HK1051621B - Method for deriving at least three audio signals from two input audio signals - Google Patents

Description

Method for deriving at least three audio signals from two input audio signals

Technical Field

The present invention relates to audio signal processing. In particular, the invention relates to "multi-directional" (or "multi-channel") audio decoding, using an "adaptive" (or "active") audio matrix approach, to derive more than 3 audio signal streams from a pair of audio input signal streams (or "signals" or "channels"). The invention is used for recovering multi-channel audio, wherein each channel of signals corresponds to a direction and is combined into a signal with less channels through a coding matrix. Although the invention is described in terms of such a pre-designed matrix coding, it will be appreciated that the invention does not necessarily have to employ any particular matrix coding method, but can also be used to produce satisfactory directional effects from material originally recorded for binaural playback.

Background

Audio matrix encoding and decoding is well known in the art. For example, in so-called "4-2-4" audio matrix encoding and decoding, 4 source signals, typically corresponding to 4 cardinal directions (e.g., left, center, right, and surround, or left front, right front, left back, and right back), are encoded as two signals in an amplitude-phase matrix manner. The two signals are transmitted or stored and then decoded by an amplitude-phase matrix decoder to recover approximately the original 4 source signals. Since matrix decoders suffer from a well-known drawback of crosstalk in the decoded audio signals, the decoded signals are only an approximation. Ideally, the decoded signals should be identical to the source signals with a significant degree of isolation between the signals. However, the crosstalk inherent in various matrix encoders results in only 3dB of isolation between signals corresponding to adjacent directions. Audio matrices whose matrix properties do not change are known in the art as "passive" matrices.

In order to overcome the crosstalk problem in various matrix decoders, it is known in the prior art to improve the isolation between decoded signals by adaptively changing the characteristics of the decoding matrix to more closely approximate the source signals. A well-known example of such an active matrix decoder is the Dolby Pro Logic decoder described in us patent No. 4,799,260, which is hereby incorporated by reference in its entirety. The' 260 patent cites a number of prior art patents, many of which describe various different types of adaptive matrix decoders. Other prior art patents include patents by the present inventor, including U.S. patent nos. 5,625,696, 5,644,640, 5,504,819, 5,428,687, and 5,172,415. Each of these patents is also incorporated herein by reference in its entirety.

While prior art adaptive matrix decoders are expected to reduce crosstalk in the reproduced signal and to reproduce the source signals more closely, the prior art has achieved this objective in different ways, many of which are complex and cumbersome, they have not noticed the expected relationships between the intermediate signals in the decoder, which can be used to simplify the decoder and improve the accuracy of the decoder.

Accordingly, the present invention is directed to methods and apparatus that identify and utilize previously unachieved relationships between intermediate signals in an adaptive matrix decoder. By utilizing these relationships, particularly by using an automatic self-canceling configuration using negative feedback, undesired crosstalk components can be easily canceled.

Disclosure of Invention

According to a first aspect of the present invention, the present invention provides a method for deriving at least 3 audio output signals from two input audio signals, wherein 4 audio signals are derived from the two input audio signals by means of a passive matrix, said passive matrix generating two pairs of audio signals in response to the two audio signals: the 1 st pair of derived audio signals represents directions in the 1 st axis (e.g., "left" and "right" signals), and the 2 nd pair of derived audio signals represents directions in the 2 nd axis (e.g., "center" and "surround" signals), the 1 st and 2 nd axes being substantially orthogonal to each other. Each pair of the derived audio signals is processed to generate a 1 st pair and a2 nd pair (left/right pair and center/surround pair, respectively) of intermediate audio signals such that the magnitude of the relative amplitudes of the audio signals in each pair of intermediate audio signals tends to be equal. Generating a 1 st output signal (e.g. a left output signal L) by at least combining at least one component of each of the 2 nd (mid/surround) pair of intermediate audio signals with the same polarity_out) This signal represents the 1 st direction on the axis of the derived audio signal pair (left/right pair) from which the 1 st pair (left/right pair) of intermediate signals is generated. Generating a2 nd output signal (e.g., a right output signal R) at least by combining at least one component of each of the 2 nd (center/surround) pair of intermediate audio signals with opposite polarity_out) This signal represents the 2 nd direction on the axis of the derived audio signal pair (left/right pair) from which the 1 st pair (left/right pair) of intermediate signals is generated. Generating a3 rd output signal (e.g., a center output signal C) by at least combining at least one component of each of the 1 st pair (left/right pair) of intermediate audio signals with the same or opposite polarity_outOr surround output signal S_out) The signal represents being inThe 1 st direction of the axis of the derived audio signal pair (mid/surround pair) from which the 2 nd pair (mid/surround pair) intermediate audio signal is generated. Alternatively, if the 3 rd output signal is generated by a combination of the same polarities, then the opposite polarity, if the 3 rd output signal is generated by a combination of the opposite polarities, then the 4 th output signal is generated by the same polarities, at least by combining at least one component of each of the 1 st pair of intermediate audio signals (left/right pair) (e.g., if the 3 rd output signal is a center output signal C)_outIt is the surround output signal S_outOr if the 3 rd output signal is S_outIt is C_out) It represents the 2 nd direction on the axis of the derived audio signal pair (mid/surround) from which the 2 nd pair (mid/surround) mid signal is generated.

