CZ2000799A3 - Decoder and encoder of sound reproduction system - Google Patents

Abstract

The system is for reproducing sound for converting stereo signals to two input channels (92, 94), wherein at least one component the signal is directionally coded and correlated and at least one the signal component is not coded and is not correlated in two input channels, to signals for several output channels channels. The system includes a decoder (90) for amplifying the correlated the components of the input signals in the desired direction and reduce the force such signals in channels that are not associated with coded direction, while keeping the distance between the respective left and right input channels (92, 94) and total energy of the uncorrelated input channel component (92, 94) in each output channel, so the tools recorded on the right input channel remain on the right side output channels and tools recorded on the left remain on this left side, and the apparent volume of all tools in all output channels remain the same, without regardless of the direction of the coded component of the input signals, the system further includes an encoder for encoding the five input ones channels by decoding them in the right direction a levels in decoders (90) according to the solution and in the decoders of current film standard. The decoder (90) adjusts level of central energy conservation channel ratios in input signals for film reproduction, or for suppressing the central component of input signals from left and right right front output for music signals; plus includes frequency-dependent elements in surround outputs for improving the reproduction of the side and back signal components at five

Description

Decoder and encoder of a sound reproduction system

Technical field

The present invention relates to audio reproduction systems comprising decoding a pair of input audio signals into a plurality of output signals for reproduction, after appropriate amplification, through the same plurality of speakers arranged to surround the listener, as well as encoding the multi-channel material into two channels. This application is based on U.S. Provisional Patent Application No. 60 / 058,169 entitled 5-2-5 Matrix Encoder and Decoder System, filed September 5, 1997.

BACKGROUND OF THE INVENTION

The present invention is concerned with an improved set of design criteria and their solutions for creating a decoding matrix having optimal psychoacoustic performance in reproducing encoded multichannel material as well as standard two-channel material, including maintaining a large distance between left and right stereo components under all circumstances, even with forward or reverse bias for input signals, or when there is a strong audio component in a particular direction, while maintaining a large distance between the different outputs for signals in a defined direction, while keeping non-directionally coded components at a constant acoustic level regardless of directional coded components of the audio input signals, including frequency-dependent circuits that improve the balance between the front and rear signals, move the sound around the seven channel version of the system and produce a five channel version sound that is close to the seven channel version.

• · 9

9999

9 • 99

• »

In addition, the present invention is concerned with an improved set of design criteria and their solutions for creating a coding circuit for encoding multi-channel audio into two channels for reproduction in standard two-channel receivers and by means of the decoders of the present invention.

The present invention is part of an ongoing effort to improve the means of encoding multichannel audio signals into two separate channels and then the means for splitting the resulting two channels back into multichannel signals from which the two channels were derived. One of the objectives of this coding / decoding process is to re-create the original signals as perceptually identical to these original signals as possible. Another important goal of the decoder is to extract five or more separate channels from a two-channel 15 source that has not been encoded from a five-channel original. The resulting five-channel presentation must be at least as musical and enjoyable as the original two-channel presentation.

The present invention is concerned with improvements in deriving suitable variable matrix coefficients. For a better understanding of these improvements, reference may be made in this specification to: US Patent No. 4,862,502 (Griesinger, 1989), which will be referred to as the '89 Patent; U.S. Patent No. 5,136,650 (Griesinger), 1992, which will be referred to as the '92 Patent; US

5 Patent Application No. 08 / 684,948 (Griesinger), July 1996, which is commonly referred to as July '96; and US Patent Application No. 08 / 742,460 (Griesinger), November 1996, which will be referred to as November '96.

Commercial versions of the decoder based on this last application will be referred to as Version 1.11 (or VI.11). Some others • 99 9 • 99 9 • 99 * 9

Improvements have been described in Provisional Patent Application 60 / 058,169, filed in September 1997, which will be referred to as Version 2.01 (or V2.01). These versions VI.11 and V2.01 and the present invention will be collectively referred to as decoders

Logic 7.

Other technical references cited in this description are: [1] Multichannel Matrix Surround Decoders for Two-Eared Listeners, David Griesinger, AES Preprint # 4402, October 1996, and [2] Progress in 5-2-5 Matrix Systems, θ David Griesinger, AES Preprint # 4625, September 1997. Summary of the Invention

The means used to realize the two objectives for recreating the original signals encoded from five to two channels and creating a perceptually satisfactory reproduction of the two channel material in a five channel format have been developed as we continue to better understand the respective physical and psychoacoustic phenomena. The foregoing patents and patent applications cited above have presented a constructional θ philosophy that has created usable decoding devices.

The present invention is concerned with implementing an active matrix having certain properties that maximize its psychoacoustic performance. In another aspect, the present invention discloses a frequency-dependent modification of certain outputs from,.

this active matrix. In another aspect, the invention provides active circuits that encode five input channels to dual channel output that will operate optimally with both standard dual channel decoders of the present invention and Dolby Pro-Logic industry standard decoders.

• φ • · φ φ φφφφ φ

φ ·· φ φ φ φ • Φ

In part, the invention is an active matrix decoder having matrix elements that vary depending on the directional component of the incoming signals. The matrix elements vary in such a way as to limit the volume of the directionally coded signals in the outputs that are not engaged in the designated direction, while increasing the volume of these signals in the outputs that are engaged in the reproduction in the designated direction, the sides of other signals that may be present at the inputs at the same time. In addition, the matrix elements of the present invention restore the separation of the right and left sides of a separated (uncorrelated) two-channel material that has been directionally coded by increasing or decreasing the color of the audio between the two inputs - for example, with stereo bandwidth control. In addition, the matrix elements of the present invention are designed to maintain the highest possible energy balance between the various components of the input signal, so that the balance between vocals and accompaniment is maintained in the decoder outputs. Consequently, the matrix elements of the present invention maintain both the volume of the non-directionally coded components of the input sound as well as the right and left side spacing of these components.

In addition, the decoders of the present invention include frequency dependent circuits that improve compatibility

5 decoder outputs when playing standard 2-channel material, which converts two-to-five-channel decoder outputs to four to seven-channel decoders, and modifies the rear channel spectrum in the five-channel decoder so that sound routing seems much more likely to route audio from a seven-channel decoder .

* * * * · · · · ·

I ·· * ·

The encoders of the present invention mix five (or five full range plus one low frequency) input channels into two output channels such that when the input level of a particular input is strong, the energy present at that input is maintained at the output; such that the direction of the strong input is coded in the phase / amplitude ratio of the output signals; such that strong signals can be input between any two encoder inputs and the output will be properly directionally encoded. In addition, the separate material fed to the two rear encoder inputs will be coded into two channels in such a way that the right and left side spacing of these inputs will be retained when the encoder output is decoded by the decoders of the present invention. a 2-channel output that will be decoded into the back channels of the decoders of the present invention and the Dolby standard decoders, and so that the inputs in counter-phase to the two back channels of the encoder will produce outputs corresponding to non-directional signals for the decoders of the present invention and Dolby standard decoders, and also that low-level reverberation signals applied to the two rear encoder inputs will be encoded with a 3 dB reduction in the 2-channel output.

New features characterizing the present invention are set forth in the appended claims. The present invention, including its other features and its various advantages, is most readily understood by reading the following detailed description of an illustrative embodiment thereof with reference to the accompanying drawings.

• · · · ♦ ··

- · · · · · · · ···· · ··

T »9 9 I t t t t t t t t t

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 shows a block diagram of a directional detection section and a two-to-five-channel conversion section of a decoder according to the present invention, but without including the aspects further illustrated in Figs. 2 and 3;

Fig. 2 shows a block diagram of a five-channel frequency-dependent signal processor active circuit that can be connected between the outputs of the matrix section of Fig. 1 and the outputs of the decoder;

Fig. 3 shows a block diagram of a frequency-dependent active signal processor for converting five channels to seven, which may alternatively be connected between the outputs of the matrix section of Fig. 1 and the outputs of the decoder;

Fig. 4 shows a block diagram of an active five channel to two channel encoder according to the present invention;

Giant. 5 is a three-dimensional graphical representation of a prior art left front left (LFL) matrix element of the '89 patent and Dolby Pro-Logic on a scale such that the maximum value is one;

Fig. 6 shows a three-dimensional graphical representation of a prior art left front right (LFR) matrix element of the '89 patent and Dolby Pro-Logic in '0' 0 0 0 0 0 0 * ··· 0 0 0 · 0 0 »· 0 * · ··· 0 0 0 0 •• 0 0« 0 «· 0 00 scale changed 0.71 so that the minimum value is -0.5 and the maximum value is +0.5;

Figure 7 shows a three-dimensional graphical representation of the square root of the sum of squares of LFL and LFR of the prior art from the '89 patent on a scale such that the maximum value is one;

Fig. 8 shows a three-dimensional graphical representation of the square root of the sum of squares of LFL and LFR of the prior art from Application No. 08 / 742,460 on a scale such that the maximum value is one;

Figure 9 shows a three-dimensional graphical representation of a left front left (LFL) matrix element in VI.11;

Fig. 10 shows a three-dimensional graphical representation of a partially completed left front left (LFL) matrix element according to the present invention;

Fig. 11 is a graph illustrating the behavior of LFL and LFR in the present invention along the posterior boundary between left and rearmost;

Fig. 12 shows a three-dimensional graphical representation of the fully completed left front left (LFL) matrix element of the present invention viewed from the rear left side;

»»

000

* 0 0 0 0 • ♦ 0 0 0

0 0 0 0 0 0 0 0

Fig.13

Fig.13

Fig.14

Fig.14

Fig.15

Fig.15

Fig. 16

Fig. 16

Fig.17

Fig.17

Fig.18

Fig. 18 shows a three-dimensional graphical representation of a fully completed left front right (LFR) matrix element according to the present invention;

shows a three-dimensional graphical representation of the rms value of the LFL and LFR sum of the present invention;

shows a three-dimensional graphical representation of the square root of the sum of the squares of LFL and LFR of the present invention, including back level correction, viewed from the rear left side;

is a graph illustrating the solid line central matrix elements that would be used in a prior art Dolby Pro-Logic decoder as a function of CS in dB and the dotted line the actual value of the central matrix elements in a Dolby Pro-Logic decoder;

is a graph illustrating the solid line ideal value of the central matrix elements and the dotted line of the actual value of the central matrix elements in a prior art Dolby Pro-Logic decoder;

shows a three-dimensional graphical representation of the square root of the sum of the squares LRL and LRR using elements of VI. 11 of the prior art;

• · · * * t t t t t t t t t t t »» »

Φ · ·

ΦΦ a ·

Figure 19 shows a graphical representation of a numerical solution for GS (lr) and GR (lr) for a constant power level along the cs = 0 axis, and a zero output along the border between the left side and the center;

Fig. 20 shows a three-dimensional graphical representation of the square root of the sum of the squares LRL and LRR using the values for GR and GS of the present invention;

Fig. 21 shows a three-dimensional graphical representation of a prior art central left (CL) matrix element of a four-channel decoder of the '89 patent (Dolby Pro-Logic decoder), which can also represent a censal right (CR) matrix element interchangeably with the left and right sides;

Fig. 22 shows a three-dimensional graphical representation of the central left (CL) matrix element in Logic 7 VI.11 decoders;

Fig. 23 is a graph illustrating the solid line mean attenuation attenuation required for the new LFL and LFR and the dotted mean attenuation line for a standard prior art Dolby Pro-Logic decoder;

Fig. 24 is a graph illustrating a solid line ideal mean attenuation for the film strategy of the present invention, a dashed line value that operates. Considerably better, and with a dotted line to compare the mean attenuation for a standard Dolby decoder;

Figure 25 shows the mean attenuation used in the music strategy of the present invention;

Fig. 26 is a graph illustrating a fixed line GF value required for constant power ratios with music mean attenuation GC of the present invention, the dashed line sin (cs) * corrl of the preceding LFR element, and the dotted line sin (cs);

Fig. 27 shows a three-dimensional graphical representation of a left front right (LFR) matrix element according to the present invention, including a mid-level correction along the axis lr = 0;

Fig. 28 shows a three-dimensional graphical representation of a central left (CL) matrix element with a new central gain function; and

Fig. 29 shows a graph that plots the output level from the left front output (dotted line) and the central output (solid line) as a strong signal region from center to left.

DETAILED DESCRIPTION OF THE INVENTION

The construction presented in this description retains much of the design philosophy of the previous decoders, but fefe •··· · · · · · ··· · · · · · · · · 9 · 9

9 9 9 9 * »·· The actual design has been changed in many ways. It is not possible to fully describe the evolution of this construction within the limits of a document of acceptable length. With regard to the logical structure of the document, the most important elements of constructional philosophy will be presented here, mathematical solutions to the presented problems will be outlined, and the original solutions in this submission will be presented. It may be appropriate to read the same applicant's prior documents in this field to facilitate understanding of the description, but this should not be necessary.

The experience with the decoder and encoder as described in the June '96 and November '96 applications and the provisional patent application of September '97 led to further improvements that have not yet been described. This application will present the essential features of the improved encoder and decoder of the present invention and will claim new features that have been added since US Patent Application No. 08 / 742,460.

1. General description of the decoder

The decoder will be described in this application as consisting of two separate parts. The first is a nut that divides two input channels into five output channels, which are commonly referred to as center, left front, right front, left rear, and right rear. The second part consists of a series of delays and filters that modify the spectrum and levels of the two rear outputs. One function of the second part is to derive an additional pair of outputs, for the left side and for the right side, when a seven-channel version of the decoder is required. In the application No. 08 / 742,460 the second part was not explicit - two additional channels • * · · · · · · φ • φ φ φ φ φ φ φ φ φ φ φ φ φ φ φ φ φ φ φ · Were derived from an additional pair of matrix elements in the original matrix.

In the mathematical equations describing the decoder and the encoder, standard typographic conventions will be used for most variables, simple variables will be in italics, while vector quantities are shown in bold lower case, and matrices are represented by bold upper case. Matrix elements that are coefficients from said input channels resulting from said output channels will be shown in normal capital letters. Some simple variables, such as Ir and cs, are indicated by two-letter names that do not represent the results of two separate simple variables. Other variables such as 1 / r and c / s in some sense represent left to right and center (front) to surround (back) values, but in terms of control signal voltages derived from these ratios. These conventions have also been used in previous US patents and patent applications cited herein above. Program segments in language

Matlab will also be differentiated through the use of a different font shape and point size and through space separation. The equations will be numbered to distinguish them from the Matlab assignment statements and to reference them for the specific characters described in this description.

