1989 patent and Dolby elements is that there is only a single rear exit. The description given in US Patent No. 5,136,650 of 1991 by Griesinger created two independent outputs on the sides, and the mathematics in that patent was incorporated in the left front quadrant of United States Patent Application No. 08. / 684,948 July 1996. The goal of the elements in this quadrant was to eliminate the output of a signal directed from the left to the center, while maintaining some output from the left rear channel for the unmanaged material present at the same time. To achieve this goal we assume that the LRL matrix element should have the following form: For the left front quadrant: LRL = 1 - GS (Ir) - 0.5 * G (cs) ... (152) LRR = - 0.5 * G (cs) - G (lr) ... (153) As can be seen, these matrix elements are very similar to those of the United States Patent No. 4,862,502 from 1989 by Griesinger, but with the addition of a term G (lr) in LRR and a term GS (Ir) in LRL. G (lr) was included to add signals from the input channel B of the decoder to the left rear output, to provide some unaddressed signal strength as the directed signal was removed. Then we solved the GS (Ir) function, using the criterion that there should not be any signal output with a completely directed signal moving from left to center. The formula for GS (Ir) turned out to be equal to G? 2 (Ir), although a more complicated formula is given in the 1991 patent (5,136,650). It can be shown that the two formulas are identical. In the July 1996 application these elements are corrected by having given an increase of (sin (cs) + cos (cs)) to make them closer to the constant sound intensity for the non-directed material. Although it is completely successful in the right frontal quadrant, the correction is not very successful in the left frontal quadrant. For the right front quadrant, the matrix elements are identical to the back elements given in the 1989 patent (4,862,502), and were implemented in the decoder of version 1.11. In Figure 30 you can see the problems of the left frontal quadrant, which shows the square root of the sum of the squares of LRL and LRR, using the matrix elements implemented in the decoder of version 1.11 as above. Note that in the left frontal quadrant, there is a 3 dB slope along the line from the middle to the left vertex, and an almost 3 dB increase in the level along the boundary between the left and the center. Subsequently the mountain range in the rear quadrant will be discussed. This drawing includes the decline of 3 dB "TV matrix" in the middle of the plane, which is difficult to see in this projection. First we consider the decline in the sum of the squares along the axis cs = 0. This decline exists because of the use of G (lr) in LRR. This choice was entirely arbitrary - although this makes the implementation of the analog circuit system easy. Ideally, we would like to have a function GR (Go) in this equation, and choose GS (Go) and GR (Go) in such a way as to keep the sum of the squares of LRL and LRR constant along the axis cs = 0, and keep the zero output along the boundary between the left and the center. This can be done. We would also like to be sure that the matrix elements are identical to the matrix elements in the right front quadrant along the axis lr = 0. Therefore, we assume that: LRL = eos (cs) - GS (Ir) ... (154) LRR = -sen (cs) - GR (Ir) ... (155) We want the sum of the squares to be one along the axis cs = 0, such that (l-GS (Ir))? 2 + (GR (lr)) ~ 2 = 1 ... (156) and the output must be zero at a signal directed, or as t varies from zero to 45 degrees, LRL * cos (t) + LRR * sin (t) = 0 ... (157) These two equations result in a disordered quadratic equation for GR and GS, which it can be solved numerically. Figure 31 shows the numerical solution for GS and GR for a constant power level along the cs = 0 axis, and a zero output along the boundary between the left and the center. The use of GS ,. and GR as shown, results in a great improvement along the axis lr = 0, as intended. However, the peak remains in the sum of the squares along the boundary between the left and the center. In a practical design it is probably not very important to compensate for this error, but we can try to do it with the following strategy. We will divide both matrix elements between a factor that depends on a new combined variable based on Ir and cs. Let's call the new variable xymin. (In practice the division can be replaced by a multiplication by the inverse of the factor described later). A procedure to define xymin (in Matlab notation) is:% find the minimum of x or y xymin = x; yes (xymin > y) xymin = y; end yes (xymin > 23) xymin = 23; end Note that xymin varies from zero to 22.5 degrees. If we multiply it by four, it will vary from zero to 90 degrees, and it can be used later. In the left front quadrant LRL = (cos (cs) - GS (Go)) / (l + .29 * sin (4 * xymin)) ... (158) LRR = (-sen (cs) - GR (Go) ) / (l + .29 * sin (4 * xymin)) ... (159) 'In the right frontal quadrant, LRL = cos (cs) ... (160) LRR = -sen (cs) ... ( 161) As explained in the previous document, these elements are further multiplied by the correction of "TV matrix", which reduces the amplitude when the direction is near the middle of the plane. We will call the correction for the matrix TV tvcorr (| Ir | + I cs |). Tvcorr (| Ir | + | cs |) is -3 dB at zero, and 1 when the argument is 22.5 degrees and higher. Figure 32 shows the square root of the sum of the squares of LRL and LRR using the new values for GR and GS. This factor is shown in Figure 32 as a valley centered in the zero direction. Note that, except for the valley created by the "TV matrix" correction, the sum of the squares is close to one and continuous. In the present invention, the correction of "TV matrix" has been modified to depend only on the absolute value of Ir when cs is frontal. This will cause the top surface to remain at .71 along the axis lr = 0 on the front. In this case the correction for the TV matrix becomes tvcorr (| Ir |). Tvcorr (| lr |) is -3 dB at zero, and 1 when the argument is 22.5 degrees and higher. The rear matrix elements were discussed previously during the rearward direction, with reference to Figures 14-16. The back matrix elements given in the 1991 patent (5,136,650) were not suitable