A hitherto unaddressed relationship between the decoded signals is that undesired crosstalk components in the decoded output signal can be substantially suppressed by striving to equalize the amplitudes of the intermediate audio signals in each pair of intermediate audio signals. The present principles do not require perfect equality in order to achieve substantial crosstalk cancellation. Such processing is readily and optimally achieved by using a negative feedback arrangement that results in automatic cancellation of the undesired crosstalk components.

The invention includes embodiments having equivalent topologies. In each embodiment, as described above, the intermediate signals are derived from a passive matrix acting on a pair of input signals and the aim is to equalize these intermediate signals. In embodiments employing the first topology, one cancellation component of each intermediate signal is combined with the passive matrix signal (from the passive matrix acting on each input signal or other signal) to produce each output signal. In one embodiment employing the second topology, the pairs of intermediate signals are combined to produce output signals.

Other aspects of the invention include deriving additional control signals for generating additional output signals.

A primary object of the present invention is to achieve a high degree of cancellation of measurable and perceptible crosstalk under a variety of input signal conditions, using only circuitry with no special requirements in terms of accuracy, and not requiring an exceptionally complex design in the control path, both of which are present in the prior art.

It is another object of the invention to achieve such high performance with a circuit that is simpler or less expensive than prior art circuits.

Drawings

Fig. 1 is a schematic diagram of the operating principle of a prior art passive decoding matrix useful for understanding the present invention.

Fig. 2 is a schematic diagram of the operating principle of a prior art active matrix decoder useful for understanding the present invention, in which the result of the scaling of the outputs of the passive matrix is added to those of the unchanged outputs of the passive matrix in linear combiners.

Fig. 3 is a schematic diagram of the operation of a feedback-derived control system according to the present invention for the left and right VCA and the sum-difference VCA of fig. 2, and for the VCA in other embodiments of the present invention.

FIG. 4 is a schematic diagram of the principle of operation of an arrangement according to the invention, equivalent to the combination of FIGS. 2 and 3, in which the output combiners are responsive to input signal L_tAnd R_tInstead of receiving them from the passive matrix, the components of the output signal of the passive matrix are generated, from which the cancellation components are derived.

Fig. 5 is a schematic diagram of the working principle of an arrangement according to the invention, equivalent to the combination of fig. 2, 3 and 4. In the configuration of fig. 5, the signals that need to be kept equal are the respective signals applied to the output derived combiner and to the feedback circuit for controlling the VCA; the output of each feedback circuit comprises a respective component of the passive matrix.

Fig. 6 is a schematic diagram of the operating principle of an arrangement according to the invention, equivalent to the combination of fig. 2, 3, 4 and 5, in which the variable gain circuit gains (1-g) provided by one VCA and subtractor are replaced by one VCA whose gain is changed in the opposite direction to that of the VCA in the VCA and subtractor arrangement. In this embodiment, the components of the passive matrix are implicit. While in other embodiments the components of the passive matrix are distinct.

FIG. 7 is an idealized graph showing L_t/R_tLeft and right VCA gains g for feedback derived control systems_lAnd g_r(vertical axis) to the pan angle α (horizontal axis).

FIG. 8 is an idealized graph showing the sum and difference VCA gains g for the sum/difference feedback-derived control system_cAnd g_s(vertical axis) to the pan angle α (horizontal axis).

FIG. 9 is an idealized graph showing left/right and reverse and/or differential control voltages (vertical axis) versus the pan angle α (horizontal axis) in a proportional relationship where the maximum and minimum values of the control signals are +/-15V.

Fig. 10 is an idealized graph showing the relationship between the smaller value of the two curves of fig. 9 (vertical axis) and the pan angle α (horizontal axis).

Fig. 11 is an idealized graph showing the relationship between the smaller value of the curves (vertical axis) and the pan angle α (horizontal axis) obtained before the smaller value is taken out of the two curves of fig. 9 and after the/difference voltage is multiplied by a scale factor of 0.8.

FIG. 12 is an idealized graph showing the left rear and right rear VCA gains g for the left rear/right rear feedback derived control system_lbAnd g_rb(vertical axis) to the pan angle α (horizontal axis).

Fig. 13 is a schematic diagram of the operation of a portion of an active matrix decoder according to the present invention in which 6 outputs are obtained.

Fig. 14 is a schematic diagram of the operation of the derivation of the 6 cancellation signals for the 6 output active matrix decoder shown in fig. 13.

Fig. 15 is a schematic diagram showing an actual circuit implementing various aspects of the present invention.

Best Mode for Carrying Out The Invention

Fig. 1 shows a passive decoding matrix in functional principle. Each of the following equations establishes each of the outputs and inputs L_tAnd R_t(left all and right all) relationship:

L_out＝L_t(equation 1)

R_out＝R_t(equation 2)

C_out＝1/2*(L_t+R_t) (equation 3)

S_out＝1/2*(L_t-R_t) (equation 4)

(the symbol "+" in all equations herein represents multiplication.)