Giant. 1, which is identical to FIG. 4 in US Patent Application No. 08 / 742,460, shows a block diagram of a first part of a decoder, a matrix 90 for converting two channels to five channels. The left half of FIG. 5, separated by a vertical dashed line 30, illustrates a means for deriving two

* • ··· control voltages 1 / r and c / s. These voltages represent the extent to which the input signals have their own or coded directional components in the left / right and front / rear directions, respectively. This part of Fig. 1 will not be explicitly discussed in this application because it has been fully described in the above-mentioned patent application, which is incorporated herein by reference.

In Fig. 1, a directional detection means of a decoder 90 comprising elements 92 to 138 is followed by a 5x2 matrix to the right of the vertical dashed line. The elements 140-158 of this matrix determine the size of each input channel that is linearly combined with the second input channel to form each output channel. These matrix elements are considered real. (The case of complex matrix elements has been described in US Patent Application No. 08 / 742,460 and will not be discussed in this specification.) Matrix elements are functions of two control voltages 1 / r and c / s. US Patent Application No. 08 / 742,460 presented mathematical formulas for these functions. Part of the novel solution of the present invention lies in improvements to these mathematical formulas. In this description, these formulas are presented graphically and an explanation is given as to why these graphical representations have the shape shown.

2. Brief description of control voltages

As shown in Fig. 1, the control voltages c / s and 1 / r are derived from the logarithm of the ratio of the left input amplitude at terminal 92 to the right input amplitude at terminal 94, and the logarithm of the sum of amplitude and amplitude difference. When • 0 • 0

0 ·

0 «• · · 0000 * 0 0

0 0 0

0 0 ♦

0 0 0

In describing the matrix elements, it is preferable to express 1 / rac / s as angles that vary from +45 degrees to -45 degrees. In decoders VI.11 and V2.01, these voltages have decibel units. It is possible to convert these control parameters to angles where ⁵ lr = 90 - arctane (10 ^Λ ((1 / r) / 20)) ... (la) en = 90 - arctane (10 ^Λ ((c / s) / 20)) ... (la)

The angles Ir and cs determine the extent to which the input signals have a directional component. For example, when the inputs to the decoder are separated, both Ir and cs are zero. For a signal coming only from the center, Ir is zero and cs is 45 degrees. For the signal coming from behind, Ir is zero and cs is -45 degrees. Similarly, for the signal coming from the left, my Ir value is 45 degrees and the cs value is zero, and for the signal coming from the right, the Ir value is -45 degrees and the cs value is zero. In our construction, we assume that the encoder that produces the encoded signal has such properties that when encoding the left rear signal, Ir = 22.5 degrees and cs =

22.5 degrees. Similarly, the signal applied to the right rear input to the encoder produces Ir = 22.5 degrees and cs = 22.5 degrees.

It can be seen from the definitions 1 / r and c / s and the derivation of Ir and cs that the magnitude of the absolute value of Ir and cs cannot be greater than 45 degrees. The allowed values of Ir and cs create a surface bounded by the geometric spots described by the formula abs (lr) -abs (cs) = 45 degrees. Any input that produces lr and cs values that lie along the boundary of this surface is fully localized - that is, it consists of one sound that has been encoded to come from a particular direction.

::. ··························. · * I •. ; »·· · · ·

In this application, matrix graphs will be extensively used as a function of this two-dimensional surface. Generally, the derivation of matrix elements will be different in the four quadrants of this surface. In other words, the nut elements are described differently depending on whether the steering is forward or rearward and whether it is left or right. Significant efforts are being made to ensure that this surface is continuous across the boundaries between quadrants. Any loss of this continuity is one of the problems of the VI.11 decoder, which will be addressed in this application.

3. Frequency-dependent elements

The matrix elements shown in FIG. 1 are real and therefore frequency independent. All signals in the inputs will be routed to the outputs depending on the derived angles Ir and cs. (In the prior art, low frequencies and very high frequencies are attenuated when deriving Ir and cs from the input signals through the filters shown in Fig. 1.

2Q are not shown. But the matrix itself is broadband.)

In practice, it has been found that there are several advantages when using frequency-dependent circuits for signals beyond the matrix. One of these frequency-dependent circuits - a phase shift network 170 at the right side output 180 of FIG. 1 - has been described in

US Patent Application No. 08 / 742,460 and will not be further discussed in this specification.

Giant. 2 shows a five-channel version of additional frequency-dependent circuits. These circuits do not have fixed parameters. Frequency and level behavior is dependent on the control values Ir and cs. These circuits are used for several I * I * IV * I * ^B * I * I * I * I * I * II I • I III I ti I · Purposes. First, in both the five-channel and seven-channel decoders, additional elements allow the apparent volume of the rear channels to be adjusted when the control is neutral (Ir and cs are 0) or forward (cs> 0). In US patent application

No. 08 / 742,460, this attenuation was performed as part of the matrix itself and was frequency independent. It has been found by theoretical studies and listening tests that it is highly desirable that the low frequencies be reproduced from the listener. In the decoder presented in this description, only the high frequencies are attenuated by the variable low pass filters 182, 184,

188 and 190.

This is accomplished by attenuating frequencies above 500 Hz in the rear channels by elements 188, 190, and above 4 kHz by elements 182, 184 when the control is almost always 15 neutral or forward, using the background control signal 186, which in this application will be defined later. The occasional occurrence of backward-controlled sounds reduces attenuation, a feature that automatically differentiates the surround encoding material from conventional 2-channel material.

Other elements 192, 194 in the five-channel version modify the audio spectrum when the control is directed backwards (cs <0) using the c / s signal 196 so that the speakers appear to be placed behind the listener, even when their actual position is on the sides. This modified left surround and right surround signals appear in terminals 198 and 20Q, respectively. Further details of this circuit will be given in this description below.

Giant. 3 shows a seven-channel version of frequency dependent elements. As before, the first set of filters 182,184; * · · · · · · Ί.;. · · · · · · · · · · · · · · · · · · · · · · · · · · ·. · · · · · * · * * *

188, 190 attenuates the upper frequencies of the side and rear outputs when the control is neutral or forward, again under control by the background control signal 186. This attenuation also results in a more forward-acting audio image and can be adjusted to the listener's requirements. As the control represented by the c / s signal 196 moves backward, the additional circuits 202, 204, 206, 208 act to distinguish the side outputs from the rear outputs. As the control moves backward, the aforementioned attenuation in the side speakers is first removed by the elements 204 and 206 to produce side-oriented sound. As the control moves further backward, the attenuation of the elements 204 and 206 is then restored and increased. As a result, the sound moves smoothly from the front speakers to the side speakers and then to the rear speakers, which have a delay of approximately 10 ms, generated by delay elements 202 and 208. Because low frequencies are not affected by these circuits, low volume The frequency in the side speakers (which affects the perception of space) is not affected by this shifting sound. Further details of the circuit of FIG. 3 will be given below.

4. General description of the encoder

Giant. 4 is a block diagram of an encoder designed to automatically mix five input channels into two output channels. The architecture of this encoder is relatively different from that described in US Patent Application No. 08 / 742,460. The aim of this new construction is

2Q to maintain the musical balance of the five-channel original while providing phase / amplitude markers that

0

0

0 • 0 «· 00 0 0 0

0 0 0 0

0 0

0 0 ·

allows you to remove the original five channels by a decoder. The previous encoder had similar goals, but improvements were made in the procedures used to achieve these goals. Maintaining the music balance is very important in the encoder. One of the primary purposes of the encoder is to automatically create a two-channel mix of five-channel recording that will be played back in a conventional two-channel system with the same artistic value as the five-channel original. The new encoder design includes active elements to ensure that the musical balance is maintained.

Compared to the encoder of the November '97 patent application, the new design allows input signals to be inserted between any of the five encoder inputs. For example, sound may be input from the front left input to the rear right input. When the resulting two-channel signal is decoded by the decoder described in this application, the result will be relatively close to the original audio. Decoding with the previous surround decoder was also similar to the original.

A detailed description of the encoder will follow in this description below.

5. Construction targets for active matrix elements of the decoder 25

The most basic objectives of the present invention are identical to those of the previous decoders of the same applicant, in particular the decoder described in US patent application No. 08 / 742,460 - The invention is a surround sound decoder having variable matrix values designed to limit directionally encoded audio components in outputs. 0000 0 0 0 ' _A 0000 0 00 · ·· ·· not directly involved in their reproduction in the intended direction; improve directional encoded audio components in the outputs that are directly involved in reproducing them in a designated direction so as to maintain a constant overall power for these signals; while maintaining a large distance between the left and right channel components of the non-directional signals, regardless of the control signals; and keeping the volume, defined as the overall sound power level of the non-directional signals, as effectively constant, whether or not the directional coded signals are present and regardless of their intended direction, if present.

Most of these goals seem to be shared with all matrix decoders. The novelty of the solution in this application lies in part in knowing how to implement the above rules 15 accurately, and in part in knowing not to apply the above rules. However, much of the methodology of US Patent Application No. 08 / 742,460 is retained. One of the most important previous goals is to explicitly maintain a large distance between the left and right channels of the decoder at all times.

All previous four-channel decoders are not able to maintain back separation since they provide only one rear channel. Five-channel decoders from other manufacturers solve the distance in many ways. The decoder described in this application fulfills this objective in a similar way to the VI.11 decoders - but at the same time 25 also fulfills the additional objectives.

US Patent Application No. 08 / 742,460 also describes many minor design improvements, such as circuits for improving control signal accuracy and a variable phase shift network for switching the phase of one of the rear channels during strong rearward (rearward) steering. These decoder characters * »44 4444

4444 4 44 «« 444

VI. 11 are retained in the new construction, but will not be covered in this document.

In Fig. 4, the front input signals L, C and R are applied to the input terminals 50, 52 and 54, respectively. The signals L and R arrive directly into the adders 278 and 282, respectively, while the C signal is first attenuated by the female factor in attenuator 372 before being applied to the inputs of both adders 278 and 282. element 374 and then fed to both adders 278 and 282.

Surround input signals LS and RS are applied through two input terminals 62 and 64, respectively, into two separate paths: for the LS signal, the path through attenuator 378 has a gain fs (ls, ls) and the RS signal passes through the corresponding attenuator 38Q. with gain fs (r, rs). Outputs from these elements are fed to interconnecting elements 374 and 386 having a gain factor -crx, where crx is nominally 0.383. The coupler signals from these elements are applied to the adders 392 and 394, which also receive the attenuated L5 and RS signals from the attenuation elements 388 and 392 with a factor of 0.91. The outputs of the adders 392, 394 are fed to the inputs of the adders 278. 282. This places the respective elements within 45 degrees to the left and right of the center rear in the decoded space.

The next signal branch conducts the LS and RS signals through attenuation elements 376 with gain fc (1,1s) and 382 with gain fc (r, rs) and then through a similar arrangement of the elements 396, 398, 402, 404, 406 and 408, where Adders 406 and 408 have outputs representing left rear and right rear inputs, which are located 45 degrees to the left and right of the center rear as mentioned above. However, these signals now pass through the phase shift elements 234 and 246, respectively, while the left and right signals from the adders 278 and 282 pass through the element. 286 and 288 phase shifts, respectively. Each of these phase shift elements is a phase adjuster, the phase response being φ (ί) for elements 286 and 288 and cp (f) -90 ° for elements 234 and 246. Calculating the values of the components required in these filters is well within the art known and will not be further discussed in this description. As a result, the outputs of the adders 406 and 408 are delayed by the outputs of the adders 278 and 282 by 90 ° at all frequencies after passing through the phase adjustment networks as shown in Fig. 4. The network outputs 234 and 286 are now combined in the adder 276 to produce output signal A (left) at terminal 44., while the outputs of filters 246 and 288 are combined by an adder 280 to produce output signal B (right) at terminal 46.

The fs and fc amplification functions are designed to allow strong surround signals to be presented in phase with other sounds, while weak surround signals pass through a 90 degree phase shift path to maintain constant power for separate music signals. The crx value can also change and change the angle at which the surround signals are heard.

6, Constructional improvement since Application No. 08 / 742,460

One of the most noticeable improvements in the present invention with respect to the solution of US Patent Application No. 08 / 742,460 is a change in the central matrix elements and the left and right matrix elements when the signal is controlled in the central direction. It has been shown that there are two problems for coding and decoding for the central channel with the preceding ftft ft ft ft ft ft ft ft coding and decoding.

The most obvious problem is that in a five-channel matrix system, the use of the central channel is inevitably in conflict with the aim of keeping the left / right spacing value as high as possible. If the matrix is to produce a sensitive output from a conventional two-channel stereo material when the two input channels do not have a left / right component, the center channel must be controlled by the sum of the left and right input channels. Thus, both the left decoder input and the right decoder input will be reproduced by the center speaker, and sounds that were originally only in the left (or right) channel will be reproduced from the center. The result must be that the apparent positions of these sounds are pulled into the middle of the room. The extent to which this happens depends on the volume of the central channel.

US Patent Nos. 4,862,502 and 5,136,650 used matrix elements having a minimum value of 3 dB compared to the left and right channels. When the inputs to the decoder were separated, the volume of the center channel was the same as the volume of the left and right channels. As the control moved forward, the central matrix elements added an additional 3 dB. The effect of this high volume is a severe limitation of the front audio image width. Tools that should sound left and right in the audio image are always pulled toward the center 25 of the audio image.

US Patent Application No. 08 / 742,460 used central matrix elements that had a minimum value of 4.5 dB less than the previous values. This minimum value was chosen based on listening tests. This attenuation ³⁰ causes a comfortable stretching of the front image when the input is

........

• 9 9 9 · · · 9 * 9 · material separated - as with orchestral music. The front image is not dramatically narrowed. In US Patent Application No. 08 / 742,460, as the control moves forward, these matrix elements add, ultimately reaching the values used in the Dolby matrix.