for a 5-channel decoder, and were heuristically modified in the Lexicon CP-3 product. The patent application of July 1996 (08 / 684,948) used a mathematical method to derive these elements along the boundary of the left rear quadrant. As described in the previous document, this procedure resulted in discontinuities along the axis lr = 0, and along the axis cs = 0. In the decoder of version 1.11, these discontinuities were repaired (mostly) by means of additional corrections to the matrix elements, which retained their behavior along the direction limits. Previously in this application the software error was shown in the LRR element, as well as the small discontinuity along the limit cs = 0 for the LRL element (see Figures 14, 15). For the new elements described here these errors have been corrected, first by using an interpolation along the limit cs = 0 for LRL, where the value is made to match the value of GS (Ir) when cs is zero, and it rises gently to the value given by the previous mathematics as it increases cs negatively towards the back. In the new software, LRR is interpolated along the cs = 0 to GR (Ir) axis. In the decoder of version 1.11 LRR is interpolated to G (lr). First we will consider the Left-Back-Left and Left-Back-Right matrix elements when the direction is neutral, or anywhere between the full right and the right rear part. That is, Ir may vary from 0 to -45 degrees, and cs may vary from 0 to -22.5 degrees. Under these conditions the directed component of the input of the left outputs must be removed - there should be no output from the left rear channel when the direction is to the right or to the rear right. The matrix elements given in the 1991 patent (5,136,650) achieve this goal. These are essentially the same as the back matrix elements in the 4-channel decoder, with the addition of the sin (cs) + eos (cs) correction for the non-directed sound intensity. When this is done, the matrix elements are simple. We will define two new functions that are simply equal to the sine and cosine of cs over this range. LRL = eos (-es) = sri (-cs) ... (162) LRR = sin (-cs) = sric (-cs) ... (163) To complete LRL and LRR over the range of cs = 0 to -22.5, we must add a gain reduction for the "TV matrix" mode. Once again, in the "TV tint" mode we want 3 dB less output when the address is neutral, but raising the value in the decoder of the "Logic 7" version, when the address is more than 22.5 degrees the back part. The operation is improved in some way by means of making this reduction sensible to the sum of | Ir | and | cs |. This is achieved in the current design by reducing both RRR and RRL elements by 3 dB, when the sum is zero, and raising them back to their original values, as the sum goes to 22.5 degrees. Again, the tilt of this gain change is relatively arbitrary, as long as both RRR and RRL are altered in the same way. We can call the correction for the tv Matrix TV Matrix (| I | + | cs |). Tvcorr (| Ir | + | cs |) is -3 dB at zero, and 1 when the argument is 22.5 degrees and higher. LRL = eos (-cs) * tvcorr (| lr | + | cs |) = sri (-cs) * tvcorr (| Ir | + | cs \) ... (164) LRR = sin (-cs) * tvcorr (| lr | + | cs |) = sric (-cs) * tvcorr (| Ir \ + | cs |) ... (165) Notice that we have defined a new function sric (x), which is equal to sin ( x) over the range of 0 to 22.5 degrees, and sri (x) which is equal to cos (x). We will use these functions again to define the Left Back matrix elements, during the left direction. Now let's consider the same matrix elements as cs becomes larger than -22.5 degrees. As mentioned in Griesinger's previous AES document, and in the patents and applications cited above, LRL must rise to one or more over this range, and LRR must decrease to zero. Simple functions comply with this: (remember that cs is negative) LRL = (cos (45 + cs) + rboost (-cs)) = (sri (-cs) + rboost (-cs)) ... (166) LRR = sin (45 + cs) = sric (-cs) ... (167) The Left Left matrix elements are now complete during the right direction. The behavior of the Left-Rear-Left and Left-Rear-Right elements is much more complex. The Left Rear Left element must rise rapidly from zero to almost the maximum, as Ir decreases from 45 to 22.5 or to zero. The matrix elements given in the appendix application of November 1996 do this, but as we showed earlier, there are problems with continuity in the limit cs = 0 (see Figure 15). For the decoder of version 1.11, a solution was found that uses the functions of one variable and many conditional ones. In the Griesinger AES document and in the patent applications cited above, the problem at the limit cs = 0 arises because on the front side of the limit the matrix element LRL is given by GS (Ir). On the back side the function given by the AES document has the same endpoints, but it is different between them. The mathematical method in the AES document provides the following equations for the Left Back matrix elements over the range of 22.5 <; Go < Four. Five; (remember that t = 45 -Ir). LRL = cos (45-lr) * sin (4 * (45-lr)) -sen (45-lr) * cos (4 * (45-lr)) = sra (Go) ... (168) LRR = - (sin (45-Ir) * sin (4 * (45-lr)) + cos (45-lr) * cos (4 * (45-lr))) = -srac (lr) ... (169) If cs < = 22.5, Ir can still vary from 0 to 45. The AES document defines LRL and LRR when the Ir has the range 0 < Go < 22.5, as shown in Figure 6 in the AES document. LRL = eos (Ir) = sra (lr) ... (170) LRR = -sen (lr) = -srac (lr) ... (171) Define the two functions sra (x) and srac (x) for 0 < Go < 45. In the decoder of version 1.11 the following technique is used to set the discontinuity across the limit cs = 0. In the AES document near cs = 0, LRL and LRR are both functions of a single variable. To fix the lack of continuity along the limit cs = 0, we add a function of a compound of Ir and cs. The new variable is lr_bounded, the limited difference between Ir and cs. The definition of this variable is complicated enough, so I will present it in a pseudo-c form (MATLAB). lr_bounded = Ir - cs; % find the difference if (lr_bounded < 0) %% only if Go > cs lr_bounded = 0; if (45- | cs | < lr_bounded)% use the smaller of the two values lr_bounded = 45-cs: We define a new function that is equal to the difference between the previous equations when cs = 0. This is rear_active_correct (lr_bounded). For 0 < x < 