The mid-set output is the sum of the inputs and the surround output is the difference between the inputs. Furthermore, both have a scaling factor; this scale factor is arbitrary and is chosen to be 1/2 for ease of illustration. Other scale factor values are also possible. By scaling L by a scale factor of +1/2_tAnd R_tIs applied to a linear combiner 2 to obtain an output C_out. By scaling L with scale factors of +1/2 and-1/2, respectively_tAnd R_tIs applied to a linear combiner 4 to obtain an output S_out。

Thus, the passive matrix of FIG. 1 produces two pairs of tonesFrequency signal, 1 st pair being L_outAnd R_outIn the 2 nd pair are C_outAnd S_out. In this example, the cardinal directions of the passive matrix are designated "left", "center", "right", and "surround". The adjacent cardinal directions lie on mutually perpendicular axes such that for these directional labels, left is adjacent to center and surround, surround to left and right, and so on. It should be understood that the present invention is applicable to any orthogonal 2: 4 decoding matrix.

According to an invariant relationship (e.g., in FIG. 1, C_outAlways 1/2X (L)_t+R_t) A passive matrix decoder derives n audio signals from m audio signals, where n is greater than m. In contrast, an active matrix decoder derives n audio signals according to a variable relationship. One way to configure an active matrix decoder is to combine signal-dependent signal components with the output signals of a passive matrix. For example, as shown in the functional diagram of fig. 2,4 Voltage Controlled Amplifiers (VCAs) 6, 8, 10 and 12 providing variable ratios of the outputs of the passive matrix are added in linear combiners 14, 16, 18 and 20 to the outputs of the passive matrix that are unchanged (i.e., the two inputs themselves together with the two outputs of combiners 2 and 4). Since the inputs to each VCA are derived from the left, right, center, and surround outputs of the passive matrix, respectively, their gains can be designated as g_l、g_r、g_cAnd g_s(all positive values). The output signals of the VCA constitute cancellation signals and are combined with passively derived outputs having crosstalk from directions from which the cancellation signals are derived in order to improve the directional performance of the matrix decoder by suppressing crosstalk.

Note that in the configuration of fig. 2, the paths of the passive matrix still exist. Each output is a combination of the outputs of the passive matrix plus the outputs of the two VCAs. The VCA outputs are selected and multiplied by a scaling factor to provide the required crosstalk cancellation for the passive matrix outputs, respectively, taking into account crosstalk components present in the outputs representing adjacent cardinal directions. For example, one center signal produces crosstalk in the passively decoded left and right signals, and one surround signal produces crosstalk in the passively decoded left and right signals. Accordingly, the left signal output should be combined with the cancellation signal component derived from the passively decoded mid and surround signals and be similar for the other 4-way. In fig. 2, the signals are scaled, polarized, and combined in a manner that provides the desired crosstalk suppression. By varying the respective VCA gains (e.g., the scaling factors in fig. 2) in the range of 0 to 1, unwanted crosstalk components in the passively decoded outputs can be suppressed.

The configuration of FIG. 2 has the following equation:

L_out＝L_t-g_c***(L_t+R_t)-g_s***(L_t-R_t) (equation 5)

R_out＝R_t-g_c***(L_t+R_t)+g_s***(L_t-R_t) (equation 6)

C_out＝**(L_t+R_t)-g_l***L_t-g_r***R_t(equation 7)

S_out＝**(L_t-R_t)-g_l***L_t+g_r***R_t(equation 8)

If all VCAs have a gain of 0, the configuration will be the same as the passive matrix. The configuration of fig. 2 will only differ from the passive matrix by a proportionality constant for any equal value of gain for all VCAs. For example, if all VCAs have a gain of 0.1, then:

L_out＝L_t-0.05*(L_t+R_t)-0.05*(L_t-R_t)＝0.9*L_t

R_out＝R_t-0.05*(L_t+R_t)+0.05(L_t-R_t)＝0.9*R_t

C_out＝**(L_t+R_t)-0.05*L_t-0.05*R_t＝0.9***(L_t+R_t)

S_out＝**(L_t-R_t)-0.05*L_t+0.05*R_t＝0.9***(L_t-R_t)

the result is a passive matrix multiplied by a scaling factor of 0.9. It is therefore apparent that the exact value of the static VCA gain, which will be described below, is not important.

Considering an example, for the case where only the cardinal directions (left, right, center, and surround) are considered, the inputs are respectively only L_tOnly R is_t，L_t＝R_t(same polarity), and L_t＝-R_t(opposite polarity) and the corresponding useful signal output is L only_outOnly R is_outOf only C_outAnd only S_out. In each case, ideally, one output should provide only one signal and the remaining outputs should provide no signal.

It is evident from inspection that if the VCAs can be controlled such that the gain of one VCA corresponding to the desired cardinal direction is 1 and the gain of the remaining VCAs is much less than 1, then at all outputs except that required, the VCA signals will cancel the undesired outputs. As described above, in the configuration of fig. 2, each VCA output cancels crosstalk components in each adjacent fundamental direction (into which crosstalk enters the passive matrix).