Experience with the decoder Vl.11 has shown that although the volume reduction in the central channel resolved the spatial problem, the power balance in the input signals was not maintained through this matrix. Mathematical analyzes have shown that the error occurred only in the Vl.ll decoder, but that the error was also in the Dolby decoder and in the previous decoders of the same applicant. Paradoxically, although the center channel was too strong to reproduce the Front Image Width, it was also too weak to maintain power balance. This problem was particularly serious for the Mandel decoder, which is a standard Dolby decoder. In a standard Dolby decoder, the rear channels are thicker than in the decoder of U.S. Patent No. 4,862,502 of the same applicant. As a result, the central channel must be stronger to maintain the power balance. Loss of power balance in the central channel was a continuous problem for the Dolby decoder. For a Dolby decoder, it is recommended that the person mixing the sound should always listen to the balance achieved by the matrix so that the power loss in the matrix can be compensated during the mixing process. Unfortunately, modern movies are blended for 5-channel distribution, and automatic two-channel encoding can lead to audio-level dialogue problems.

Further analysis and listening tests have shown that movies and music require different solutions to the power balance problem. For films, it has been found to be most useful to retain the left and right front matrix elements of the U.S. patent application Serial No. 08 / 742,460. These elements eliminate the central channel information from the left and right front channels as much as possible. This minimizes the penetration of dialogue into the left and right front channels. In this new film construction, power balance is corrected by changing the central matrix elements so that the center channel volume increases faster than a standard decoder when the control moves forward (as cs becomes greater than zero). the value of the central matrix elements was higher than the values in the standard decoder, because this state is only achieved when the central channel is active. It is only necessary that the central level be thicker than that of a standard decoder when there are approximately the same levels in the central channel and in the left and right channels.

With this film strategy, the central channel volume is increased to maintain power balance in the input signals, while the central channel component in all other outputs is minimized. This strategy seems to be optimal for films where the main use of the central channel is for dialogues and where dialogue from places other than the center is not expected. The main drawback of this strategy is that at any moment there is considerable steering to the center - as is the case with many types of popular music, thereby narrowing the front picture. But the advantages for the film - that is, minimal penetration of dialogue into the front channels and excellent performance balance - counterbalance this disadvantage.

A different strategy is possible for music. In this case, the volume of the central channel will be allowed to increase at the same rate as in the US patent application 9,999,989 9,999,989 9,999,999,999

No. 08 / 742,460 up to the middle steering value where cs> 22.5 degrees. To maintain the musical balance, the left and right front matrix elements have been changed so that the central component of the input signals is not completely removed. The amount of center channel component in the left and right front elements is set so that the audio power from all decoder inputs corresponds to the audio power in the input signals - without excessive center volume.

With this strategy, all three front speakers θ reproduce the center channel information presented in the original coded material. The most usable version of this strategy limits the control action when the central input component is 6 dB stronger in the central output than in any of the two other front outputs. This is achieved by 5 simply limiting the positive cs value.

This new strategy - allowing the center channel component to come from all three front speakers, and limiting control when the center is 6 dB louder than the front left and right channels - is an excellent solution for all types of music. Both the encoded five-channel mixtures and the conventional two-channel mixtures are decoded with a stable center and a corresponding distance between the center channel and the left and right channels. It should be noted that compared to previous decoders, the distance between the center and the left and right sides is not intentionally complete. The signal to be received from the left is eliminated from the central channel, but vice versa. For music, the large lateral spacing and stable front image, as offered by this strategy, offsets this loss of complete spacing. Listening tests with this setting for movies have shown that although some of the dialogs come from • · · · »«. _w w 000 000 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 The result is pleasant and does not disturb the listener. Therefore, a listener listening to a movie with a decoder set for music will not have a reduced artistic level of the movie. However, listening to music with a decoder set for a movie is much more problematic.

Obviously, another most obvious improvement in the present invention is to increase the spacing between the front channels and the rear channels when the signal is controlled to the left front or rear channel.

LQ of the left rear direction. Decoder VI. 11 used matrix elements of US Patent No. 4,862,502 for the front channels under these conditions. These matrix elements do not completely eliminate the rear control signal unless it has been driven to the fully rearward position - halfway between the left rear and right rear channels. When the control was left rear or right rear (not fully rear), the left or right front output had an output signal 9 dB less than the corresponding rear output. In the present invention, the front matrix elements are modified to eliminate front sound when the steering is anywhere between the left rear and right rear channel.

7, Improvement of rear matrix elements

Improvements of the rear matrix elements are not immediately apparent to the ordinary listener. These improvements correct various errors in the continuity of matrix elements across the boundaries between quadrants. These improvements also improve the power balance between controlled signals and uncontrolled signals under different conditions. The mathematical description of the matrix elements, which will be described in the description below, includes these improvements of the rear matrix elements.

8. Detailed description of active matrix elements

Matiab

The mathematical apparatus used to describe matrix elements is not based on the continuous functions of the variables cs and Ir. Generally, there are conditions, absolute values, and other non-linear modifications of formulas. For this reason, the matrix elements will be described in the programming language. Matiab provides a simple way to validate patterns graphically. Matiab is very similar to Fortran or C. The main difference is that Matiab variables can be vectors - that is, each variable can represent an array of numbers in a sequence. For example, you can define the variable x as follows:

x = 1:10;

this Matiab specification produces a string of ten numbers with values from one to ten. The variable x includes all ten values. The vector is described by a matrix of 1 times 10. One can access or manipulate individual numbers within each vector. For example:

x (4) = 4;

sets the fourth member of vector x to 4. Variables can also represent a two-dimensional matrix. Individual elements in a matrix can be labeled in a similar way:

X (2.3) = 10;

0 assigns a value of 10 to the second row and the third column of matrix X.

The detailed description of the matrix elements that follows is almost identical to the description published in reference [2]. However, this text has been somewhat improved. The main differences are:

1. Reference [2] contains the tv character of the matrix. This feature reduces the level of rear outputs when the control is front or neutral. In this application, this function is achieved by means of frequency dependent circuits following the matrix. Therefore, the tv matrix correction has been omitted.

2. The section describing the central matrix elements has been modified to include references to a film strategy, a music strategy, and a strategy that limits the effect of the music setting. Reference [2] described only musical settings without limitation.

9. Matrix decoders in equators and graphical expressions

In reference [1] a matrix decoder construction was presented that could be described by elements of the matrix n x 2, where n is the number of output channels. Each output can be viewed as a linear combination of two inputs, where the coefficients of the linear combination are given by elements in the matrix. In this application, elements are identified by a simple combination of letters. Reference [1] described a five-channel and seven-channel decoder. The conversion from five channels to seven channels is now performed in the frequency-dependent part of the decoder, so that only a five-channel decoder will be described in this description.

φ φ

• Φ • ΦΦΦ »φ φ φ • · φ φ

Φ · Φ · φ · φ ·

From the given symmetry, it is clear that the behavior of only six elements - the central elements, the two left front elements and the two left rear elements - should be described. These elements are therefore:

CL: Matrix element for the left input channel to the central output

CR: Matrix element for the right input channel to the central output

LFL: Left input channel to the front left output

LFR: Right input channel to the left front output

LRL: Left input channel to the left rear output

LRR: Right input channel to left rear output

These elements are not constant. Their values change as a two-dimensional function of the apparent direction of the input sounds. Most phase / amplitude decoders determine the apparent direction of the input by comparing the amplitude ratio of the input signals. For example, the degree of control in the right / left direction is determined from the ratio of the amplitude in the left input channel to the amplitude in the right input channel. In a similar way, the degree of control in the front / rear direction is determined from the ratio of the sum amplitude to the difference of the input signals. The procedure for determining these control directions will not be discussed in this description, although the Logic 7 decoders differ from standard decoders considerably in how this determination is realized. This is based on the assumption that the guidelines have already been determined. In this description, these control directions will be represented as angles - one angle for left / right direction (lr) and one angle for front / rear (center / surround) direction (cs). These two control directions are variables marked with a sign. When both lr and cs · · jsou ♦ · · · · · · · · jsou jsou jsou jsou jsou jsou jsou jsou jsou jsou jsou zero, the input signals are uncontrolled - that is, the two input channels are separate.

When an input consists of one signal that has been directionally encoded, the two control directions take their maximum values. However, under these conditions they are not independent. The advantage of representing control values as angles is that when there is only one signal, the absolute value of the two control values must give a sum of 45 degrees. When the input contains some separate material together with a strongly controlled signal, the sum of the absolute values of the control values must be less than 45 degrees.

Ir | + | en | <45 ... (1)

If matrix element values are plotted in the two-dimensional plane formed by the control values, then the center of that plane will have a value of (0,0) and the valid values for the sum of the control values will not exceed 45. In practice, the behavior of nonlinear filters - Application No. 08 / 742,460 claimed a circuit that limited lr or cs so that their sum did not exceed 45 degrees. This claim will not be further discussed in this description. We assume that the mathematical apparatus for matrix elements will work well when values are exceeded. When plotting matrix elements graphically, values will be automatically reset when the valid sum of input variables is exceeded. This allows direct monitoring of the behavior of elements along the boundary trajectory - a trajectory followed by a strongly controlled signal. Graphic expressions were created using Matlat language. In Matlab it is an uncontrolled position (46, 46, because Matlab requires angular variables to be 1 larger »00 0 • 00 0 • 0 00

0 · 0 · 00 00 0 0 0 0 0

0000> 0 · 0 than the actual angle value. Perhaps this circumstance will not be excessively confusing.

Previous structures for matrix decoders tended to consider only the matrix behavior for a strongly controlled signal, that is, the behavior of matrix elements around the boundary of the respective surface. But this is a fundamental mistake in the preview. When studying real signals - either film or music - it is possible to find that the boundary of this surface is rarely reached. For the most part, the signals fluctuate around the center of the plane - slightly toward the central channel. The behavior of the matrix under these conditions is essential for sound. When comparing the elements of the present invention with the preceding elements, a significant increase in surface complexity in the intermediate regions can be seen. It is this complexity that is the basis for improvements in sound reproduction.

Such complexity has a certain value. The original design of the same applicant from 1987 - see the 1989 patent - was simple to implement with analog components. The new features are designed to be almost entirely described by one-dimensional lookup tables that are trivial in digital realization. Construction of an analog version with similar effect and power is possible, but it is definitely not trivial.

In this application, several different versions of matrix elements are compared. The oldest are the elements of the 1989 patent of the same applicant. These elements were used in the first surround processor of the same applicant and are identical to those of a standard (Dolby) surround processor in the left, center and right channels (but are not identical in the surround channels). In the present construction, the surround channel is processed symmetrically with respect to the central channel. In a standard (Dolby) decoder, the surround channel is treated differently, and this will be discussed later in the description below.

The elements presented here do not always have the appropriate scale. In general, they are presented such that the uncontrolled value of the non-zero matrix elements is one for any given channel. In practice, the elements are usually on a scale such that the maximum value of each element is one or less. In any case, in the final result, the scale of the elements is additionally changed during the calibration procedure. The matrix elements presented in this description are to be considered as scalable elements by means of suitable constants.

IQ. Left front matrix elements ’= 9 of the same applicant's patent

Suppose that cs and Ir are control directions in degrees in the center / surround and left / right respectively.

In the '89 patent, the equations for the front matrix elements are given as:

In the front left quadrant:

LFL = 1 - 0.5 * G (en) + 0.41 * G (ir) ... (3a)

LFR = -0.5 * G (en) ... (3b)

In the right front quadrar.ru:

LFL = 1 - 0.5 * G (en) ... (3c)

LFR = -0.5 * G (axis). · (3d)

In the left rear quadrant (s is negative!):

LFL = 1 - 0.5 * G (en) + 0.41 * G (Ir) ... (3e) • · ftft * ftftft ft «·· • ft ftft ftftft • ft ·

(3g). (3h) of the patent:

LFR = -0.5 * G

In the right rear quadrant:

LFL = 1-0.5 * G

LFR = -0.5 * G

The function G (x) was determined experimentally in 1989 and is specified mathematically in the '91 patent. This function varies from 0 to 1 as x varies from 0 to 45 degrees. When the control is in the left front quadrant (Ir and cs are both positive), G (x) is equal to 1- | r | / | l | where | r | and | 1 | are the right and left input amplitudes. G (x) can also be described in terms of control angles using different formulas. One such pattern is outlined in the '91 patent and others will be listed later in this document. See Fig. 5 and Fig. 6 for graphical representations of LFL and LFR matrix elements plotted in a three-dimensional representation against the Ir and Cs axes.

In reference [1], these elements have been improved by adding the requirement that the volume of the uncontrolled signal should be constant regardless of the direction of control. Mathematically, this means that the mean quadratic sum of the LFL and LFR matrix elements should be constant. In this description it was stated that this target should be changed in the direction of control, i.e. when the control is fully left, the sum of squares of these elements should increase by 3 dB. Giant. 7 shows the sum of squares of these elements, thereby demonstrating that the above matrix elements do not meet the constant loudness requirement. In Fig. 7, it should be noted that the value is constant at 0.1 along the axis from uncontrolled to right. Uncontrolled to the left increases by 3 dB to one and uncontrolled to the central or rear decreases by 3 dB to 0.5. This part of the graph is obscured by the tip · * · V »-» - ·

9999 9 9 9 9 9 9 · · 9

9,999 * 999

9 · 9 · 9 9 «9 9« · 9 left. The rear direction level is the same as the central direction level.

In Application No. 08 / 742,460 and reference [1], the amplitude errors seen in Figure 7 were corrected by replacing the function G (x) in matrix equations with the functions sine and cosine. See Figure 8 for the resulting graph of the sum of squares of the corrected LFL and LFR elements described by Equations (4a) to (4h) below.

Note the constant value of 0.71 throughout the right half of the plane and a slight increase to one towards the left peak.

For left front quadrant: LFL = cos (en) + 0.41 * G (lr) ... (4a) LFR = -sin (en) ... (4b) For right front quadrant: LFL = cos (en) ... (4c) LFR = -sin (en) ... (4d) For left rear quadrant: LFL = cos (-en) + 0.41 * G (lr) ... (4e) LFR = sin (-en) . . · (4f) For right rear quadrant: LFL - cos . . (3g) LFR = sin (-en) ... (2h)

999 »* • · 999

9 9

99 «

IIM • 9

9 «9

11, Improvement of left front nut elements

In March 1996, certain changes to these matrix elements were made by the same applicant. The basic functional dependence was retained, but reinforcement along the cs axis at the front was added and a weakening along the cs axis at the rear was added. The reason for the amplification was to improve the operation of stereo music that was directed forward. The purpose of attenuating the rear was to increase the distance between the front channels and the rear channels when the stereo music was directed backwards.