45 Rear_active_correct (x) = sra (x) - (l-GSL (x)) ... (172) LRL = (sri (cs) + sra (Go) -rear_active_correct (lr_bounded) -l) * tvcorr (| I | + \ cs \) ... (173) The important point about this method is that it works when Go < 22.5, but this one does not work when Go is bigger. A better technique, which was not used in the decoder of version 1.11, is the interpolation technique that is used to LRR. The LRR coefficient of the decoder of version 1.11 uses a better technique. Here are two dicontinuities. Along the limit cs = 0 LRR at the rear must match the LRR for the forward direction, which shows LRR = -G (Ir) along the limit cs = 0. The choice that is used in version 1.11 of the "logic-7" decoder -although in some way it is computationally intensive- is to use an interpolation based on the value of cs over the range of 0 to 15 degrees. In other words, when cs is zero, we use G (Ir) to find LRR. As cs increases to 15 degrees, we interpolate to the value of srac (Go). There is also the possibility of a discontinuity along the axis lr = 0. We can solve this by adding a term to LRR, which is found by using cs_bounded. The term is simply sric (cs_bounded). This term will ensure continuity through the axis lr = 0. First we define cs_bounded cs_bounded = Irs; if (cs_bounded < l)% this limits the maximum value cs_bounded = 0; end if (45- | Ir | < cs_bounded)% use the smaller of the two values cs_bounded = 45-lr; end for cs = 0 to 15 LRR = (- (srac (lr) + (srac (lr) -G (lr)) * (15-cs) / 15 + sric (cs_bounded)) * tvcorr (| Ir | + | cs I); for cs = 15 to 22.5 LRR = (- (srac (lr) + (sric (cs_bounded)) * tvcorr (| Ir | + | cs |); In the decoder according to the present invention LRL is calculated with interpolation, exactly as for LRR, for cs = 0 to 15 LRL = (- (srac (lr) + (sra (Ir) -GS (Ir)) * (15-cs) / 15) + sri (-cs) ) * tvcorr (| Ir | + | cs I); for cs = 15 to 22.5 LRL = (- (sra (lr) + (sri (-cs)) * tvcorr (| lr | + | cs |); that the direction goes from the left rear to the full rear, the elements follow the dice in the AES document, with the addition of the corrections for the rear sound intensity For cs> 22.5, Go <22.5 LRL = (sra (Go) + sri (cs) + rboost (cs)) ... (174) LRR = -srac (lr) + sric (cs_bounded) ... (175) This completes the matrix elements LRL and LRR during the left direction, the values for the dir Right ection can be found by means of exchanging left and right definitions. The following improvements to be discussed are the implementation of the central array elements in the present invention. The central matrix elements in the decoder of version 1.11 have major differences from the core elements in the patent application of July 1996. The patent '89 and Dolby Pro-Logic have the following matrix elements: For the front direction CL = 1 + .41 * G (cs) - G (lr) ... (176) CR = 1 + .41 * G (cs ) ... (177) For the rear direction CL = 1 - G (lr) ... (178) CR = 1 ... (179) Since the matrix elements have a symmetry towards the left / right axis , you can find the values of CL and CR for the right direction by means of exchanging CL and CR. Figure 33 shows the matrix element Center Left (CL) of the four-channel decoder and the Dolby Pro-Logic plotted in three dimensions. This is also the graph of the Right Center matrix element if we swap left and right. Half of the graph, and the right and rear vertices have the value 1. The central vertex has the value 1.41. In practice this element is scaled, so that the maximum value is one. In the patent application of July 1996 these elements are replaced by sines and cosines: For the frontal direction CL = eos (45-lr) * sin (2 * (45-lr)) -sen (45-lr) * cos (2 * (45-lr)) + .41 * G (cs) ... (180) CR = sin (45-lr) * sin (2 * (45-lr)) + cos (45-lr) * cos (2 * (45-lr)) + .41 * G (cs) ... (181) These equations have never been implemented. ' The implementation of version 1.11 is handled at the address in the 1989 patent, but with a different scale, and a different function from cs. We found that it was important to reduce the non-directed level of the central output, and a value of 4.5 dB less than the Pro-Logic level was chosen. The increase function (.41 * G (cs)) was changed, to increase the value of the matrix elements back to the Pro-Logic value as cs and increases towards the center. The increase function in the decoder of version 1.11 was selected relatively arbitrarily. In version 1.11 the new cs increase function starts at zero as before, and rises with cs in the same way that CL and CR increase 4.5 dB as cs goes from zero to 22.5 degrees. In version 1.11 the increment is a constant number of dB for each dB increase in cs. The function then changes the inclination, so that in the next 20 degrees the matrix elements rise another 3 dB, and then remain constant. In this way, when the address is the "front half" (8 dB or 23 degrees) the new matrix elements are equal to the neutral values of the old matrix elements. As the direction continues to move forward, the old and new matrix elements become equal.

The output of the center channel is therefore 4.5 dB lower than the old output, when the direction is neutral, but it rises to the old value when the direction is completely to the center. Figure 33 shows the Left Center matrix element in the Logic 7 decoder of version 1.11. Note that, with respect to the graph in Figure 33, the average value and the right and rear vertices have been reduced by 4.5 dB. As cs increases, the center rises to the value of 1.41 in two inclinations. The solution for the center used in version 1.11 is not optimal. Considerable experience with the decoder in practice, has shown that the central portion of popular music recordings, and dialogue in some movies may tend to get lost when you switch between stereo playback (two channels), and playback at through the matrix. In addition, as the center channel changes in level, a listener who is not equidistant from the front speakers may notice the apparent position of a central voice moving. This problem was discussed extensively to develop the new matrix presented here. As will be seen later, there is also a problem when a signal panneates from the left to the center or from the right to the center, along the limit. The previous value gives an output too low from the central speaker, when the pan is half way.