Thus, for example, if two inputs are fed with in-phase signals of equal magnitude, R is such that_t＝L_t1, and if the result is g_c1, and g_l、g_rAnd g_sAre all made of0 or close to 0, then:

L_out＝1-1***(1+1)-0***(1-1)＝0

R_out＝1-1***(1+1)+0***(1-1)＝0

C_out＝**(1+1)-0***1-0***1＝1

S_out＝**(1-1)-0***1+0***1＝0

the only output coming from C expected_outSimilar calculations will show that the same is true for a signal taken from only one of the other 3 cardinal directions.

Equations 5,6, 7 and 8 can be equivalently transformed into:

L_out＝**(L_t+R_t)*(1-g_c)+**(L_t-R_t)*(1-g_s) (equation 9)

C_out＝**L_t*(1-g_l)+**R_t*(1-g_r) (equation 10)

R_out＝**(L_t+R_t)*(1-g_c)-**(L_t-R_t)*(1-g_s) (equation 11)

S_out＝**L_t*(1-g_l)-**R_t*(1-g_r) (equation 12)

In such a configuration, each output is a combination of two signals. L is_outAnd R_outBoth involve the gains of the sum and difference and sum and difference VCAs of the two input signals (the inputs of these VCAs are derived from the mid and surround directions, this pair being orthogonal to the left and right directions). C_outAnd S_outBoth involving actual two-way input signalsThe sign, and the gains of the left and right VCAs (the respective inputs of these VCAs are derived from the left and right directions, this pair being orthogonal to the mid and surround directions).

Consider a non-cardinal direction, where the direction R_tFeeding similar L_tHave the same polarity but are attenuated in amplitude. This case means that a signal is placed somewhere between the left and the neutral cardinal directions and should therefore be provided from L_outAnd C_outWith little or no output from R_outAnd S_outTo output of (c).

To R_outAnd S_outFor example, if the two terms in the equation are equal in magnitude and opposite in polarity, a zero output can be obtained.

To R_outThe relationship for this cancellation is, in particular, that

[1/2*(L_t+R_t)*(1-g_c)]Size of (2)

＝[1/2*(L_t-R_t)*(1-g_s)]Size of (equation 13)

To S_outIn other words, the corresponding relationship is

[1/2*L_t*(1-g_l)]Size of (2)

＝[1/2*R_t*(1-g_r)]Size of (equation 14)

Two similar relationships will also be revealed in view of the rocking (or, simply, positioning) of the signal between any two adjacent cardinal directions. In other words, when the input signal represents a sound that rocks between any two adjacent outputs, such amplitude relationships will ensure that the sound emanates from the outputs corresponding to those two adjacent cardinal directions, while the other two outputs provide no sound. To achieve such a result in essence, the magnitudes of the two terms should be strived to be equal in each of equations 9-12. This is achieved by keeping the relative amplitudes of the two pairs of signals in the active matrix equal:

[(L_t+R_t)*(1-g_c)]size of (2)

＝[(L_t-R_t)*(1-g_s)]Size of (equation 15)

And

[L_t*(1-g_l)]size of (2)

＝[R_t*(1-g_r)]Size of (equation 16)

The required relationships shown in equations 15 and 16 are the same as those shown in equations 13 and 14, except that the scaling factor is omitted. When the outputs are obtained by the combiners 14, 16, 18 and 20 of fig. 2, it should be noted that the signals depend on the polarity of the combination and their scaling factors.

The present invention is based on the discovery of these equal magnitude relationships that have not heretofore been noted and, preferably, the relationships are maintained using automatic feedback control, as will be explained below.

From the above discussion of the cancellation of undesired crosstalk signal components and from the requirements for the respective fundamental directions, it can be inferred that for the scaling factors used in this description, the maximum gain of a VCA should be 1. In static, undefined or "no control" conditions, each VCA should employ a low gain in order to effectively provide a passive matrix. When the gain of one of a pair of VCAs needs to be raised from its static value to 1, the other of the pair may maintain the static gain or may change in the opposite direction. One convenient and practical relationship is to keep the product of the gains of the pair of VCAs constant. With analog VCAs, the gain in decibels is a linear function of the control voltage, which automatically occurs if a control voltage is applied equally (but with actually opposite polarity) to a pair of VCAs. Another alternative is to keep the sum of the gains of the pair of VCAs constant. Of course, the invention can also be implemented digitally or in software form, without using analog components.

Thus, for example, if the static gain is 1/a, the actual relationship between the gains of the two pairs of VCAs may be expressed as their product, i.e., the product

g_l*g_r＝1/a²And an

g_c*g_s＝1/a²。

Typical values for "a" may be 10 to 20.

Fig. 3 shows a schematic diagram of the operation of a feedback-derived control system for the left and right VCAs of fig. 2 (6 and 12 respectively). The system receives an input signal L_tAnd R_tProcesses them to derive an intermediate signal L_t*(1-g_l) And R_t*(1-g_r) The magnitudes of the two intermediate signals are compared and, when the magnitudes are different, an error signal is generated that reduces the magnitude difference of the magnitudes for each VCA. One way to achieve this result is to rectify the intermediate signals to derive their amplitudes and apply the two amplitude signals to a comparator, the output of which controls the gain of the VCA in one direction, e.g. signal L_tWill be such that g is increased_lIncrease and make g_rAnd (4) reducing. The values of the circuit (or their equivalents in a digital or software implementation) are chosen such that when the output of the comparator is 0, the gain of the static amplifier is less than 1 (e.g., 1/a).