For left front quadrant:

LFL = (cos (en) + 0.41 * G (lr)) * boostl (en) ... (4a)

LFR = (-sin (en)) * boostl (en) ... (4b)

For right front quadrant:

LFL = (cos (en)) * boostl (en) ... (4a)

LFR = (-sin (en)) * boostl (en) ... (4d)

For left rear quadrant:

LFL = (cos (-en) + 0.41 * G (lr)) / boost (en) ... (4e)

LFR = (sin (en)) / boost (en). (4f)

For right rear quadrant:

LFL = (cos (en)) / boost (en) ... (3g)

LFR = (sin (en)) / boost (en) ... (3h)

The function G (x) is the same as in the '89 patent. When expressed with angles as input, it can be expressed as follows:

G (x) = 1 - tan (45-x) • · · (6) • · «· φ φ ί φ φ φ φ

• Φ

Φ φ • φ φφ

The boostl (cs) function, as used in March 1997, was a 3 dB linear gain applied during the first 22.5 degrees of control and decreasing back to 0 dB in the following 22.5 degrees. Boost (en) is given by corr (x) in the Matlab code below (comment lines are always indicated by the percent% symbol).

calculating the amplification function by + 3dB at 22.5 degrees% corr (x) goes up to 3 dB and stays here, corrl (x) goes up% and then down again for x = 1:24; % x has a value of 1 to 24, representing 0 to% 23 degrees corr (x) = 10 ^Λ (3 * (x-1) / (23 * 20)); % goes up to 3 dB over% this range corrl (x) = corr (χ), end for x = 25:46% goes back down for corrl over this range% 24 to 45 degrees corr (x) = 1.41;

corr1 (x) = corr (48-x);

end

The equation (5a) to (5h) resulting in the LFL is illustrated in Figure 9. It should be noted that the amplification as the control moves towards the center is applied both along the axis Ir = 0 and along the boundary between · 4 · 4 · 4 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · It may also be noted that the lowering of the level as the steering moves backwards.

The operation of the March 1997 circuit can be improved. The first problem is in the behavior of driving along the boundaries between left and central and between right and central. As the strong one signal expands from left to central, it can be seen in Fig. 9 that the LFL value of the matrix element increases to a maximum halfway between the left and central. This increase in value is an unintentional consequence of a deliberate increase in the level for the left and right main output as the central signal is added to stereo music.

When the stereo signal is directed forward, it is desirable that the left and right front outputs increase the level to compensate for the removal of the combined component by the matrix from these outputs. However, the procedure used to raise the level under these circumstances would only be implemented when the input component Ir is minimal, that is, when there is no left or right steering. The procedure chosen to implement this increase in March 1997 was independent of the Ir value and resulted in an increase in level when a strong signal crossed the boundaries between quadrants.

Gain is needed only along the axis Ir = 0. If Ir is nonzero, the matrix element should not be reinforced. This problem can be solved by using the summation member to the matrix elements instead of the multiple member. Thus, a new control index is defined, bounding the bounds with a value of cs with the following code in Matlab:

• · »

Suppose that both lr and cs> 0 - are in the left front quadrant (suppose that cs and lr take values according to Matlab conventions, ranging from 1 to 46)% finding boundary c / s 5 if (cs <24) bcs = cs- (1y-1);

if (bcs <l)% this limits the maximum value of bcs = 1;

end else bcs = 47-en- (lr-1);

if (bcs <1) bcs = 1;

end end

If cs <22,5 and lr = 0 (in the Matlab convention cs <24 and lr = 1) bcs is equal to cs. However, as lr increases, bcs will decrease to zero. If cs> 22.5, then with increasing lr the bcs also decreases.

To find the necessary correction function, a difference has now been found between the reinforced matrix elements and the non-reinforced matrix elements along the lr = 0 axis. This difference is called cos_tblj? Lus and sin_tbl_plus. Using the Matlab code:

a = 0:45; % defines the vector in one-step steps, • · • 0 • 0 · ♦

0 0 • 00 0 0 · • 0 • 000 ·% and has values from 0 to 45 degrees al = 2 * pi * a / 360; % Conversion to Tools% defines sine and cosine tables as well as amplification tables

Front quadrant sin_tbl - sin (al); cos_tbl = cos (a1); cos_tbl_plus = cos (a1) * corrl (a + 1);

cos_tbl_plus = cos_tbl_plus - cos_tbl;% that is one used cos_tbl_minus = cos (al) / corr (a + 1); sin_tbl_plus = sin (a1) * corrl (a + 1);

sin_tbl_plus = sin_tbl_plus - sin_tbl;% that is one used sin_tbl_minus = sin (al) / corr (a + 1);

The vectors sin_tbl_plus and cos_th> l_plus are the difference between conventional sine and cosine and enhanced sine and cosine. They will now be defined

LFL = cos (cs) + 0.41 * G (lr) + cos_tbljplus (bcs} .. (7a)

LFR = -sin (en) - sin_tbl_plus (bcs) ... (7b)

LFL and LFR in the front right quadrant are similar, but without the term + 0.41 * G. These new definitions lead to the matrix element shown graphically in Figure 10.

In Fig. 10, it can be seen that the new element has a corrected amplitude along the border between the left and the central as well as along the border between the central and the right.

• 0000 00 0 · 0 · 0 «0 0 0

0000 00 00 00 00 00

Control in the rear quadrant is not optimal. When the steering is rearward, the above matrix elements are given by:

LFL = cos_tbl_minus (-en) + 0.41 * G (-en) ... (8a)

LFR = sin_tbl_jninus ... (8b)

These matrix elements are almost identical to those in the '89 patent (Considering the case where a strong signal goes from left to back. The elements of the '89 patent were designed to completely cancel the output from the front left output only when this the signal was fully controlled backwards (cs = -45, Ir = 0), but in the Logic 7 decoder it is desirable that the output from the left front output be zero when the coded signal reaches the left rear direction (cs = -22.5 and lr = The left front output should remain at zero as the signal moves further backwards The matrix elements used in March 1997, the elements listed above - result in the front left channel output being approximately -9 dB when it is This level difference is sufficient for a good matrix operation but not as good as it could be.

This activity can be improved by changing the LFL and LFR matrix elements in the left rear quadrant. It should be noted that here we are looking at how matrix elements change along the boundary between left and rear. The mathematical procedure referred to in reference [1] can be used to find the behavior of elements along this boundary. Suppose the left front output amplitude decreases with the function F (t) as t varies from 0 (left) to -22.5 degrees (left rear). This procedure defines matrix elements:

LFL = cos (t) * F (t) -7 + sin (t) * (sqrt (1-F (t) ^Λ 2)) ... (9a)

0 0 »0 0 Β 0 ♦ φ φ φ φ 0

ΦΦ 00

LFR = (sin (t) * F (t) +/- cos (t) * (sqrt (1-F (t) ^Λ 2))) ... (9b)

If we choose F (t) = cos (4 * t) and select the correct sign, the following equations are simplified to:

LFL = cos (t) * cos (4 * t) + sin (t) * sin (4 * t) ... (9c)

LFR = (sin (t) * cos (4 * t) -cos (t) * sin (4 * t) ... (9d)

Figure 11 is a graphical representation of these LFLs (solid line) and LFR (dotted line) versus t. (A small jump in the middle is due to the absence at the point at

22.5 degrees because all Matlab angles are integers.)

These elements work well - the front left output is confined smoothly to zero as t varies from 0 to 22.5 degrees. It is desirable that the output remains zero as control continues from

22.5 degrees to 45 degrees (far back) along this part of the boundary between quadrants:

LFL = -sin (t) ... (9a)

LFR = cos (t)

... (10b)

It should be noted that these matrix elements are something completely different from matrix elements along the boundary lr = 0, where in reference [1] these values were:

... (10c)

...(ship)

LFL = cos

LFR = sin

It can be noted that the matrix elements are designed to behave correctly with a strongly controlled signal - when both cs and lr have maximum values. Previous matrix elements were successful for signals where lr is close to zero - these are stereo signals that were * 0 0 0 • ♦ «·

0 0 ·

0 · 0 ·

0000 0 «0 moved back. However, there is a need for a smooth transformation of the previous matrix elements into new matrix elements as lr and cs approach the boundary. Linear interpolation could be used. In the processor used in Lexicon products, where multiplication is challenging, it is better to define a new variable - minimum lr and cs, as defined by the Matlab section below:

finding a new border parameter bp = x;

if (bp> y) bp = y;

end and a new correction function that depends on bp:

for x = 1:24 x = 2 * pi (46-x) 360;

front_boundary_tbl (x) = (cos (ax) - sin (ax)) / (cos (ax) + sin (ax));

end for x =: 46 x = 2 * pi * (x-1) / 360;

front_boundary_tbl (x) = (cos (ax) - sin (ax)) / (cos (ax) + sin (ax));

end

Then LFL and LFR in this quadrant will be defined as:

• 4 4

4 4 4

44

LFL = cos (en) / (cos (en) + sin (en)) -front_boundary_tbl (bp) + 0.41 * G (lr). . . (lla)

LFR = sin (en) / (cos (en) + sin (en)) -front_boundary_tbl (bp)

... (lib)

It is worth noting the correction cos (cs) + sin (cs). When cos (cs) is divided by this factor, the function l-0,5 * G (cs) is achieved, which is the same as the Dolby matrix in this quadrant. When this factor is divided by sin (cs),] _0 will achieve the previous function + 0.5 * G (cs).

Similarly in the right rear quadrant:

LFL = cos (en) / (cos (en) + sin (en) = l-0,5 * G (en) ... (12a)

LFR = sin (en) / (cos (en) + sin (en)) = 0,5 * G (en) ... (12b)

See Fig. 12 and Fig. 13 for a graphical representation of these values.

In Fig. 12, which follows a graphical representation of the coefficients from left to rear, one can notice a large correction along the border between left and rear. This causes the front left output to go to zero when the control goes from left to left rear. The output remains zero as the control continues fully backward. Along the lr = 0 axis and in the right rear quadrant, this function is identical to the Dolby matrix.

In Fig. 13, one can notice a large peak at 25 boundary between left and rear. This is due to how the LFL matrix element keeps the front output at zero along this boundary as the control passes from the left rear to the fully rear. In the rear direction along the lr = 0 axis and in the rear right quadrant, this element is identical to the Dolby matrix.

· · V

9

9 9 9

9 · 99 »» «· in • 9999 * · 9 · 9 9 9 9 • 999 9 99 9

One of the main design objectives of the Logic 7 matrix design is that the volume at any given output of the uncontrolled signal presented to the decoder inputs is constant regardless of the direction of the controlled signal that is present at the same time. As explained above, this requirement means that the sum of the squares of the matrix elements for each output should be one regardless of the direction of control. As explained earlier, this requirement must be changed when there is strong control in the direction of the respective output. That is, if we observe the left front output, the sum of the squares of the matrix elements must be increased by 3 dB when the control goes all the way to the left. The above elements also change this requirement somewhat when the steering is moved back and forth along the axis Ir = 0.

However, the success of the present construction can still be tested by plotting the square root of the sum of the squares of the matrix elements. See Fig. 14 and Fig. 15 for graphical representations for the revised structure.

In Fig. 14, it is possible to notice a 3 dB peak in the left direction and a somewhat smaller peak as the signal passes from uncontrolled to 22.5 degrees in the central direction. (For this graph, correction 1 / (sin (cs) + cos (cs)) in the back quadrant has been cleared to show exactly how the resulting sum goes to one.) This peak is due to intentionally amplifying the left and right ascents in half front steering. It should be noted that in the other quadrants, the rms sum is very close to one, as intended by this construction, the value in the rear left quadrant is not exactly one, since the procedure used to create the elements is approximative, but the match is still fairly good.

Fefe fefe fefe fefe fe feef feef feef feef feef feefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefefef

In Fig. 15, the uncontrolled (middle) to the right axis is one, the central peak is 0.71, the rear peak is 0.5, and the left peak is 1.41. It is worth noting the vertex along the axis between the middle and the central.

12, The rear matrix elements during the front steering of the rear matrix elements in the '89 patent (except for scale-up by means of the constant 0.71 to show the effect of the standard calibration procedure) are given by:

For front left quadrant

LRL = 0.71 * (1-G (Ir))

LRR = 0.71 * (-1)

(13a). .. (13b)

For rear left quadrant

LRL = 0.71 * (1-G (Ir) + 0.41 * G (-en)) ... (13c)

LRR = -0.71 * (1 + 0.41 * G (-cs)). .. (13d) (the right half of the plane is identical, but with the confusion of LRL and LRR)

The rear nut elements in Dolby Pro-Logic are (after similar calibration):

For front left quadrant

LRL = I-G (14)

LRR = -1 ... (14p)

For rear left quadrant

LRL = I-G (14)

LRR = -1 (14d) ftft · ftftftftftftftftftftftftftftftftftftftftftftftftftftftftft (right half of the plane is the same, but with confusion between LRL and LRR)

It should be noted that the Dolby elements and the elements of the '89 patent are calibrated to be the same in the rear left quadrant when cs = -45 degrees.

13, A brief turn to the surround level in Dolby Pro-Logic

Dolby elements are similar to the '89 patent elements of the same applicant, but without the cs-dependent rear reinforcement. In fact, this difference is quite important because, after a standard calibration procedure, the elements have completely different values for uncontrolled signals. Generally, this description of matrix elements does not consider a calibration procedure for these decoders. All matrix elements were derived to a relatively arbitrary scale. In most cases, the elements are presented as having a maximum value of 1.41. In fact, for technical reasons, the matrix elements are all scaled as necessary to have a maximum value of less than one. In addition, when the decoder is finally put into use, the gain (gain) of each speaker output is adjusted. To set them, a signal that has been coded from four main directions - left, center (front), right and surround (back) with the same sound power is played back and the gain of each output is adjusted until the sound power is the same at the listening position. In practice, this means that the actual level of the matrix elements is scaled so that the four decoder outputs are the same under full control conditions. This calibration was in this • · φ φ φφφφφφφφφφφ · · · ·φφφφφφφφφφφφφφφφφ

The description is explicitly incorporated into the equations for the rear elements mentioned above.