Now let's consider the central channel in the new design. The output of the center channel must be derived from the A and B inputs to the decoder. Although it is possible to remove the input either A or B from the output of the center channel, using the matrix techniques, at any time the address is not biased to the left or right, the center channel must reproduce the sum of inputs A and B with some gain factor, either an increase or a cut. How strong should it be? The answer to this question depends on the behavior of the left and right main outputs. The matrix values presented above for LFL and LFR are designed to remove the central component of the input signals, as the direction moves forward. We can show that if the input signal has been coded forward, with some kind of cross-mixer, as a stereo width control, the matrix elements given above (the 1989 elements of U.S. Patent No. 4,862,502, the elements of the patent application of July 1996, decoder elements of version 1.11, and new ones in accordance with the present invention) completely restore the original separation. The elements of version 1.11 (with the level increase when cs is approximately 22.5) also restore the original separation. However, if the input to the decoder consists of uncorrelated left and right channels, to which an unrelated central channel has been added: A = Lin + .71 * Cin ... (182) B = Rin + .71 * Cin ... (183) then, as the level of Cin in relation to Lin and Rin increases, the C component of the decoder front outputs L and R is not completely eliminated, unless Cin is large compared to Lin and Rin. In general, there is a left Cin bit in the front outputs L and R. What does a listener hear? There are two ways to calculate what a listener hears. If a listener is exactly equidistant from the Left, Right, and Center speakers, you will hear the sum of the sound pressures of each speaker. This is equivalent to adding the three frontal outputs. Under these conditions it is easy to show that any reduction of the central component of the left and right speakers will result in a net loss of sound pressure from the central component, regardless of the amplitude of the central speaker. This is because the signal from the central speaker is always derived from the sum of the inputs A and B, and as its amplitude rises, the amplitude of the Lin and Rin signals must be increased, together with the amplitude of the Cin signal. However, if the listener is not equidistant from each speaker, the listener is much more likely to hear the sum of the sound power from each speaker, which is equivalent to the sum of the squares of the three front outputs. In fact, listening extensively has shown that, in fact, the sum of the powers of all the speakers is really what is important, so we must consider the sum of the squares of all the outputs of the decoder. If we want to design the matrix in such a way that the ratio of the amplitudes of Lin, Rin, and Cin is conserved, when switching between stereo reproduction and matrix reproduction, the sound power of the Cin component of the central output is it must raise in exact proportion to the reduction in its sound power of the left and right outputs, and its reduction in the rear outputs. An additional complication is that the left and right front outputs have the level increase described above. This will cause the center to be somewhat stronger to keep the proportions constant. We can write this requirement as a set of equations for sound power. These equations can be solved for the gain function we need for the center speaker. Previously we gave graphs showing the energy ratios for a Dolby Pro-Logic decoder under different conditions. The Pro-Logic decoder is not optimal. We can not do the same for our new decoder. In Figure 35, the solid curve shows the central attenuation that is needed for the new LFL and LFR, if the energy of the central component of the input signal in the three front channels is to be conserved, as the direction increases towards the front. The dotted curve shows the core values for a standard decoder. As can be seen in the solid curve, the necessary elevation at the level of the central channel is very steep - the elevation is many dB of amplitude per dB of address value. This steep change in amplitude is audible in practice. In addition, although the relative balance of central channel information in a popular recording is well preserved, if one is standing near the central speaker, sudden changes in the level can be annoying. On the other hand, the sound intensity of the central channel is extreme. We tested this curve and found that the central balance is excellent, but the center speaker dominated the frontal sound level, and the left-right separation was minimal. There is a better solution. The central attenuation shown in Figure 35 is derived by assuming the matrix elements previously given for LFL and LFR. How about we use different elements? Specifically, do we have to be aggressive about the removal of the central component of the left and right frontal exits? The listening tests show that the decoder elements of version 1.11 are unnecessarily aggressive to remove the central component. Acoustically, there is no need for you to do that. The energy removed from them should be given to the center speaker. If we do not remove this energy, it comes from the left and right front speakers, and the sound field is similar. The trick is to get a balance between the center power in the left and right speakers, and the central power in the center speaker. We create the optimal system by first selecting a smoother function for the level increase at the central output, as cs increases along the axis lr = 0. Then we can solve the decrease in the level J ^ needed in the front left and right outputs, to keep constant the power of the component Cin in the sound field in the room. Assume that the central channel is reduced in level by 4.5 dB below the level in the Griesinger 1989 decoder 15 (4,862,502), or -7.5 dB of total attenuation. This is a factor of 0.42. For the front direction, ^ CL = .42 - .42 * G (lr) + GC (cs) ... (184) CR = .42 + GC (cs) ... (185) For the rear direction, CL = .42 - .42 * G (Ir) ... (186) CR = .42 ... (187) Many functions were attempted for GS (cs). The one given below- is specified in terms of the angle cs in 25 degrees, and was obtained by some trial and error.

In MATLAB notation: center_max = .65; center_rate = .75; center_max2 = 1; center_rate2 = .3; center_rate3 = .1; if (cs < 12) gc (cs + l) = .42 * 10? (db * center_rate / (20)); tmp = ge (cs + 1); otherwise, if (cs < 30) ge (cs + 1) = tmp * 10? ((cs-11) * center_rate3 / (20)); if (ge (cs + 1)> center_max) (ge (cs + 1) = center_max; end otherwise, ge (cs + 1) = center_max * 10? ((cs-29) * center_rate2 / (20) ); if (ge (cs + 1) > center_max2) ge (cs + 1) = center_max2; end end This function is plotted as the solid curve in Figure 36. We can solve the necessary function for LFR if we assume functions for LFL , LRL, and LRR We want to solve the speed that the Cin component must decrease in the left and right outputs, and then design the matrix elements that provide this rate of decrease.These matrix elements must also provide some increase in the components Lin and Rin, and they must have the current form in the limit of left to the center, as well as the limit of right to the center We assume that: LFL = GP (cs) ... (188) LFR = GF (cs) .. (189) CL = .42 - .42 * G (Go) + GC (cs) ... (190) CR = .42 + GC (cs) ... (191) The power from the left front speakers and right is given by: PLR = (GP? 2 + GF? 2) * (Lin? 2 + Rin? 2) + (GP-GF)? 2 * Cin? 2 ... (192) The power from the central speaker is: PC = GC? 2 * (Lin) ? 2 + Rin? 2) + 2 * GC? 2 * Cin 2 ... (193) The power from the rear speakers depends on the matrix elements we use. We will assume that the back channels are attenuated by 3 dB during the forward direction, and that LRL is eos (cs) and LRR is sin (cs). From a single horn, then, PREAR = (.71 * cos (cs) * (Lin + .71 * Cin) -sen (cs) * (Rin + .71 * Cin)))? 2 ... (194) If we assume that Lin? 2 and Rin? 2 are approximately equal, for two speakers, PREAR = .5 * Cin? 2 * ((eos (cs) -sen (cs))? 