In analog mode, one practical way to implement this comparison function is to convert the two amplitudes into the number domain so that the comparator subtracts them, rather than determining their ratio. Many analog VCAs have a gain proportional to the exponent of the control signal and therefore they themselves conveniently derive the inverse logarithm of the control output of the logarithm-based comparator. In contrast, however, if implemented digitally, it may be more convenient to divide the two magnitudes and use the result as a direct multiplier or divisor of the VCA function.

More specifically, as shown in FIG. 3, input L_tIs applied to the "left" VCA 6 and to one input of the linear combiner 22, where it is applied with a scale factor of + 1. The output of the left VCA 6 is applied to a combiner 22 (thus forming a subtractor) at a scale factor of-1, and the output of the combiner 22 is applied to a full-wave rectifier 24. Input R_tIs applied to the "right" VCA 12 and to one input of the linear combiner 26, where it is applied with a scale factor of + 1. The output of the right VCA 12 is applied to a combiner 26 (thus forming a subtractor) at a scale factor of-1, and the output of the combiner 26 is applied to a full-wave rectifier 28. The outputs of rectifiers 24 and 28 are applied to non-inverting and inverting inputs, respectively, of an operational amplifier 30 (acting as a differential amplifier). The output of the amplifier 30 provides a control signal in the nature of an error signal that is applied in-phase to the gain control input of the VCA 6 and in polarity-reversed to the gain control input of the VCA 12. The error signal indicates that the two signals whose amplitudes are to be equalized differ in amplitude. The error signal is used to "steer" the two VCAs in the correct direction to reduce the difference in amplitude of the intermediate signal. The outputs to the combiners 16 and 18 are taken from the outputs of VCA 6 and VCA 12. Thus, only one component of each intermediate signal is applied to each output combiner, i.e., -L_t g_rand-R_t g_l。

By providing sufficient loop gain for steady state signal conditions, the difference in amplitude can be reduced to a negligible value. However, in order to achieve substantial crosstalk cancellation, it is not necessary to reduce the difference in amplitude to 0 or a negligible value. For example, theoretically, it is sufficient to reduce the decibel (dB) difference to one tenth of the loop gain, and in the worst case, it is possible to reduce crosstalk by 30 dB. For dynamic situations, the time constant in the feedback control arrangement should be chosen to equalize the amplitudes in such a way that it is substantially inaudible, at least under most signal conditions. In the various configurations described, the specifics regarding the selection of the time constant are outside the scope of the present invention.

Preferably, the gain of the VCA cannot be greater than 1 by selecting the circuit parameters to provide negative feedback of about 20 dB. For the example of scale factors described herein in connection with the configurations of fig. 2,4, and 5, the gains of the VCA may be from some very small value (e.g., much less than 1/a of 1)²) To 1 but not more than 1. The configuration of fig. 3 will make the signals entering the rectifiers approximately equal due to negative feedback.

Since the exact gain (value) is not important when the gain is small, any other relationship will lead to similar acceptable results, the relationship being: when the gain of one of the pair (VCA) increases towards 1, the gain of the other (VCA) can be forced to a small value.

The feedback-derived control system for the mid and surround VCAs of FIG. 2 (8 and 10, respectively) is substantially the same as the configuration of FIG. 3 described above, but receives something other than L_tAnd R_tBut their sum and difference and their outputs from VCA 6 and VCA 12 (which constitute one component of the respective intermediate signals) are applied to combiners 14 and 20.

Thus, by using a circuit without special requirements in terms of accuracy, while using a simple control path integrated into the signal path, a high degree of crosstalk cancellation can be achieved under a variety of input signal conditions. The feedback derivation control system processes pairs of audio signals from the passive matrix in an effort to equalize the magnitudes of the relative amplitudes of the intermediate audio signals in each pair of intermediate audio signals.

The feedback-derived control system shown in fig. 3 controls the gains of VCA 6 and VCA 12 in opposite directions in an effort to equalize the inputs to rectifiers 24 and 28. The degree to which these two terms tend to equalize depends on the characteristics of the rectifiers, the comparator 30 following them, and the gain/control relationship of the VCAs. The higher the loop gain, the closer to equality, but regardless of the characteristics of these components, an attempt to equalize is always made (assuming, of course, that the polarity of the signals helps to reduce the level difference). In practice, the comparator cannot have infinite gain, but it can be implemented as a subtractor with finite gain.

If the rectifiers are linear, that is, their outputs are proportional to the input amplitude, then the output of the comparator or subtractor is a function of the signal voltage or current difference. If instead the rectifiers are logarithmically responsive to the input amplitude (i.e. to the level in dB), the subtraction performed at the comparator inputs is equivalent to taking the ratio of the input levels. Advantageously, the result depends only on the signal difference in dB, independent of the absolute signal level. Considering that the source signal level in dB more closely reflects human hearing, this means that other aspects of equal loop gain are loudness independent, and thus the degree of equalization sought is absolute loudness independent. Of course, at some very low levels, the logarithmic rectifier will not work exactly and therefore there will be an input threshold below which the operation intended to be equalized will cease. As a result, however, control can be maintained over a range of 70dB or more for high input levels without requiring an unusually high loop gain, but there is still a potential problem with loop stability.