The 3 dB difference in forward or uncontrolled elements is not trivial. During uncontrolled conditions, the elements of the '89 patent are 0.71 and the sum of the squares of these elements is one. This does not apply to Dolby rear elements when calibrated. The LRL has an uncontrolled value of one and the sum of squares is 2 or 3 dB higher than the outputs of the '89 patent. Note that the calibration procedure results in a matrix that does not match the passive Dolby Surround matrix when the matrix is uncontrolled. The Dolby Surround passive matrix specifies that the rear output should be 0.71 * (A _in - B _in ), while the Pro-Logic matrix does not meet this specification. As a result, when inputs A and B are separate, the rear output will be 3 dB stronger than the others. If there are two speakers sharing the rear output, each will be set 3 dB less than one rear speaker, causing all five speakers to have approximately the same audio performance when the decoder inputs are separated. When matrix elements according to the '89 patent are used, the same calibration procedure results in 3 dB less sound power at the rear output when the decoder inputs are separated.

The question of how loud the rear channels should be when the inputs are separate (uncorrelated) ends as a matter of choice according to the listener's taste. When a surround encoded recording is played back, it would be desirable to reproduce the balance as heard by the producer when the recording was mixed. Achieving this balance is a design goal of the combination of a decoder and an encoder according to the present invention. With standard stereo

......... • • * * 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 em em em em em em em Μ Μ em em the original recording while creating a pleasant and unobtrusive surround effect. A problem with Dolby matrix elements is that the power balance in a normal two-channel record is not maintained through this matrix. The surround channels are too thick and the center channel is too weak.

To see the importance of this question, consider what happens when we have an input to a decoder that consists of three components, separate (uncorrelated) left and right components, and separate and separate central components.

A _in - L _in + 0.71 * C _in ... (15a)

B _in = R _in + .71 * C _in ... (15b)

When A _in and B _{in are} played back through a conventional stereo system, the sound output in the room will be proportional to L _in ² + Rin ² + Cln ² . If all three components have roughly equal amplitudes, the power ratio of the central component to the left plus the right component will be 1: 2.

It is an object that the decoder of the present invention reproduces the sound power in a room with approximately the same power ratio as stereo, regardless of the power ratio C _in to L _in and R _in . This can be expressed mathematically. Essentially, the requirement of the same power ratio will specify the functional expression of the central matrix elements along the cs axis if all other matrix elements are considered as given. Suppose Dolby matrix elements calibrated so that the rear sound power is 3 dB less than the other three outputs when the matrix is fully controlled - that is, 3 dB less than the standard calibration, then the central matrix elements should have the shape shown in FIG. 16. The same thing is possible • · 9 9 »999» 9999 99 99 9 t 99 9

9,999,999 for standard calibration and combine the results in Fig. 17.

In Fig. 16, assuming that the power ratios in the decoder outputs are the same as the power ratios for stereo and with a Dolby rear matrix element calibrated at 3 dB below the level normally used, it may be noted that while actual values provide acceptable values the results for the uncontrolled signal and the fully controlled signal are approximately 1.5 dB too low in the middle.

¹⁰

In Figure 17, given the same power ratios for stereo and matrix elements and the calibration actually used in Dolby Pro-Logic (dotted curve), it may be noted that the actual values are more than 3 dB too low for all cs values.

These two pictures show something that mixing technicians are often worried about - especially that a mix ready for playback on Dolby Pro-Logic may need more central volume than a mix ready for playback in Dolby Pro-Logic

2o stereo. Conversely, a stereo-ready mix will lose the clarity of the vocals when played through a Pro-Logic decoder. Ironically, this is not true of a passive Dolby Surround decoder. This issue will again be addressed in the discussion of the central matrix elements.

14. Creating two independent rear outputs

The main problem with both the '89 patent elements and the Dolby elements is that they have only one rear outlet. The '91 patent describes a procedure for generating two independent side outputs, and the mathematical apparatus of this patent has been incorporated in • 9 9 9 9 9 «9 ··· 9 ·

9

9 ··· 9 99 9 »9 9« front left quadrant in ref. [1] of 1996 and Application No. 08 / 742,460. The aim of the elements in this quadrant was to eliminate the output of the left-to-central signal while maintaining some output from the left rear channel for the uncontrolled material that was present at the same time. To achieve this goal, it was assumed that the LRL matrix element would have the following expression:

For left front quadrant

LRL = l-GS (lr) -0.5 * G (en) ... (16a)

LRR = -0.5 * G (en) -G (2y) ... (16b)

As can be seen, these matrix elements are very similar to those of the '89 patent, but with the addition of G (2r) in LRR and GS in LRL. G (2r) was added to add a signal from the decoder input channel B to the left rear output to provide some uncontrolled signal power with the controlled signal removed. It was then solved for the GS function (2r). the criterion is that there is no signal output with a fully controlled signal shifting from left to central. Formula for

GS (2r) was changed to equal G ² (2r), although a much more complex representation of this formula is disclosed in the '91 patent. These two representations can be described as exactly the same.

In reference [1], these elements are corrected by giving them gain (sin (cs) + cos (cs) so that they are closer to constant loudness for uncontrolled material. Although absolutely successful in the right front quadrant, this correction is not too successful in the left front quadrant, see Figure 18. (For the right front quadrant, matrix elements are identical to LRL and

3 ° LRR elements in the '89 patent.) · · · · ···· · ···· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·. «« ··

In Fig. 18, it can be noted that in the left front quadrant there is a 3 dB decrease along a line drawn from the center to the left peak and nearly 3 dB gain at a level along the border between the left and central. The high range in the back quadrant will be discussed later. In this drawing, the tv matrix correction in the decoder VI.11 has been removed to allow a better comparison with the present invention in Fig. 20.

Giant. 18 illustrates several problems associated with audio performance. The first is associated with a decrease in the sum of the squares along the cs = 0 axis. This decrease exists because the functional form G (lr) in LRR is not optimal. The choice of G (lr) was arbitrary - in the previous construction this function already existed in the decoder and its implementation in analog circuits is easy.

Ideally, in this equation, it would have the function GR (lr) and select GS (lr) and GR (lr) in such a way that the sum of squares LRL and LRR is kept constant along the cs = 0 axis, and keeping output zero along the boundary between the left and central.

This can be done. It would also be appropriate to ensure that matrix elements are identical to matrix elements in the right front quadrant along the Ir = 0 axis. Therefore, assume:

LRL = cos (en) -GS (Ir). . . (17a)

LRR = -sin (en) -GR (Ir). .. (17b)

Furthermore, it is desirable that the sum of the squares be one along the cs = 0 axis:

(l-GS (lr)) ² + (GR (lr)} ² = 1 ... (18) and that the output be zero for the controlled signal, or when t varies from zero to 45 degrees:

• 0

0

0

0 ·

0000 0 0

0 «

0 0 0 • 00 0

LRL * cos (t) + LRR * sin (t) = 0

Equations (18) and (19) result in a complicated quadratic equation for GR and GS, which is solved numerically and represented graphically in Figure 19. The use of GS and GR, as shown, results in a large improvement in power sum along the axis cs = 0, as is the intention. However, the sum remains in the sum of squares along the border between the left and the central.

In practical construction, it is probably not very important to compensate for this error, but the present inventor has decided to make this correction heuristically with the following strategy. Both matrix elements were divided by a factor that depends on a new combined variable based on lr and cs. This new variable was called xymin. (In practice, not the division but the multiplication of the inverse of the factor described below.) In Matlab notation:

% finding minimum x or y xymin = x;

if (xymin> y) xymin = y;

end if (xymin> 23) xymin = 23;

end% it can be seen that xymine varies from zero to 22.5 degrees.

· * 0

0 »·

0 · 000

0

0000 0 • 0 • 0

Then, correction of matrix elements along the border was found using xyiain.

In the front left quadrant

LRL = (cos (cs) -GS (lr)) / (1 + 0.29 * sin (Vxymin)) ... (20a)

LRR = (-sin (en) -GR (lr)) / (1 + 0.29 * sin (4 * xymin)) .. (20a)

In the front right quadrant

LRL = cos ... (20c)

LRR = -sin ... (20d)

In reference [2], these elements are additionally multiplied by the tv matrix correction. Giant. 20 in this application shows matrix elements without correcting the tv matrix. In this application, this correction is processed by frequency-dependent circuits that follow the matrix and will be described later in the description.

In FIG. 20, it should be noted that the sum of squares is close to one and continuous, except for a deliberate increase in the back quadrant level.

15. Rear nut elements during rear steering

The back matrix elements described in the '91 patent were not suitable for a five-channel decoder and were modified 25 heuristically in CP-3 of the same applicant. Reference [1] and US Patent Application No. 08 / 742,460 presented a mathematical procedure for deriving these elements along the left rear quadrant boundary. This procedure worked along the boundary, but resulted in discontinuities along the lr = 0 axis and along the cs - 0 axis.

Většinou · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Φ · · · · borders.

For the elements described in this application, these errors 5 have been corrected by interpolation. The first interpolation corrects discontinuities along the cs - 0 boundary for LRL. This interpolation causes the value to coincide with the GS (Ir) value when cs equals zero, and allows the value to increase smoothly to the value given by the previous mathematical apparatus as cs increases negatively backwards. The second interpolation causes the LRR to approach GR (lr) along the cs = 0 axis.

16. Left side / rear output during rear steering from right to right rear

Consider first the left rear left and the left rear right matrix element when the steering is neutral or anywhere between the right and right rear. That is, Ir can vary from 0 to -45 degrees and cs can vary from 0 to -22.5 degrees.

Under these conditions, the controlled input component should be removed from the left outputs - there should be no output from the rear left channel when the steering is to the right or rearward.

The matrix elements described in the '91 patent achieve this objective. These matrix elements are substantially the same as the rear matrix elements in a four-channel decoder with the addition of a correction (sin (cs) + cos (cs)) for uncontrolled volume. When this is done, matrix elements are simple sine and cosine functions:

9 9 99

• * · 9

9 9

LRL = cos (-en) = sri (-en)

LRR = sin (-en) = sric (-en)

... (21b)

... (21b)

Note that a new function sric (x) equal to sin (x) ranging from 0 to 22.5 degrees and sri (x) equal to cos (x) has been defined. These functions will again be used to define the left rear matrix elements during the left control.

17. Left side / rear exit during rear steering from right rear to rear

Now consider the same matrix elements when cs becomes greater than -22.5 degrees. As stated in reference [1] and the two patent applications, the LRL should be increased to one or more in this range, and the LRR should be reduced to zero. Simple functions meet these requirements (keep in mind that cs is negative and varies from -22.5 to -45 degrees in these equations):

LRL = (cos (45 + en) + rboost (-en)) = (sri (-en) + rboos (-en))

LRR = sin (45 + en) = sric (-en)

... (22a)

... (22b)

Rboost (cs) is defined in reference [1] and Application No. 08 / 742,460. It is almost equivalent to the function of 0.41 * G (cs) in the preceding matrix elements except that rboost (cs) is zero for 0> cs> -22.5, and varies from zero to 0.41 as cs varying from -22.5 degrees to -45 degrees. Its precise functional shape is determined by the requirement to keep the rear output volume constant when the sound is moved from the left rear to the fully rear.

• 00 0 0 • 0000 · · 0 • 0 0 0 000 · · «Φ • 000 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 now

The left rear nut elements are fully described during the right steering.

j sou

18. Left Rear Drvkv while driving from left to left rear

The behavior of the left rear left and left rear right elements is much more complicated. The left rear left element must rise rapidly from zero to almost maximum as the lr decreases from 45 to 22.5 or to zero. Matrix elements described in

The reference IQs [1] do this, but, as previously shown, there are continuity problems at the csi = 0 boundary.

For elements from March 1997, a solution was found that uses functions of one variable and several conditions. In reference [1], the problem at the cs = 0 boundary increases because at the front of the boundary (cs> 0) the matrix element LRL is given by GS (2r). On the back (cs <0) the function described in reference (1) has the same endpoints, but is different when lr is not zero or 45 degrees.

The mathematical procedure in reference [1] provides the following

0 equations for left rear matrix elements in the range 22.55 <lr <45. (When rewriting these equations from reference [1], remember that t = 45-lr).

LRL = cos (45-lr) * sin (4 * (45-lr)) -sin (45-lr) * cos (4 * (45-lr)) = sra (lr). . . (23a)

LRR = - (sin (45-lr) * sin (4 * (45-lr)) + cos (45-lr) * cos (4 * (45-lr))) = -srac (lr) ... ( 23b)

Note that two new functions sra (lr) and srac (lr) have been defined in the given range.

• · · · * fefefe · fefefe · fefe fefe fefe fefe «• · fefefe fefefe • fefefe fe fefefe fefe ··

If cs> 22.5, Ir can still vary from 0 to 45. Reference [1] defines LRL and LRR (when lr has a range of 0 <lr <

22.5) (see Fig. 6 in reference [1]) as:

LRL = cos (lr) = sr (lr) ... (23c)

LRR = -sin (Ir) = -srac (1r). . . (23d) these two functions sra (x) and srac (x) are now defined for the range 0 <lr <45.

19. Version of March 1997

The March 1997 version uses interpolation techniques to correct LRR along borders. Here are two discontinuities. Along boundary cs = 0, where the LRR in the back quadrant must match the forward direction LRR, where LRR = -G (Ir) along the boundary cs = 0.

The choice used in March 1997 - though somewhat computationally intensive - is to use interpolation based on cs over a range of 0 to 15 degrees. In other words, when cs is zero, it is used to find the LRR. As cs increases

2o to 15 degrees, interpolation to srac (lr) is used.

There is also the possibility of discontinuity along the lr = axis

0. In March 1997, this discontinuity was corrected (somewhat) by adding a member to the LRR that was found through the use of the new cs_bounded variable. The correction member simply becomes sric (cs_bounded). This member ensures continuity around the axis lr = 0.