2 + Lin? 2 ... (195) The total power from all the speakers is PLR + PC + PREAR, therefore: PT = (GP? 2 + GF? 2 + GC? 2) * (Lin? 2 + Rin? 2) + ((GP - GF)? 2 + 2 * GC? 2 ) * Cin? 2 + PREAR ... (196) The ratio of the power of Cin to the power of Lin and Rin is, therefore (assuming Lin? 2 and Rin? 2 equal): RATIO = ((GP? 2 + GF? 2 + GC? 2) * (Lin? 2 + Rin? 2) + ((GP-GF)? 2 + 2 * GC ~ 2) * Cin? 2 + PREAR) / PREAR = (((gp (cs) -gf (cs))? 2 + 2 * gc (cs)? 2 + .5 * (eos (cs) - sin (cs)) ~ 2) * Cin? 2/2 * (gp (cs) ? 2 + gc (cs) "2 + gf (cs) ~ 2) + l) * Lin? 2 = (Cin? 2 / Lin? 2) * ((gp (cs) -gf (cs))? 2+ 2 * (gc (cs)? 2) + .5 * (eos (cs) -sen (cs))? 2) / 2 * (gp (cs)? 2 + gc (cs)? 2+ gf (cs) ? 2) + l ... (197) For normal stereo, GC = 0, GP = 1, and GF = 0. The power ratio of the center to LR is, then: RATIO = (Cin? 2 / Lin? 2 ) * 0.5 ... (198) If this ratio is going to be constant, regardless of the value of Cin? 2 / Lin? 2 for the active matrix, which is desirable, then: ((gp (cs) -gf (cs))? 2- 2 * gc (cs)? 2) + .5 (eos (cs) -sen (cs))? 2) = (gp (cs)? 2 + gc (cs)? 2 + gf (cs)? 2) - .5) ... (199) Equation 199 can be solved numerically. If we assume the previous GC value, and GP = LFL as above, then Figure 36 shows the resulting values for GF in the solid curve, sin (cs) * corrl (the previous LFR element) in the line curve, and sin (cs) in the dotted curve. Note that GF remains close to zero until cs reaches 30 degrees, and then increases sharply. In practice we arbitrarily increase GF beyond 30 degrees, to reach the value of 0.71 as the dashed and dotted curves. This causes the complete cancellation of the center channel on the left and right during the strong direction. In addition, GF must be softly interpolated to the previous value along the limits. All these curves have a negative sign in practice. GF gives the shape of the LFR matrix element along the axis lr = 0, as cs increases from zero to the center. We need a method to combine this behavior with that of the previous LFR element, which should be kept along the boundary between the left and the center, as well as from the right to the center. One method to do this when cs = 22.5 degrees, is to define a function of difference between GF and sin (cs). So we limit this function in different ways. In Matlab notation, gf_diff = sin (cs) - gf (cs); for cs = 0:45; yes (gf_diff (cs) > sen (cs)) gf_diff (cs) = sin (cs); end yes (gf diff (cs) < 0) gf_diff (cs) = 0; end end% find c / s limited if (and < 24) bes = y- (x-1); if (bcs < l)% this limits the maximum value bes = 1; order otherwise bes = 47-y- (x-1); yes (bes < 1)% > 46) bes = 1; % 46; end order Now you can write the element LFR:% this neat trick makes an interpolation to the limit% the cost, of course, it's a division! ! ! if (and <23)% this is the easy way for half the region lfr3d (47-x, 47-y) = -sin_tbl (y) + gf_diff (bes); otherwise tmp = ((47-y-x) / (47-y)) * gf_diff (y); lfr3d (47-x, 47-y) = -sin_tbl (y) + tmp; end Notice that the sign of gf_diff is positive in the previous equation. Therefore, gf_diff cancels the value of sin (cs), reducing the value of the element to zero, along the first part of the axis lr = 0. Figure 37 shows the left-front-right matrix element (LFR) with the correction for the central level along the axis lr = 0. Note that the value is zero in the middle of the plane (without direction), and remains at zero as cs increases to 22.5 degrees along the axis lr = 0. Then the value falls to coincide with the previous value along the limit from the left to the center and from the right to the center. Now consider the panning error in the central output. Figure 38 shows the center-left matrix element (CL) with the new central magnification function GC (cs). Note the correction for panning along the boundary between the left and the center. As it is presented, the new central function (if we write it this way): CL = .42 - .42 * G (Jr) + GC (cs) ... (200) CR = .42 + GC ( cs) ... (201) works well along the axis lr = 0, but causes a panning error along the boundary between the left and the center, and between the right and the center. The values in the paten-te application of July 1996 give a smooth function of cos (2 * cs) along the left boundary, which creates a smooth pan between the left and the center. We would like our new central function to have a similar behavior along this limit. We can make a correction to the matrix element that will do the job by adding an additional function of xymin (in Matlab notation): center_fix_tbl = .8 * (corrl-1); CL = .42 - .42 * G (lr) + GC (cs) + center_fix_table (xymin) CR = .42 + GC (cs) + center_fix_table (xymin) Figure 39 shows the levels of the central output and the left output , as a signal pans from the center to the left. Note that with the correction, the panning of the center, although not perfect, is reasonably close to the inverse of the left output. (The values are inverted in the cs axis). Now consider a new five-channel encoder (called "Logic 7") that is designed to operate correctly with the decoder specified by the equations and algorithms given above. There are two major goals for this encoder. First, it is capable of encoding a 5.1-channel tape, in a manner that allows the encoded version to be decoded by means of a Logic 7 decoder, in accordance with the present invention, with minimal inaccuracy. Second, the encoded output must be stereo compatible - that is, it must sound as close as possible to a manual mix of two channels of the same material. A factor in this stereo compatibility should be that the encoder output, when played back in a standard stereo system, should give an identical perceived sound intensity for each sound source in an original five-channel mix. The apparent position of the stereo sound source should also be as close as possible to the apparent position on the five original channels. In discussions with the Institute for Broadcast Technique (IRT) in Munich, it became clear that the goal of stereo compatibility of the stereo signal as described above, it can not be fulfilled by means of a single encoder setting. A recording of five channels, where all the channels have equal foreground importance, must be encoded as described above. This coding requires that the surrounding channels be mixed within the encoder output, in such a way that energy is conserved. That is, the total energy at the output of the encoder must be the same, regardless of which input is being driven. This technique will include most movie sources and 5-channel music sources, where the instruments have been assigned to all five speakers. Although those sources of music are not currently common, it is the opinion of the inventor that these will become common in the future. But music recordings in which the foreground instruments are placed on the three front channels, with repercussion primarily on the back channels, require a different coding technique. After a series of tests at the IRT and elsewhere, it was determined that music recordings of this type were successfully coded in a stereo compatible manner, when the surrounding channels were mixed with 3 dB of less power than the other channels. This -3 dB level has been adopted as a standard for surrounding coding in Europe, but the standard specifies that other surrounding levels can be used for special purposes. As we will see later, the new encoder contains active circuits that detect strong signals in the surrounding channels. When those signals are occasionally present, the encoder uses the entire surrounding level. If the surrounding inputs are continuously down to -6 dB or less, compared to the front channels, the surrounding gain is gradually decreased by 3 dB, to correspond to the European standard. During tests with the IRT in Munich, it was found that the encoder described in the AES document (pre-printing No. 4402) incorrectly encoded a particular tape. A new architecture was developed to solve the problem with this tape. Although the coding of this particular tape was only marginally improved, the new encoder is superior in its operation over a variety of difficult material. The