Similarly, VCAs 6 and 12 may have gains (i.e., multipliers or dividers) that are proportional or inversely proportional to their control voltages. This will have the effect of: when the gain is small, a small absolute change in the control voltage will result in a large change in gain in dB. For example, consider a VCA with a maximum gain of 1 as required in a feedback-derived control system architecture, and a control voltage V that varies from, say, 0 to 10V_cThus, the gain may be expressed as a ═ 0.1 × V_c. When V is_cNear its maximum, a 100mV change from say 9900 to 10000mV will provide a gain change of 20 log (10000/9900) or about 0.09 dB. When V is_cVery small, a 100mV change from say 100 to 200mV will provide a gain change of 20 log (200/100) or 6 dB. As a result, the effective loop gain and corresponding speed of response will vary depending largely on whether the control signal is large or small. Furthermore, there may be problems with loop stability.

This problem can be solved by using VCAs whose gain, expressed in dB, is proportional to the control voltage, or in other words, by using VCAs whose voltage or current gain depends on the exponential or inverse logarithm of the control voltage. A change in the control voltage, such as 100mV, will give the same dB change in gain as long as the control voltage is within its range. Such devices are readily available as analog integrated circuits, and in digital implementations such characteristics or similar characteristics are readily available.

Thus, the preferred embodiment uses a logarithmic rectifier and exponentially controlled variable gain amplifier to provide closer uniformity in the strive for equality (considered in dB) over a wider range of input levels and ratios of the two input signals.

Since the perception of direction cannot be kept constant in human hearing as the frequency changes, it is desirable to frequency weight the signals entering the rectifiers to emphasize those frequencies that contribute most to human direction and to deemphasize those frequencies that may cause improper steering. Thus, in practical embodiments, an empirically derived filter is installed in front of the rectifiers 24 and 28 of fig. 3, providing a response that attenuates low and very high frequencies and provides a gently rising response in the middle of the audible range. Note that these filters do not change the frequency response of the output signals, they only change the control signals and VCA gains in the feedback-derived control system.

Fig. 4 represents functionally and in a schematic way a configuration equivalent to the combination of fig. 2 and 3. It differs from the combination of fig. 2 and 3 in that each output combiner is responsive to an input signal L_tAnd R_tInstead of receiving the components of the output signal from the passive matrix, the cancellation components are derived from the passive matrix. Such a configuration provides the same result as the combination of fig. 2 and 3, assuming that the summation coefficients are substantially the same in the passive matrices. Fig. 4 incorporates the feedback arrangements described in connection with fig. 3.

More specifically, in FIG. 4, input L_tAnd R_tIs first applied to the passive matrix containing combiners 2 and 4, as in the passive matrix configuration of fig. 1. Input L_tWhich is also the "left" output of the passive matrix, is applied to the "left" VCA 32 and to one input of the linear combiner 34 with a scale factor of + 1. The output of the left VCA 32 is applied to the combiner 34 with a scale factor of-1 (thus forming a subtractor). Input R_tWhich is also the "right" output of the passive matrix, is applied to the "right" VCA 44 and to one input of the linear combiner 46 at a scale factor of + 1. The output of the right VCA 44 is applied to a combiner 46 with a scale factor of-1 (thus forming a subtractor). The outputs of the combiners 34 and 46, respectively, are the signal L_t*(1-g_l) And R_t*(1-g_r) And it is desirable to keep the amplitudes of these signals equal or strive to equalize them. To achieve such a result, it is preferred that those signals be applied to a feedback circuit as shown in and described in connection with fig. 3. The feedback circuit then controls the gain of each VCA 32 and 44.

Further still referring to fig. 4, the "center" output from the passive matrix of combiner 2 is applied to a "center" VCA36 and to one input of a linear combiner 38 at a scale factor of + 1. The output of the central VCA36 is scaled by a scale factor of-1 (and thusForming a subtractor) is applied to combiner 38. The "surround" output from the passive matrix of combiner 4 is applied to a "surround" VCA 40 and to one input of a linear combiner 42 with a scale factor of + 1. The output of the surround VCA 40 is applied to a combiner 42 with a scale factor of-1 (thus forming a subtractor). The outputs of the combiners 38 and 42 are respectively the signal 1/2 (L)_t+R_t)*(1-g_c) And 1/2 (L)_t-R_t)*(1-g_s) And it is desirable to keep the amplitudes of these signals equal or strive to equalize them. To achieve such a result, it is preferred that those signals be applied to a feedback circuit as shown in and described in connection with fig. 3. The feedback circuit then controls the gain of each VCA 38 and 42.