Let's first define cs_bounded in the following Matlab notation:

cs_bounded = lr - en:

00 00 0 0 0 000 000 00 00 00 0 0 0000 0 00 0 0 0 if (cs_bounded <l)% this limits the maximum value of cs_bounded = 0;

end if (45-Ilr | <cs_bounded)% use less of two values cs_bounded = 45-lr;

end for en = 0 to 15

LRR = (- (srac (lr) + (srac (lr) -G (lr) *)

15 + sric (csjoounded));

For cs = 15 to 22.5

LRR = (-srac (lr) -sric (csjoounded));

20. LRL implemented in Logic 7 of August 1997

In the present invention, LRL is calculated by interpolation as well as LRR. In Matlab notation:

for en 0 to 15

LRL = ((sr (lr)) (sr (lr-GS (lr)) * (15-en) / 15) + sri (-en)) for cs = 15 to 22.5

LRL = (sra (lr) + sri (-cs));

21. Rear outputs while steering from left rear to full rear

As the control changes from left rear to full rear, the elements follow what was referred to in reference [1], adding back volume corrections. In Matlab notation:

0 0 • ··· ♦ • 0 0 • · «0

0 0 0 0

0000 · 00 0 0 0 0

0 0 0

0 0 0

00 for en> 22.5, lr <22.5

LRL = {sra (ir) -sri (en) + rboost (en)

LRR = -srac (lr) -sric

This fully describes the LRL and LRR matrix elements during the left steering. Values for right steering can be obtained by confusing left and right in the definitions.

CL = 1-G (Ir) + 0.41 * G (en)

CR = l + 0.41 * G

For rear steering:

CL = 1-G

CR = 1

22, the '89 patent central matrix elements and Dolby Pro-Logic have both of the following matrix elements:

For front steering:

... (25a)

(24b). . . (24c). . . (24d)

Since the matrix elements have symmetry about the axis between the left and right, the CL and CR values for the right steering can be detected by confusing CL and CR. See Fig. 21 for a graphical representation of these elements.

In Fig. 21, the center of the graph and the right and rear peaks are 1. The central peak is 1.41. In practice, this element is scaled so that the maximum value is one.

In application No. 0,8 / 742,460 and reference [1], these elements are replaced by sine and cosine functions.

· · ♦ ·· • 43 «

4 • 4 4

4 4 4 •> 4 4

4 4 4 4

4 4 4

44

For front steering:

CL = cos (45-2r) * sin (2 * (45-2r)) - sin (45-2r) * cos (2 * (45-Ir)) + 0,41 * G (en) ... ( 25a)

CR = sin (45-1r) * sin (2 * (45-2r)) + cos (45-2r) * cos (2 * (45-2r)) + 0.41 * G (en). . . (25b) these equations have never been realized. The product of March 1997 used the elements of the '89 patent, but with a different scale and different boost function than G (cs). It was found to be important to reduce the uncontrolled level of the central output and a value of 4.5 dB less than the Pro-Logic level was selected. The amplification function (0.41 * G (cs)) was changed to increase the value of the matrix elements back to the Pro-Logic value as cs increases towards the central one. The amplification function in the March 1997 version was chosen heuristically through listening tests.

In the March 1997 version, the amplification function for cs starts at zero, as before, and increases with cs in such a way that CL and CR increase by 4.5 dB as cs moves from zero to zero.

22.5 degrees. This increase is a constant number in dB for each dB increase cs. The amplification function then changes the slope so that in the next 20 degrees the matrix elements increase by a further 3 dB and then remain constant. Thus, if the control is half-front (8 dB or 23 degrees), the new matrix elements are the same as the neutral values of the old matrix elements. As the control continues to move forward, the new and old matrix elements are aligned. Thus, the central channel output is 4.5 dB smaller than the old output when the control is neutral, but increases to the old value when the control is completely in the central (forward) direction. See Fig. 22 for a three-dimensional graphical representation of these elements.

ftftft ftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftftft

In Fig. 22, it should be noted that the mean value and the right and rear peaks were reduced by 4.5 dB. As cs increases, the central increases to 1.41 at the two slopes of the curve.

Later, it was found that the central elements used in March 5, 1997 were not optimal. Extensive experience with the decoder in practice has shown that the central part of popular music recordings and dialogues in some films may tend to get lost when switching between stereo (dual channel) and reproduction through this matrix. In addition, as the center channel changes in level, a listener that is not equidistant from the front speakers may notice a shift in the apparent position of the center voices. This problem has been extensively analyzed in the development of the new central matrix elements presented in this application. As will be shown later on, there is also a problem when the signal shifts from left to central or from right to central along the border. Matrix elements in application. 08 / 742,460 provide too low output from the center speaker when the audio image is in the middle halfway.

23. Central channel in new construction

Although it is possible to remove a strongly controlled signal from the central channel output using matrix techniques, whenever the control is front but not shifted to the left or right, the central channel must reproduce the sum of inputs A and B with a certain gain member. In other words, it is not possible to remove the separated (uncorrelated) left and right material from the central channel. The only option is to adjust the center speaker volume. How loud should it be?

9 9 9 β 9 9 ···

9999 98 «·« 9 · «9

9 * 9 9999

9999 9 99 9 9

This question depends on the behavior of the left and right main output. The matrix values presented above for LFL and LFR are designed to remove the central component of the input signals when the control moves forward. It can be shown that if the input signal has been coded to come from the front, using a cross-mixer such as a stereo stereo control, all the matrix elements listed above (elements according to the '89 patent, AES elements from the 1996 application, elements of March 1997 and the elements presented earlier in this application) fully retain the original distance.

However, if the input to the decoder consists of separate left and right channels to which a separate central channel has been added, that is:

A _in = L _in + 0.71 * C _in ... (26a)

B _in = R _in + .71 * C _in. ... (26b), then as the level of C _in increases relative to L _in and R _ln C component L and R front outputs of the decoder is not completely eliminated unless C _in is large compared to L _in and R _in. In general, a little of C _{in is} left in the L and R front outputs. What does the listener hear?

There are two calculations of what the listener hears. If the listener is exactly equidistant from the left, right and center speakers, then the listener will hear the sum of the sound pressure from each speaker. This is equivalent to summing the three front outputs. Under these conditions, it is easy to show that any reduction in the center component of the left and right speakers will result in a net loss of sound pressure from the center component regardless of the amplitude of the center speaker. This is because the center speaker is always derived from • • * 0 0 0 0 0 * · 0 0 0 4 • ···· 00 00 * 0 00 0 · 000 0000 • 000 0 0 · 4 ·· 00 sum of inputs A and B and, as its amplitude is increased, the amplitude of the signals L _in and R _{in must} increase together with the amplitude of the signal C _in .

However, if the listener is not equidistant from every 5 speakers, the listener is likely to hear the sum of the sound power from each speaker that is equivalent to the sum of the squares of the three front outputs. In fact, extensive listening tests have shown that it is the sum of the power of all speakers that is really important, so you need to consider the sum of squares of all the decoder outputs, including the rear outputs.

If it is desirable to design a matrix so that the amplitude ratio of L _in , R _in and C _{in is} maintained when switching between stereo reproduction and matrix reproduction, the sound power 1 of the C _in component from the central output must increase in exact proportion to reduce its sound power from the left and right outputs and its reduction in the rear outputs. A further complication is that the left and right outputs have a level gain of up to 3 dB as described above. This causes the central to be somewhat

i. 0 louder to keep the ratios constant. This requirement can be written as a series of equations for audio performance. These equations can be solved for the profit function that is needed for the center speaker.

Graphs showing energy relations for a Dolby Pro-Logic decoder under different conditions have been reported previously. The Pro-Logic decoder is not optimal. The same graphs can now be plotted for the new decoder according to the invention.

Giant. 23 shows the necessary central gain (solid line) if the energy of the central component of the input signal has to be maintained in the front three channels as the control shifts in the direction forward. As can be seen, the necessary increase at the central channel level is quite steep - this increase is many dB amplitude per dB control value. The gain in the standard decoder (dotted curve) is also shown.

As mentioned above, there are two solutions to this problem. The film solution will be described first. The solution here is not entirely mathematical. In practice, it has been found that the function shown in Fig. 23 increases too steeply. The change in central channel levels is too obvious.

It was decided to slightly release the power demands - about 1 dB less central than the ideal shape. When the central values were recalculated, the solid line results in Figure 24 were found. In practice, a linear increase in the initial part of the curve as shown by the dashed line in Figure 24 can be achieved. In practice, the results with these central values were excellent for films .

Referring to Fig. 24, in practice the solid line rises too steeply. The linear slope given by the dashed line works better.

Music requires a different solution. The central attenuation shown in Figures 23 and 24 is derived from the assumption of matrix elements previously reported for LFL and LFR. What if different elements were used? More specifically, is it necessary to vigorously remove the central component from the left and right front outlets?

Listening tests have shown that the previous left and right front matrix elements are unnecessarily emphatic when removing the central component during music playback. From the sound point of view here, there is no need for these to be the following: hlediska * hlediska hlediska hlediska hlediska hlediska hlediska hlediska hlediska hlediska hlediska hlediska hlediska hlediska hlediska hlediska elements have worked that way. The energy removed from the front left and right must be supplied to the center speaker. If this power is not removed, it will come from the left and right speakers, and the center speaker will not have to be as strong. The sound performance in the room is the same. The trick is to supply just enough power to the center speaker to create a convincing front view for off-axis listeners while minimizing the reduction of the stereo width for the listeners equidistant from the front and right speakers.

As was the case in Application No. 08 / 742,460, optimal central loudness can be found by trial and error. Then it is possible to solve the matrix elements that are needed in the front left and right to maintain the performance of the C _in component in the room. As before, suppose that the central channel is reduced by 4.5 dB below the decoder level of the '89 patent of the same applicant, or a total attenuation of -7.5 dB. -7.5 dB equals 0.42. The matrix elements for the central may be multiplied by this factor and a new central gain function GC may be defined.

For front steering

CL = 0.42 * (l-G (lr)) + GC (en)

CR = 0.42 + GC

For rear steering

CL = 0.42 * (l-G (lr))

CR = 0.42

... (27a)

... (27c) (27c!

... (27d)

0 0

0000

0

0 0 0 Φ

0 »

· * 0 0 0 · · · ·

• 0 ··

Several functions have been tested for GC (cs). The function below may not be ideal, but seems appropriate. This function is specified by the angle [alpha] in degrees and was obtained by trial and error.

In Matlab notation:

center_max = 0.65; center_rate = 0.75; center_max2 = 1;

o η n center_rate2 = 0.3;

center_rate3 = 0.1;

if (en <12) gc (en-1) = 0.42 * 10 (db * center_rate / (20));

tmp = gc (en + 1);

else if (cs <30) gc (cs + 1) = tmp * 10 ^and ((E -11) * center_rate3 / (20)); if (gc (en-1)> center_max) gc (en + 1) = center_max;

end end gc (cs + 1) = 10 * cener_max ^A ((cs-29) * center_rate2 / (20));

if (gc (en + 1)> center_max2) gc (donkey) = center_max2;

end end · 9 • * ·· ♦

The function (0.42 + GC (cs)} is shown graphically in Fig.

25. Note the rapid increase from 0.42 (4.5 dB lower than Dolby surround), followed by a slight increase, finally followed by a steep increase to 1.

It is possible to solve the necessary function for the LFR provided the functions for the LFL, LRL and LRR. It is desirable to solve the rate at which component C .. in the left and right outputs should decrease, and then to construct matrix elements that provide this rate of decrease. These matrix elements would also provide some reinforcement of the components L _in and R _in and should have the current common shape at the left and central boundaries as well as at the right and central boundaries.

Suppose that:

LFL = GP (28a)

LFR = GF ... (28b)

CL = 0.42 * (l-G (lr) + GC (en) ... (28c) / Ί)

8d)

The power from the front left and right can then be calculated as follows:

PLR = (GP ² + GF ² ) * (Lin + R) e (GP-GF) ² * Cin ² ... (29a)

Power from the central is:

PC = GC ² * (Lin ² + Rin ² ) + 2 * GC ^: ' _in ² ... (29b)

Power from the rear depends on the matrix elements that will be used. Suppose that the rear channels are attenuated by 3 dB during the front steering, and that LRL is cos (cs) and LRR is sin (cs). From one speaker:

PREAR = (071 * (cos (en) * (L _in + 0.71 * R _in ) -sin (en) * (R _in + Q, 71 * C _in ))) ² ... (29c)

Assuming that L _in ² «R _in ² then for two speakers:

PREAR = 0.5 * C _and ² * ((cos (en) -sin (en)) ² ) + L _in ² . . . (29d)

The overall power of all three speakers is PLR + PC +

PREAR:

PT = (GP ² + GF ² + GC ²⁾ * (Lin + Rin ^{2 2)} + ((GP-GF) ² + 2 * GC ² * C _in ² + PREAR 10 ... (30)

The ratio of power C. _n to power L _in and R _in is:

(assuming L. _n ² = R _in ² )

RATIO = (((GP (en) -GF (en)) ² + 2 * (GC (en) ² +

0.5 (cos (en) -sin (en)) ² )) * _in ² / ((2 * (GP (en) ² + GC (en) ² + GF (en) ² ) +1) * L _in ² .. (31a)

RATIO = (C _in ² / i) * {(GP (en) -GF (en)) ² 22 * (GC (en) ² ) +

0.5 * (cos (en) -sin (en)) 7 (2 * (GP (en)) ² +

GC (en) ^: + GF (en) ² + 1) ... (31b)

For conventional stereo GC = 0, GP = 1 and GF = 0. The power ratio of central to LR is then:

RATIO- (C _in ² / Lj) / * 05 ... (31)

If this ratio is to be constant regardless of the 25 value (C _ir // L _ln ² ) for the present active matrix, then ((GP (cs) -GF (cs)) ^z + 2 * (GC (cs) ² +0 , 5 (cos (en) -sin (en)) ² ) = ((GP (en) ² + GC (en) ² + GF (en) ² ) + Q, 5) ... (33)

The equation above can be solved numerically. Assuming GC above and GP-LFL as before, the results are shown in Figure 26.

a a a a a a a a a a a a a a a a aaaa a «a a aa <·

In Fig. 26, the solid line is a graph of GF needed for constant power ratios with the new music central attenuation GC. The dashed line is the LFR element of March 1997, sin (cs) * corrl. The dotted line is sin (cs), an LFR element without a correction term corrl. It should be noted that GF is close to zero until cs reaches 30 degrees and then increases sharply. In practice, it has been found best to limit the cs value to approximately 33 degrees. In practice, the LFR derived from these curves has a negative sign.