original encoder was first developed as a passive encoder, and it worked reasonably well with a variety of input signals. The new encoder will also operate in a passive mode, but is primarily intended to function as an active encoder (ie, one in which the encoding depends on the types of signal presented to its inputs). The active circuitry corrects many small errors inherent in the design. However, even without the active correction, the operation is better than that of the encoder previously described. After listening extensively, many small problems with the first encoder were discovered. In the new encoder, attention has been given to many, but not all of these problems. For example, when the stereo signals are applied to both the front and rear inputs of the encoder, at the same time, the resulting output of the encoder is biased too much to the front. The new encoder compensates for this effect by slightly increasing the rear biasation. In the same way, we have found that when the film is encoded with substantial surrounding content, there is a net backward bias, which may tend to reduce the power of the dialogue signal in the center channel. This can be important in a film, where the intelligibility of the dialogue is of supreme importance. The new encoder compensates for this effect by slightly raising the input of the center channel to the encoder, under these conditions. The new encoder, shown in the schematic block form in Figure 40, handles the left, middle and right channels identically to the previous design, and identically to the Dolby encoder, with the proviso that the central attenuation function fen in the attenuator 302 is equal to 0.71 0 -3 dB. In accordance with the previous designs of the encoder shown in Figures 10 and 11, the left (L), middle (C) and right (R) signals are presented to the input terminals 50, 52 and 54 of the circuit system of the encoder, respectively. The left side (LS) and right side (RS) signals are presented to the input terminals 62 and 64, respectively. An additional signal LFE (for effects of low frequency in a mixture of 5 + 1) is applied to a new input terminal 370. The signals C and LFE pass through the attenuator / gain elements 372, 374, respectively , where C is amplified by a factor fen and LFE by a factor of 2.0. These signals are applied, each one, directly to the summing circuit 278, and the signal R is applied similarly to the summing circuit 282. The surrounding signals also apply to these summing circuits, but only after some manipulation, which seems be more complex than it really is. In the surrounding channel attenuators 376, 378, 380, and 382, the functions fc () and fs () direct the surrounding channels either to a path (through the phase shift elements 234 and 246) with a phase shift of 90 degrees, relative to the front channels (which proceed through the phase shifters 286 and 288), or to a path without any relative phase shift. In the basic coder, fc is one and fs is zero, so that the active path is through the 90 degree phase shifters. In this way, the signal LS passes without change through the block 376 to the attenuator 396, where it is multiplied by a factor of 0.91, then it passes to the adder 406, where it is mixed with the signal RS coupled transversely of attenuator 404, which has a gain of -crx. The value of crx is typically 0.38. This controls the amount of negative cross-feed for each surrounding channel. The signal then passes through a phase shifter 234 of 90 degrees and an adder 276, where it is mixed with the other signals of the phase shifter 286 to this adder, and passes to the output terminal 44 as the "A" signal. As in the previous encoder, when there is only one input to one of the surrounding channels, the outputs A and B at terminals 44 and 46 respectively, have an amplitude ratio of -.38 / .91, which gives a steering angle of 22.5 degrees to the rear. The signal RS applied to the terminal 64 similarly passes through the attenuator 382 with unity gain to the inverting element 400, and then through an attenuator 402 with a gain of 0.91, as for the channel LS. This signal is then added to -crx times the unmodified RS signal in adder 408. As for the LS channel, the signal passes through a phase shifting element 246 of 90 degrees, and thence to an adder 280. The signals R, C, and LFE, after the combination in the adder circuit 282, pass through a phase-shifting element 288, within the adder 280, where these are mixed with the signals RS shifted in phase and LS fed transversely, to provide the output signal "B" at terminal 46. As usual, for each of the output signals at terminals 44 and 46, the output level is unity, since the sum of the squares of 0.38 and 0.91 is one. Although the output of the encoder is simple when only one channel is driven, it becomes problematic when both surrounding inputs are operated at the same time. If we drive the LS and RS inputs with the same signal, a common practice in movies, all the signals in the summing nodes are in phase, such that the total level at the exit is .38 + .91, or 1.29. This output is too strong by the factor of 1.29, or 2.2 dB. The active circuit system (not shown, but similar to the active circuitry in the decoder) is included in the encoder to reduce the gain by the factor of 2.2 dB when this situation occurs, that is, when the two surrounding channels They are similar in amplitude and are in phase. Another error occurs when the two surrounding inputs are equal in level and out of phase. In this case, the two attenuation factors subtract, so that the output level is .91 - .38, or .53. This signal will be decoded as a central sense signal, at a reduced level. This error is severe. The previous encoder produced a non-directed signal under these conditions, which is reasonable. It is not reasonable that signals applied to the rear input terminals should result in a central oriented signal. In this way, the active circuit system (not shown, but similar to that in the decoder) increases the value of fs when the two rear channels are similar in level, but of opposite phase. The result of mixing both the actual trajectory and the offset path in phase for the rear channels, is a phase difference of 90 degrees, between the output channels A and B, which represents a non-directed signal, this effect being the desired one. In discussions at the IRT in Munich, it was noted that there is a European standard surround encoder, this encoder simply attenuates the two surrounding channels by 3 dB, and adds them within the front channels. In this way, the left rear channel is attenuated by 3 dB and added to the left front channel. This encoder has many inconveniences when it encodes multi-channel movie sound, or recordings that have specific instruments assigned to the surrounding channels. Both the sound intensity and the direction of these instruments will be encoded incorrectly. However, this encoder works rather well with classical music, where the two surrounding channels mainly have repercussion. The 3 dB attenuation was carefully chosen through ear tests, to produce a stereo compatible coding. It was decided that the new encoder of the present invention should also incorporate this 3 dB attenuation when classical music is being encoded, and that one can detect this condition by monitoring the relative levels of the front and surround channels in the encoder . A major function, therefore, of the fc function in the surrounding channels is to reduce the level of the surrounding channels in the output mix by 3 dB, when the surrounding channels are much softer than the front channels. A circuit system similar to that in the decoder is provided, to compare the front and rear levels, and when the rear part is smaller by 3 dB, the value of fc is reduced to a maximum of 3 dB. this maximum attenuation of 3 dB is reached when the rear channels are 8 dB less strong than the front channels. It seems that this active circuit works well. This makes the new encoder compatible with the European standard encoder for classical music. However, the instruments that are intended to be strong in the back channels, are encoded with full level.