Each output signal L_out、C_out、S_outAnd R_outProduced by respective combiners 48, 50, 52 and 54. Each combiner receives two VCA outputs (these VCA outputs constitute one component of the respective intermediate signals, in an effort to keep their magnitudes equal) to provide respective cancellation signal components, and one or both of the input signals to provide the respective passive matrix signal components. More specifically, the input signal L_tIs applied to L with a scale factor of +1_outCombiner 48, applied to C with a scale factor of +1/2_outCombiner 50 and is applied to S with a scale factor of +1/2_outA combiner 52. Input signal R_tIs applied to R with a scale factor of +1_outA combiner 54 applied to C with a scale factor of +1/2_outCombiner 50 and is applied to S with a scale factor of-1/2_outA combiner 52. The output of the left VCA 32 is applied to C with a scale factor of-1/2_outThe combiner 50, again with a scale factor of-1/2, is applied to S_outA combiner 52. The output of the right VCA 44 is applied to C with a scale factor of-1/2_outCombiner 50 and is applied to S with a scale factor of +1/2_outA combiner 52. The output of the center VCA36 is applied to L at a scale factor of-1_outA combiner 48 and is applied to R with a scale factor of-1_outA combiner 54. The output of the surround VCA 40 is applied to L at a scale factor of-1_outA combiner 48, and is applied to R with a scale factor of +1_outA combiner 54.

It should be noted that in different figures, such as fig. 2 and 4, it may be initially found that there is no inverse correlation between the cancellation signals and the passive matrix signals (e.g., some cancellation signals are applied to the combiners with the same polarity as the passive matrix signals). However, in operation, when a cancellation signal becomes significant, it will have the opposite polarity of the passive matrix signal.

Fig. 5 represents, functionally and in a schematic way, another configuration equivalent to the combination of fig. 2, 3 and 4. In the configuration of fig. 5, the signals to be kept equal are the signals applied to the output derived combiners and the feedback circuits to control the VCAs. These signals include components of the passive matrix output signals. In contrast, in the configuration of fig. 4, the signals from the feedback circuits that are applied to the output combiners are the output signals of the VCA and the passive matrix components are excluded. Thus, in fig. 4 (and in the combination of fig. 2 and 3) the passive matrix components should be explicitly combined with the output of the feedback circuits, whereas in fig. 5 the output of the feedback circuits comprises the passive matrix components and is sufficient by themselves. It is also noted that in the configuration of fig. 5, it is the intermediate signal outputs that are applied to the output combiners, rather than the VCAs (each VCA output constitutes only one component of the intermediate signal). Nevertheless, the configurations of fig. 4 and 5 (along with the configurations of fig. 2 and 3) are equivalent, and if the summing coefficients are accurate, the outputs from fig. 5 are the same as those from fig. 4 (and the combination of fig. 2 and 3).

In fig. 5, the 4 intermediate signals in equations 9, 10, 11 and 12 are obtained by processing the passive matrix outputs and then adding or subtracting to derive the required outputs: [1/2 (L)_t+R_t)*(1-g_c)]，[1/2*(L_t-R_t)*(1-g_s)]，[1/2*L_t*(1-g_l)]And [ 1/2R_t*(1-g_r)]. These signals are also fed to the rectifiers and comparators of two feedback circuits which seek to keep the magnitude of the signal pairs equal, as explained above in connection with figure 3. As applied to the (signals of the) configuration of fig. 5, the outputs of the feedback circuit of fig. 3 to the output combiners are taken from the outputs of combiners 22 and 26, rather than from VCAs 6 and 12.

Still referring to fig. 5, the connections between combiners 2 and 4, VCAs 32, 36, 40 and 44, and combiners 34, 38, 42 and 46 are the same as in the configuration of fig. 4. Also, in both the configurations of fig. 4 and 5, the output of each combiner 34, 38, 42, and 46 is preferably applied to two feedback control circuits (the output of the combiners 34 and 46 is directed to the 1 st feedback control circuit to generate the control signals for each VCA 32 and 44, and the output of the combiners 38 and 42 is directed to the 2 nd feedback control circuit to generate the control signals for the VCAs 36 and 40). In FIG. 5, the output of the combiner 34, signal L_t*(1-g_l) Is applied to C with a scale factor of +1_outA combiner 58 and is applied to S with a scale factor of +1_outA combiner 60. The output of the combiner 46, i.e. the signal R_t*(1-g_r) Is applied to C with a scale factor of +1_outA combiner 58 and is applied to S with a scale factor of-1_outA combiner 60. The output of the combiner 38, signal 1/2 (R)_t+L_t)*(1-g_c) Is applied to L with a scale factor of +1_outA combiner 56, and is applied to R with a scale factor of +1_outA combiner 62. The output of the combiner 42, signal 1/2 (L)_t-R_t)*(1-g_s) Is applied to L with a scale factor of +1_outA combiner 56 and is applied to R with a scale factor of-1_outA combiner 62.

Unlike prior art adaptive matrix decoders, in which the control signals are generated from inputs, the present invention preferably employs a closed loop control in which the magnitude of the signals providing the outputs are measured and fed back to provide the adaptation. In particular, unlike prior art open loop systems, the desired cancellation of unwanted signals for non-cardinal directions does not rely on an exact match of the signal to the characteristics of the control path, and the closed loop configuration greatly reduces the requirement for accuracy in the circuit.