GF specifies the LFR shape of the matrix element along the axis lr = 0 as cs increases from zero to central. There is a need to combine this behavior with the behavior of the preceding LFR element, which must be maintained along the boundaries between left and central as well as from right to central. The procedure to realize this property, when cs <22.5 degrees, is to define the difference function between GF and sin (cs). This function is then limited in a variety of ways. In Matlab notation:

gf_diff = sin (en) -gf (en);

for cs = 0:45;

if (gf_diff (en)> sin (en)} gf_diff (en) = sin (en);

end if (gf_diff (en) <0) gf_diff (en) 0 ';

end ^ find boundaries / s if (y <24) «0 ·

0000

Φ 0 0 <

0 0 · ·· • 0 0 00 0 0 bcs = y- (x-1);

if (bcs <l) maximum bcs limit = l end else bcs = 47-y- (x-1);

if (bcs <1)%> 46) bcs = 1;% 46;

end end

The LFR element can now be written in language notation

Matlab:

% this solution interpolates to the limit% price is, of course, division! 1!

if (y <23)% this is an easy way to halve the region lfr3d (47-x, 47-y) = - sin_tbl (y) + gf_diff (bcs);

else tmp = ((47-x-y) / (47-y)) * gf_diff (y); lfr3d (47-x, 47-y) = -in_tbl (y) + tmp;

end

Note that the sign for gf_diff is positive in the equation above. Thus, 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. See Fig. 27, Fig. 27 ·· · ·· · ftft

In Fig. 27 it should be noted that the value is zero at the center of the straight (no control) and remains zero as cs increases to ~ 30 degrees along the axis Ir = 0. The value then decreases to match the previous value along the border from left to central and right to central.

24. Panoramic error in central output

A new central function created this way, if written as follows:

CL = 0.42 * (1-G (Ir)) + GC (en)

... (34a)

CR = 0.42 + GC (cs) works relatively well along the Ir = 0 axis, but causes a panoramic error along the border between left and central and between right and central. The values in reference [1] of 1996 (which have never been implemented) provide a smooth cos function (2 * en) along the left border. These values create a smooth transition between left and central. It would be desirable for the new central function to have similar behavior along this boundary.

It is possible to perform a matrix element correction, which will be realized by adding an additional function xymin, in Matlab notation:

center_fix_tbl = 0.8 * (corr1-1);

Then

CL ~ 0, 42-0,42 * G (Ir) + GC (en) + center_fix_table (xymin) ... (35a)

CR = 0.42 + GC (en) + center_fix_table {xymin} ... (35b)

9 9 ·

9 9 9

9 9 9

9 9 9

9 · 99 ···

Ο 9 9 ·· / Ζ 9 ·· · 99

See Fig. 28 for a three-dimensional representation of the CL matrix element. Although not perfect, in practice this correction works well.

In Fig. 28, it is worth noting the correction for the transition along the border between left and central, which is relatively smooth.

In Fig. 29, which is a graphical representation of the left front (dotted curve) and central (solid line) output, it is noted that the central control is left on the graph and the leftmost control is on the right. In music strategy, the cs value is now limited to approximately 33 degrees (approximately 13 on the axis as indicated), where the central is approximately 6 dB stronger than the left.

25. Encoder technical details

The Logic 7 encoder has two main goals. First, it should be able to encode a 5.1 channel tape in such a way that the decoded version can be decoded by a decoder

Logic 71 with minimal subjective change. Second, the coded output should be compatible with stereo systems - that is, it should sound as close as possible to a manual 2-channel mix of the same material. One factor of this stereo compatibility should be that the encoder output, when played back on a standard stereo system, should provide consistently perceived loudness for each audio source in the original five-channel mix. The apparent position of the audio source in stereo should also be as close as possible to the apparent position in the five-channel original.

In discussions with the Institute for Transmission Technology (IRT) in Munich, it turned out that the objectives of this stereo

The compatibility of the stereo signal as described above cannot be achieved by a passive encoder. A five-channel recording in which all channels have the same priority priority must be encoded as described above. This encoding requires that the surround channels be mixed into the encoder output in such a way as to conserve energy. This means that the total energy output of the encoder should be the same, regardless of which output is controlled. This constant energy setting will be needed for most film sources and five-channel music sources, where the instruments have been allocated equally to all five speakers. Although such musical sources are currently not common, the present inventor is of the opinion that they will become quite common in the future. Music recordings in which priority instruments are placed in the front three channels, with primary reverberations in the rear channels, require different encoding.

After a series of tests (in IRT and elsewhere), it was found that music recordings of this type were successfully encoded in stereo compatible form when the surround channels were mixed with 3 dB less power than the other channels. This -3 dB level has been adopted as the standard for surround encoding in Europe, but this standard specifies that other surround levels may be used for special purposes. The new encoder according to the invention comprises active circuits which detect strong signals in the surround channels. When such signals are occasionally present, the encoder uses the entire surround level. If the surround inputs are constantly -6 dB or less compared to the front channels, the surround gain is gradually reduced to 3 dB to comply with the European standard.

• ···· 4 · 4

4 4 ·· · • ♦ 4

4 ·

4 4

4 «

These active circuits were also present in the encoder in Application No. 08 / 742,460. However, tests with a previous encoder at the Institute of Transmission Technology (IRT) in Munich found that the direction of some audio sources was encoded incorrectly. A new architecture has been developed to solve these problems. This new encoder is obviously excellent when operating on a large number of difficult materials. The original encoder was first developed as a passive encoder. The new encoder will also work in passive mode, but is primarily designed to operate as an active encoder. Active circuits fix several small flaws that the design has. However, even without this active correction, the operation of this encoder is better than that of the previous encoder.

In the extensive listening tests, several other minor problems were discovered at the first encoder. Many (but not all) of these problems have been solved in the new encoder of the invention. For example, when stereo signals are applied to both the front and rear terminals of the encoder simultaneously, the resulting encoder output is set too forward.

The new encoder compensates for this effect by slightly increasing the rear preload. Similarly, it has been found that when encoding a movie with considerable surround content, dialogue may sometimes be lost. This problem has been greatly improved through changes in power balance as described 25 above, but the encoder is also intended for use with a standard (Dolby) decoder. The new encoder under these conditions compensates for this effect by slightly increasing the central channel input to the decoder.

• · · • • »« «« «« «0 0 0 0

4 «

26, Construction explanation

The new encoder processes the left, central and right signals in accordance with the previous design and the Dolby encoder, ensuring that the central attenuation function of the bitches is equal to 0.71 or -3 dB.

Surround channels look more complicated than they really are. The fc {) and fs () functions direct the surround channels either to a 90-degree phase offset path relative to the front channels or to a zero-phase offset path. In basic encoder operation, fc is one and fs is zero - that is, only the 90 degree phase offset path is active.

The crx value is usually 0.38. This function controls the amount of negative lateral offset for each surround channel. As with the previous encoder, when only one of the surround channels is input, the A and B outputs have an amplitude ratio of -0.38 / 91, resulting in a 22.5 degree backward control angle. As usual, the total power in the two output channels is unitary - that is, the sum of squares 0.91 and 0.38 is one.

While the output of this encoder is relatively simple when only one channel is controlled, it becomes problematic when both surround inputs are controlled at the same time. If the LS and RS inputs are controlled with the same signal (as is common with a film), all signals in the summing nodes are in phase so that the total level in each output channel is 0.38 + 0.91 or 1.29. This output is too strong by a factor of 1.29 or 2.2 dB. Active fefefe feefefe feefefe feefefeefefefe feefefeefefe feefefeefefe feefefefefe fe to reduce fc by up to 2.2 dB when two surround channels they have a similar level and phase.

Another error occurs when two surround channels have a similar level and are out of phase. In this case, the two 5 attenuation factors are subtracted so that outputs A and B have the same amplitude and phase, and a level of 0.91-0.38, or 0.53.

This signal will be decoded as a central direction signal. This error is serious. The foregoing encoder construction generated an uncontrolled signal under these conditions, which is acceptable. However, it is not acceptable that the signals input to the rear input terminals result in a centrally oriented signal. Thus, active circuits are provided that increase the value of fs when the two rear channels have a similar level while being out of phase. Mixing both the real path and the phase shift path for the rear channels results in a 90 degree phase difference between the output channels A and B. This results in an uncontrolled signal, which is also the intention.

As mentioned above, a European standard surround encoder was defined during the discussions at IRT in Munich. This encoder simply mutes two 3 dB surround channels and adds them to the front channels. Thus, the left rear channel is muted and added to the left front channel. This encoder has many drawbacks when encoding multichannel movie sound, or for recordings that have specific instruments in surround channels. Both the volume and direction of these instruments will be incorrectly coded. But this encoder works quite well with classical music, where the two surround channels are the primary reverberations. The 3 dB attenuation has been carefully selected through listening tests to create stereo compatible coding. It was therefore desirable that the encoder according to the present H

** The invention included this attenuation of 3 dB when coded classical music, and to be able to detect this state through the relative levels of the front channels and surround channels in the encoder.

The main function of fc in surround channels is to reduce the level of surround channels in the output mix by 3 dB when the surround channels are much softer than the front channels. Circuits are provided for comparing the front and rear levels, where, when the rear level is less by 3 dB, fc is reduced to a maximum of 3 dB. Maximum attenuation is achieved when the rear channels are 8 dB less powerful than the front channels. This active circuit proves to work well. This circuit makes the new encoder according to the invention compatible with the encoder according to the European standard for classical music.

The operation of the active circuits causes the instruments to be strong in the rear channels to be coded with the whole level.

For surround channels, there is another function of the real mixing path coefficient fs. When the sound moves from the left front input to the left rear input, the active circuits detect that the two inputs have a similar level and phase. Under these conditions, fc is reduced to zero and fs is increased to one. This change in the real coding coefficients results in a much more accurate decoding of this type of picture. In practice, this feature is probably not essential, but appears to be an elegant improvement.

Yet another active circuit is used which has not yet been described. Level detection circuits follow the phase relationship between the central channel and the front left and right. Some popular music recordings that use five channels mix vocals into all three front channels. When the signal is strong in all • • 99

9 9 ·

9 9 99 99 99 99 99 99

99 ·· 9 ·· three inputs, the encoder output will have excessive vocal performance because the three front channels will add together in phase. When this occurs, the active circuits increase the attenuation in the central channel by 3 dB to maintain the power balance in the encoder output.

To summarize: Active circuits are designed for:

1. Decrease the surround channel level by 2.2 dB when these two q channels are in phase.

2. Increase the real mixing path coefficients for the rear channels to a sufficient degree to create an uncontrolled state when the two rear channels are out of phase.

3. Reduce the surround channel level by up to 3 dB when the surround level is much smaller than the front levels.

4. Increase the level and negative phase of the rear channels when their level is similar to the front channels.

5. use real coefficients of the surround channel mix when the audio source is moving from the front input to q the corresponding rear input.

6. Increase the center channel level in the encoder when the center level and the level of the front and surround inputs are approximately the same.

7. Decrease the central channel level in the encoder when there is a common signal in all three front inputs.

8. Future encoder improvements are likely to include a feature similar to feature 2 above for front channels. In the present encoder, when the two front channels are out of phase, the encoding will cause the decoder to position the audio θ backward. It is therefore desirable to detect this state and produce the resulting output uncontrolled.

• * * * * * *

27, Frequency-dependent circuits in the decoder

Fig. 2 shows a block diagram of the frequency-dependent circuits that follow the matrix in the five-channel version 5 of the decoder. There are three sections: low-pass variable, band-pass variable, and HRTF (Head Related Transfer Function) filter. The HRTF filter changes its characteristic depending on the value of the rear control voltage c / s. The first two filters or pass filters change their characteristics in response to a signal to represent the average direction of the input signals to the decoder during pauses between the heavily controlled signals. This signal is called the background control signal.

28. Background control signal

One of the main objectives of the present decoder is to be able to optimally produce a 5 channel surround signal from a conventional 2 channel stereo signal. It is also highly desirable for the decoder to recreate a five-channel surround recording that has been encoded into two channels by the encoder described as part of this application. The two applications differ in the way the surround channels are perceived. With a normal stereo input, most of the sound must be in front of the listener. Surround speakers should contribute to the pleasant feeling of being surrounded by sound and a pleasant atmosphere, but should not draw attention to themselves. The coded surround recording needs the surround speakers to be stronger and more vivid.

To play both types of input optimally without any user setting, it is necessary to distinguish between 2 channel recording and coded 5 channel *

t ·· «· v * v • · ·

Záznam ♦ · · · f záznam záznam záznam záznam záznam záznam záznam záznam záznam záznam záznam záznam záznam záznam záznam. The background control signal is designed to perform this resolution. The background control signal (BCS) is similar and is derived from the rear control signal cs. BCS represents a negative peak value of cs. That is, if cs has a greater negative value than BSC, then the BCS is set to the same value as cs. When cs has a more positive value than BCS, the BCS value slowly fades away. However, the fading of the BCS signal involves further calculation.

Music of many types consists of a series of strong priority tones (foreground tones) - or in the case of a song of sung words. There is a background between these priority tones. The background may consist of other instruments playing additional tones or may consist of reverberations. The circuit that derives the BCS signal monitors the peak level of the priority tones. When the current level is ~ 7 dB less than the foreground peak level, the cs level is measured. The cs value during these meters between the foreground peaks is used to control the BCS decay. If the material in the spacing between the tones is reverberation, it may tend to have a purely rearward direction in the record that was created by coding a five-channel original. This is because the reverberations on the rear channels of the original will be coded with the rear direction. Echoes in a normal 2 channel recording will not have a purely rearward direction. The cs function for these reverbs will be zero or slightly directed from the front.

The BCS derived in this way tends to reflect the type of record. Whenever a significantly backward-controlled BCS material is present, it will always be strongly negative. However, the BCS may be negative even in the absence of strong reverse control if the reverberations in the record have a clean reverse direction. BCS can be used • * 0 • 0000 · • I

0000 to set filters that optimize the decoder for stereo vs. surround inputs.

29. Frequency-dependent circuits: five-channel version

The first of the filters in Fig. 2 is a single low pass filter with an inclination of 6 dB per octave, with an adjustable cutoff frequency. When the BCS is positive or zero, this filter is set to a value that is user adjustable, but is typically approximately 4 kHz. As the BCS becomes negative, the cut-off frequency increases until the BCS is set more than 22 degrees back, then the filter is no longer active. This low-pass filter makes the rear outputs less noticeable when playing conventional stereo material. This filter was already part of the decoder at least since the VI.11, but in earlier decoders it was controlled via cs and not via BCS.