There is another function of the real coefficient mixing path fs for the surrounding channels. Note that this trajectory, through the attenuators 378 and 380, also passes through the transversely fed elements 384 and 386 to the adders 392 and 394 in the opposed channels, with the attenuating elements 388 and 390 of 0.91 in the trajectories of main signals, before being applied to summing circuits 278 and 282. When a sound is moving from the left front input to the left rear input, the active circuit system (not shown) compares the levels at these inputs, and it detects that these signals are similar in amplitude and in phase, and under these conditions, fc is reduced to zero, while fs is increased to one. This change to real coefficients in the coding results in a more accurate coding of this type of panning. In practice, this function is probably not essential, but it seems to be an elegant refinement. In summary, then, the active circuits comprise elements to compare the level and phase between the front and rear channels on each side, and to compare the relative energy in the front and rear channels. These circuits are easily implemented in the form of logarithmic ratio detectors, and are well known to those skilled in the art. Dependent on the outputs from these detectors, these active circuits 1. Reduce the level of the surrounding channels by 2.2 dB, when the signals are in phase; 2. Increase the actual coefficient mixing path for the back channels enough to create a non-directed condition, when the two rear channels are out of phase; 3. Decrease the level of the surrounding channels by up to 3 dB, when the surrounding level is much lower (-8 dB) than the front channels; 4. Increase the level and negative phase of the back channels, when their level is similar to the front channels; and 5. They cause the surrounding channel mix to use real coefficients, when paging a sound source from a front entrance to the corresponding rear entrance. It is likely that additional improvements to the encoder include a feature for the front channels, such that when the two front channels are out of phase, the encoder does not cause the decoder to place the sound 20 in the back, as currently, but to detect this condition and make it appear that the encoded output is not directed (ie the result will be a phase shift of quadrature between channels A and B). Although the preferred embodiments of the invention have been described and illustrated herein, there are many possible embodiments, and they and other modifications and variations will be apparent to those skilled in the art, without departing from the spirit of the invention.

Claims (13)

1. REIVI DICATIONS 1. A surround sound decoder for redistributing a pair of left and right audio input signals, including both directionally and non-directionally encoded components, into a plurality of output channels for playback through loudspeakers surrounding a listening area , and incorporating means for determining the directional content of said left and right audio signals and generating from them at least one left-right directional control signal and a center-surrounding directional control signal, the decoder comprising: delay means to delay each of said left and right audio input signals to provide delayed left and right audio signals; a plurality of multiplying means equal to twice the number of said plurality of output channels, organized in pairs, a first element of each mentioned pair receiving said delayed left audio signal and a second element receiving said delayed right audio signal, each one of said multiplying means multiplying its input audio signal by a variable matrix coefficient to provide an output signal; said variable matrix coefficient being controlled by one or both of said directional control signals; and a plurality of summing means, one for each channel of said plurality of output channels, each of said summing means receiving the output signals of a pair of said multiplying means and producing in its output a signal of said plurality of signals of exit; the decoder having said variable matrix values constructed so as to reduce the audio components encoded directionally in outputs that are not directly involved in reproducing them in the intended direction; improving the audio components encoded directionally in the outputs that are directly involved in reproducing them in the intended direction in order to maintain a constant total power for such signals; while maintaining a high separation between the left and right channel components of the non-directional signals independently of said directional control signals; and effectively maintaining constant the sound level defined as the total audio power level of the non-directional signals whether or not the signals are coded non-directionally and independently of their intended address, if present. The decoder of claim 1, wherein said left and right audio input signals were originally encoded from five channels in two channels and where said plurality of output signals is five, such that said left audio input signals and right are decoded into five output signals that are reproduced by amplification and application of the aforementioned five amplified output signals to five speakers arranged to surround the listener. 3. The decoder of claim 2, wherein said speakers are placed in places on the left front, center front, right front, left rear and right rear of the listening position, said output signals being typically named according to their places directional directions in relation to the position of the listener. The decoder of claim 3, wherein a sixth reduced bandwidth audio output signal, intended for low frequency sound effects, is provided in addition to the five listed output signals. 5. The decoder of claims 1 to 4, wherein said left and right audio input signals are decoded in said plurality of output signals by means of analog circuitry. The decoder of claims 1 to 4, wherein said left and right audio input signals are decoded in said plurality of output signals by means of digital signal processing circuitry, after first converting said input signals to format digital, and finally converting said output signals back into analog format for reproduction in a similar plurality of loudspeakers surrounding the listener. 