The second filter is a variable bandpass filter. The low-frequency part (pole) of this filter is fixed at 500 Hz.

The high frequency part (zero) varies depending on the user setting and the BCS. This filter implements the audio panoramic control of the present decoder. In application No. 08 / 742,460, the sound panorama is realized by means of matrix elements using a tv matrix correction. Previous decoders based on this activity limited the overall level of the rear channels when the control was neutral or forward. In the new decoder presented in this application, the matrix elements do not include a tv matrix correction.

· · · · · · · · · · · · · · · · · · · · · ·

In new decoders, when the audio panorama control is set to the rear, the high-pass bandpass filter is set in the same way as the low-pass band, in other words, the band is not attenuated and the filter has a flat response.

When the sound panorama control is set to neutral, the high-frequency (zero) setting changes. When BCS is positive or zero, this part shifts to 710 Hz - resulting in 3 dB higher frequency attenuation. For high frequencies, this result is the same as for previous decoders. Thus, 2de is 3 dB attenuation when the control is neutral or front. However, the low frequencies are not muted. These frequencies come from the side of the room at full level. The result is greater low-frequency color and surround without disturbing the high frequencies at the rear. As the BCS becomes negative, the high-frequency portion (zero) shifts toward the low-frequency portion (pole), so that when the BCS is approximately 22 degrees rearward, the bandpass filter has no attenuation.

When the sound panorama control is set to the front, the operation is similar, but the high frequency (zero) shifts to 1 kHz when the BCS is zero or positive. This provides high frequencies attenuation of 6 dB. Again, as BCS becomes negative, attenuation decreases.

The third filter is controlled via c / s rather than via BCS. This filter is designed to emulate the frequency responses of the human head and auricles when the sound source is approximately 150 degrees from the listener's forehead. This type of frequency response curve is called the Head Related Transfer Function or HRTF. This frequency response function was measured for many angles for many • Φ Φ 0 • Φ ··· · ·

Φ · · »

0 · 0 · different people. Generally, when the audio source is approximately 150 degrees from the listener's face, there is a strong notch in frequency response at approximately 5 kHz. A similar notch exists when the audio source is directly in front of the listener - in this case, the degree is at about 8 kHz. Audio sources from the side of the listener do not create these stages. The human brain uses the presence of a notch at 5 kHz as one way of detecting that the audio source is behind the listener.

The current standard for 5 channel audio reproduction recommends that two rear speakers be placed slightly behind the listener at +/- 110 or 120 degrees from the forehead. This speaker position provides a good sense of surround at low frequencies. The sound from the side of the listener, however, does not create the same level of excitement as the sound that is completely behind the listener. Very often, the film director requires the sound effect to come from behind the listener and not from the side.

It is also often the case that the listening room does not have a size or shape suitable for placing the speakers completely behind the listener and the lateral position is the best that can be achieved.

The HRTF filter in the decoder adds the frequency slots of the rear audio source so that the listener hears sound more from behind than the actual speaker position. The filter is designed to vary with cs. When cs is positive or zero the filter is at maximum. This causes the ambient sounds and reverberations to appear more behind the listener. As cs becomes negative the filter value decreases. When cs is about -15 degrees, the filter is completely removed and the audio source appears to be coming from the side. As cs continues to become negative, the filter is again applied so that the audio source ftftft appears to be behind the listener. When cs is fully at the rear, the filter is slightly modified to match the HRTF function for the full rear sound.

30. Frequency-dependent circuits: seven-channel version

Giant. 3 shows the frequency-dependent circuits of the seven-channel version of the decoder. As shown, these circuits consist of three sections - although in actual implementation the other two sections may be merged into one circuit.

The first two sections are identical to those in the five-channel decoder and also perform the same function. The third section is unique to the seven-channel decoder. In the decoder VI.11 and application No. 08 / 742,460, the side and rear channels had separate matrix elements. The operation of these elements was such that when cs was positive or neutral, the side and rear outlets were identical except for the delay. The two outputs remained the same until cs was more negative than 22 degrees. As the control moved further backward, the side outputs were attenuated by 6 dB and the rear outputs were amplified by 2 dB. This caused the sound to appear to be moved from the listener's sides to the back of the listener.

In the present decoder, the difference between the side output and the rear output is achieved by means of a variable bandpass filter in the side output. The third bandpass filter in Fig. 3 does not have attenuation when cs is forward or zero. When cs becomes more negative than 22 degrees, the high frequency (zero) band pass filter shifts rapidly towards 110 Hz, resulting in a high frequency attenuation of approximately 7 dB. Although this bandpass filter has been described as a separate filter from the bandpass filter, which provides the audio function

4

4 • 4

4 4 4

4445 4 4 ·

4444 · ·· panorama, the operation of the two bandpass filters can be combined into one by means of suitable control circuits.

While preferred embodiments of the invention have been described and illustrated herein, there are many other possible embodiments that will be apparent to those skilled in the art, including other modifications and variations, without departing from the spirit of the invention.

Represented by:

Claims (23)

PATENT CLAIMS A surround sound decoder for redistributing a pair of left and right audio input signals, including directionally encoded and non-directional components, into a plurality of output channels for reproduction through speakers surrounding the listening area, comprising means for determining the directional content of said left and right audio signals including at least a left-right control signal and a central-surround control signal, comprising: a left and right input terminals for receiving said corresponding left and right audio input signals; left and right delay means for producing delayed left and right audio signals from said left and right audio input signals; a plurality of multipliers corresponding to twice the number of the plurality of output channels, the means being organized in pairs, a first element of each pair receiving a delayed left audio signal and a second element receiving a delayed right audio signal, each multiplying its input audio a variable matrix coefficient signal to produce an output signal; wherein said variable matrix coefficient is controlled by one or both of said control signals; and a plurality of addition means, one for each of said plurality of output channels, each addition means receiving the output signals of a pair of said multiplication means and outputting one of said plurality of output channels at its output. 0 · 0 0 · · · · 0 ··· «» · · * * · • 0 0 0 0 • 4 0 · · 0 «signals; wherein the decoder has said matrix value variables designed to limit directionally encoded audio components in outputs that are not directly involved in their reproduction at a designated 5, and amplified the directionally encoded audio components in the outputs that are directly involved in their reproduction in the designated direction, so as to maintain a constant overall power of these signals while maintaining a large separation between the left and right channels of the non-directional signals, 10 with respect to said control signals, and maintaining the volume, defined as the total audio power level of the non-directional signals, is effectively constant, regardless of whether or not the directional coded signals are present, and regardless of their intended direction, if present. A decoder according to claim 1, characterized in that said output channels are five, and are designated as front left, center, right front, left surround and right surround. 20 May The decoder of claim 2, further comprising frequency-dependent variable filter means downstream of said left surround and right surround outputs to vary the frequency response and phase response of the outputs prescribed therein, said change being controlled. 25 a plurality of control signals responsive to the presence of surround or background components detected in said left and right audio input signals. 4. The decoder of claim 2, further comprising frequency-dependent variable filtering means and additional delay means following said left. With the right surround output to create from each of the surround and rear output channel surround outputs to vary the frequency and phase responses of several outputs in the prescribed manner wherein said change is controlled by a plurality of control signals responsive to the presence of surround or background components detected in said left and right audio input signals. Decoder according to claim 3 or 4, characterized in that said control signals are: a central-surround control signal responsive to the ratio of the central signal to the surround in-phase or out-of-phase signal components contained in said left and right audio input signals; and a background control signal responsive to the presence of out-of-phase signal components 15 contained in said left and right audio input signals during periods in which strongly controlled signals are not present. Decoder according to claims 1 to 4, characterized in that at least two different modes of operation are provided, wherein the matrix coefficients are controlled differently by means of said control signals in different modes of operation. A decoder according to claim 6, wherein the film mode of operation is optimized for surround reproduction 2 5 encoded audio signals derived from movie recordings or other video sources, and the music mode of operation is optimized to reproduce music recordings or transmissions. A decoder according to claim 7, characterized in that for decoding the film sources, the matrix values for the left and right front inputs are designed to eliminate the As much as possible the central component of the input signals, and the matrix values for the central output are designed so that the central output attenuation starts at least by 4 dB greater than previous standard decoders, and decreased 5 as the central / surround control signal grows in the positive direction, the intermediate matrix values being determined by the requirement to maintain the power ratio of the central component to a separate component of the input signals the same in the decoder outputs. Decoder according to claim 7, characterized in that for decoding the music sources, the matrix values for the central output are designed such that the central attenuation starts at least 4 dB greater than that of standard decoders, and decreases gradually to the maximum value for a standard 15 decoder that is, the value achieved at the central / surround control signal value of approximately 20 degrees, and then the attenuation maintains a relatively constant value as the control signal value increases, the left and right front matrix values being designed such that the central component of the input signals is not eliminated but has been intentionally set to maintain the power ratio of the central component to a separate component of the input signals in the decoder output, while the action of the central left and right front elements is additionally limited to 25 at the central / su rround a control value that results in an approximately 6 dB level difference between the central output and any left or right front output. A decoder according to claim 1, characterized in that the left and right 30 front matrix elements are designed such that the input signal encoded backwards such that the direction lies between the left and right 9 &apos; 9 * 9 99 * 99 * 9 • * 999 «9 99 9 *» «9 * 9 9 * 9 99 * 9 * 999 9 99 9 99 99 back and right rear, does not output from the front outputs. 11. A decoder according to claim 1, wherein the left and right front matrix elements are constructed such that a level gain of about 3 dB is produced for signals that do not have a pure right / left component but have a central / surround control value. about 22 degrees, this level gain decreasing to zero when the center / surround control value decreases to zero, increases to 45 degrees, or when the left / right control value increases from zero to +/- 45 degrees. The decoder of claim 4 or 5, further comprising an additional circuit that generates a background control signal by detecting the direction of the background sound between the tones. 15 or by syllables in the input material, the background control signals being quickly set to the value of said central surround control signal when the center-surround control signal is negative, and the background control signal is slowly adjusted in the positive direction when the direction is The background sound between the tones and syllables is in the forward direction, and the background control signal maintains a negative value when the surround encoded material is played back and a positive or zero value when the standard two-channel material is played back. 13. A decoder according to claim 12, wherein said background control signal is used to control the relative volume of the front and rear outputs so that the volume of the rear outputs is reduced when the background between the tones is either neutral or positive in its direction. • · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · φ · · · φ · φ · The decoder of claim 12, wherein said background control signal controls a low pass filter at the rear outputs such that the cut-off frequency is set to a user-adjustable value when the background direction signal is positive or zero and increases to a high value. when the background direction signal is negative, causing the surround outputs to be less pronounced when playing normal 2-channel material. 15. A decoder according to claim 12, wherein said background control signal controls the variable bandpass filter so that at a background control signal of positive or zero they are attenuated by a user-set frequency value above 500 Hz at the rear outputs, and at background control signal a negative of this attenuation is reduced to zero, causing the surround outputs to be less pronounced when playing conventional 2-channel material. 16. A decoder according to claim 5, wherein the rear matrix outputs are divided into a side output and a rear output by combining an additional delay in the rear output and a variable low pass filter in the side output, wherein said low pass filter is set to a high frequency when said the center-surround control signal is more positive than -22 degrees, and as the center-surround control value increases in the negative direction to more than -22 degrees, the low-pass frequency is rapidly reduced to a final value of 500 Hz when the center-surround control the signal reaches its minimum value of -45 degrees. 17. A decoder according to claim 4, wherein the left and right surround outputs of the five-channel version of the decoder are additionally 0 0 0 00 * 0 000 • 0 0 0 0 0 0 0 0 00 0 0 0 0 * • 0 1 · 0 0 0 · 0 0 »00 provided with a variable filter that emulates the frequency response of the human head / ears system for sound sources that are more than 150 degrees obliquely from the front of the listener so that at said central-surround control signal zero or 5 positive, this filter has its maximum effect, and as the center-surround control signal changes from zero to -15 degrees, the effect of the filter is reduced to zero, and as the center-surround control signal continues in the negative direction, the filter again acts maximally, and then get alone slightly 10 modifies to match the frequency response of the human head-ear lobe system for the rear audio sources as the central-surround control signal reaches its minimum value of -45 degrees. 18. An encoder with circuits for automatically mixing five channels of audio input signals into two output channels so that the energy ratios in the input signals are maintained in the output signals, the direction of the input signals is maintained in phase and amplitude relationships of the different components in the output signals, that, as the input signal moves from any input channel to any other input channel, the phase / amplitude relations of the output signals are decoded as much as possible to the original direction by the decoder according to claim 1, and also as far as possible to the original direction when decoded by a standard movie decoder, characterized by © comprising: five input terminals for receiving said five input signals, respectively designated as left, center, right, left surround, and right surround; two output terminals, each to secure one of the v v v v • • 0 0 · 0 0 0 0 0 0 0 0 The two output channels for the external device; means for determining amplitude / phase relationships between some of said five audio input signals and for generating control signals from said relationships; and 5 means for mixing either fixed or variable ratios of each of said input channels into one of said two output channels, said variable ratios responsive to said control signals. 19. An encoder according to claim 18, wherein the circuitry comprises means for actively correcting the shuffle values such that the output signal energy corresponds to the input signal energy when the same signal is applied to the rear inputs either in phase or out of phase. ] _5 20. The encoder of claim 18, wherein the circuitry includes means for actively correcting the shuffle values such that when the same signal is applied to the rear out-of-phase inputs, the encoder outputs have a 90 degree phase to phase relationship indicating an uncontrolled state for the decoder. Encoder according to claim 18, characterized in that the circuitry comprises means for actively removing phase shifting networks in the rear channels when the signal is spread between one of the front inputs and the rear input on the same side, for example from the front left input to the rear left input. . 22. An encoder according to claim 18, wherein the circuitry comprises means for actively determining the presence of a common signal at all three inputs, and adjusting the scrambled level of the central channel to maintain the overall power of the common signal at the encoder outputs. 23. The encoder of claim 18, wherein the circuitry comprises means for comparing the rear channel level to the front three channel level, and when the rear level is less than the front level, the active circuitry limits the rear channel shuffle levels by up to three decibels, such that the circuitry. allow the encoder to optimally encode music in which the rear channels mainly contain reverberations.