7. The decoder of claim 3, wherein said variable matrix coefficients that determine the level of the rear output signal are maintained at a level of 3 dB less when said left-right directional control signal is of small magnitude and rise to magnitude full when this signal reaches a magnitude equivalent to a directional control angle of 22.5 degrees or more, but is independent of said center-surrounding directional control signal, in an operation mode intended for television sound reproduction, thereby providing less variation of the sound level of the rear outputs in relation to the front outputs when directional control occurs in the forward direction, resulting in more natural and soft surrounding sound effects. The decoder of claim 3, wherein the absolute value of one of said directional control signals is limited when the other of said directional control signals is changing rapidly, thereby providing improved responsiveness to dynamic effects. The decoder of claim 3, wherein in each frontal quadrant the coefficients calculated for the left and right components of the input signal are made such that the sum of the squares of these elements is equal to one when the directional control signal The central surrounding is close to zero, so as to reduce undesirable variations of the total power delivered to the speakers as a result of the directional control. The decoder of claim 3, wherein a central amplification function is provided such that as a result minimal apparent movement of the surrounding sound sources in the front while maintaining maximum left-right separation of sounds not subject to directional control. The decoder of claim 3, wherein a new central front matrix coefficient is made dependent on said central surrounding directional control signal such that up to an effective angle of 30 degrees toward the front the output of the central channel rises to a value of 3 dB lower than a standard type of decoder, then rise faster to reach the same maximum level used in the standard type of decoder to full frontal directional control, further cutting the level of the central component of the signal in the channels front left and right so as to preserve in the sum of the powers of each decoder output the ratio of the power of the central input signal to the total power of all other input signals to the decoder from which they were derived said left and right audio input signals to said decoder; thereby preserving in the decoder outputs the same balance between the central signal and other signals that was present before the signals were encoded, and also preserving the balance between the central signal and other signals in recordings originally made for two-channel playback . 12. An active encoder, suitable for encoding five audio channels of full bandwidth into two output channels, said five input channels being respectively left front, center front, right front, rear or left and rear surround or right surround, and said output channels comprising left or A and right or B channels, respectively, comprising: first, second, third, fourth and fifth input terminals for said left front, center front, right front, left surround and right surround channels respectively; an attenuator circuit connected to said second input terminal for attenuating said central front signal in a fen factor; first and second adder circuits, said first adder circuit receiving the attenuated central front signal of said attenuator circuit and the left front input of said first input terminal directly, and said second adder circuit also receiving said attenuated central front signal and the right front signal of said third input terminal directly; a first functional attenuator with the attenuation function fc (l, ls) for attenuating said left surrounding signal received from said fourth input terminal; a second functional attenuator with the attenuation function fc (r, rs) to attenuate said right-hand surrounding signal received from said fifth input terminal; a third functional attenuator with the attenuation function fs (l, ls) to attenuate said left surrounding signal received from said fourth input terminal; a fourth functional attenuator with the attenuation function fs (r, rs) to attenuate said right-hand surrounding signal received from said fifth input terminal; first, second, third and fourth crossfeed attenuators having an attenuation factor -crx, each receiving the output signals of said first, second, third and fourth, respectively, operating attenuators; fixed attenuators first, second, third and fourth, having an attenuation factor of 0.91, each receiving the output signal of said first, second, third and fourth functional attenuators, respectively; first, second, third and fourth adders, each receiving the output signals of said first, second, third and fourth fixed attenuators, respectively, and adding to them the output signals of said second, first, fourth and second crossfeed attenuators third, respectively; first and second phase shift circuits having a phase shift function -90 'receiving the outputs of said first and second adder, respectively; third and fourth phase shift circuits having a phase shift function -0 'receiving the outputs of said first and second adder circuits, respectively, said first and second adder circuits also receiving the outputs of said third and fourth adder, respectively, and combining them with the left front and right front input signals, respectively, with the attenuated central front signal of said attenuator circuit connected to said second input terminal; a third summing circuit for summing the output signals of said first and third phase shift circuits, respectively, to provide said left or A output signal to a first output terminal; a fourth adder circuit for summing the output signals of said second and fourth phase shift circuits, respectively, to provide said right output signal or B to a second output terminal; record amplitude detecting means first, second, third, fourth and fifth, to detect the amplitudes of the signals applied to said first, second, third, fourth and fifth input terminals; first and second comparing means for comparing the recording amplitudes of the front and rear signals on the left side and on the right side, respectively; means responsive to the output of the first comparator means to reduce the function fc (l, ls) of said first one-to-zero functional attenuator and increase the function fc (l, ls) of said third functional attenuator from zero to one of a complementary way by changing the predominant signal direction from left frontal to surrounding left; means responsive to the output of the second comparing means to reduce the function fc (r, rs) of said second functional attenuator from one to zero and increase the function fs (r, rs) of said fourth functional attenuator from zero to one of a complementary way by changing the predominant signal direction from right frontal to surrounding right; third comparative means for comparing the recording amplitudes of the frontal signals with those of the posterior signals; means responsive to the output signal of said third comparing means to reduce the fen function by up to 3 dB by exceeding the front signals the rear signals by up to 8 dB or more; fourth comparator means for comparing the phase and the amplitude between the left and right surround input signals applied to said fourth and fifth input terminals; and means responsive to the output signal of said fourth comparator means to reduce the gain of said first and second functional attenuators by up to 2.2 dB when the two surrounding signals are similar in amplitude and in phase, and to increase the gains of said attenuators functional third and fourth when the two surrounding signals are similar in amplitude but in anti-phase, thereby forcing outputs A and B to be in phase quadrature relation, representing a condition without directional control. The encoder of claim 12, further comprising: a sixth input terminal for receiving a low frequency effects signal; a gain stage of 2.0 gain to amplify said signal of low frequency effects; said amplified low frequency effects signals being applied equally to additional inputs of said first and second adder circuits so as to appear equally and in phase on both outputs A and B at said first and second output terminals, respectively.