Ideally, in addition to the disadvantages of the actual circuit, the "keep amplitude equal" configuration of the present invention is "complete" in the sense that it is supplied to the input L with a known relative amplitude and polarity_tAnd R_tWill produce signals from the desired output and negligible signals at the other output. "known relative amplitude and polarity" refers to the input L_tAnd R_tRepresenting one cardinal direction or a position between adjacent cardinal directions.

Considering again equations 9, 10, 11 and 12, it can be seen that the total gain of each variable gain circuit containing one VCA is a subtractive configuration in the form of (1-g). Each VCA gain can be changed from a small value to 1, but not more than 1. Accordingly, the gain (1-g) of the variable gain circuit may be changed from very close to 1 to 0. Thus, fig. 5 can be redrawn as fig. 6, where each VCA and associated subtractor have been replaced by a single VCA whose gain changes in the opposite direction to the VCA shown in fig. 5. Thus, the gain (1-g) of each variable gain circuit (which is implemented by, for example, subtracting its output from a passive matrix output with a VCA having a gain "g" as in fig. 2/3, 4 and 5) is replaced by the gain "h" of a corresponding variable gain circuit (which is implemented by, for example, replacing it with a separate VCA having a gain "h" for the passive matrix output). If the characteristics of gain "(1-g)" are the same as gain "h" and if the feedback circuit strives to keep the amplitudes of the required pairs of signals equal, the arrangement of fig. 6 will be equivalent to the arrangement of fig. 5 and will provide the same output. In fact, all of the described configurations, i.e., the configurations of fig. 2/3, 4, 5, and 6, are mutually equivalent.

Although the arrangement of fig. 6 is equivalent and functions exactly the same as all previous arrangements, it is noted that the passive matrix does not appear explicitly, but implicitly. Under static or uncontrolled conditions of the previous configuration, the gain g of each VCA is reduced to a very small value. In the configuration of fig. 6, when all VCA gains h are raised to their maximum values (1 or close to 1), a corresponding uncontrolled state occurs.

Referring more specifically to FIG. 6, the "left" output of the passive matrix, which is also identical to the input signal L_tIs applied with a gain h_lTo generate the intermediate signal L, to generate the "left" VCA 64_t*h_l. The "right" output of the passive matrix, which is also identical to the input signal R_tIs applied with a gain h_rTo generate the intermediate signal R, to generate the "right" VCA 70_t*h_r. The "mid" output from the passive matrix of combiner 2 is applied with a gain h_cTo generate an intermediate signal 1/2 (L)_t+R_t)*h_c. The "surround" output from the passive matrix of the combiner 4 is applied with a gain h_sTo generate intermediate signal 1/2 (L)_t-R_t)*h_s. As described above, the VCA gains h are changed in the opposite direction to the VCA gains g so that the h-gain characteristic is the same as the (1-g) gain characteristic.

Generation of control voltage

Analysis of the control signals in connection with the presently described embodiments is useful for a better understanding of the present invention, and in explaining how the principles of the present invention may be used to derive more than 5 audio signal streams from a pair of audio input signal streams, each associated with a direction.

In the following analysis, an audio source is considered to illustrate the conclusion, thisThe source rocks clockwise around the listener in a circle, starting at the rear, going to the left, middle front, right, and then back. The variable α is the angular (in degrees) magnitude of the image relative to a listener, 0 ° being at the back, and 180 ° being at the middle front. Each input amplitude L is established by the following expression_tAnd R_tConnection to α:

(equation 17A)

(equation 17B)

A one-to-one correspondence relationship exists between the parameter alpha and the amplitude ratio and polarity of each input signal; the use of alpha results in a more convenient analysis. When alpha is 90 DEG, L_tIs a finite amount and R_tIs 0, i.e., only the left way. When alpha is 180 DEG, L_tIs equal to R_tAnd have the same polarity (middle front). When alpha is 0 DEG, L_tIs equal to R_tBut with opposite polarity (middle and later). When L is used, as will be explained further below, L is_tHeel R_tA particular value of interest will occur when they differ by 5dB and have opposite polarity; at this time, alpha values of 31 deg. will be produced on both sides of 0. In practice, the front left and front right speakers are usually placed more forward than ± 90 ° with respect to the center (for example, ± 30 ° to 45 °), so α does not actually represent an angle with respect to the listener, but an arbitrary parameter for explaining panning. In the figure to be described, the center of the horizontal axisThe (α ═ 180 °) represents the middle front, and the left and right ends (α ═ 0 ° and 360 °) represent the rear.

As discussed above in connection with the description of fig. 3, a convenient and practical relationship between the gains of a pair of VCAs in a feedback-derived control system keeps their product constant. As in the embodiment of fig. 3, when the same control signal is simultaneously fed to a pair of VCAs controlled exponentially, a situation occurs automatically in which the gain of one rises and the gain of the other falls.

By L_tAnd R_tTo represent the input signal, to gain the VCA by g_lAnd g_rIs set equal to 1/a²And assuming that the loop gain is large enough and the two output amplitudes have been equalized, the feedback-derived control system of fig. 3 adjusts the gains of the VCA such that they satisfy the following equations:

l Lt | (1-gl) ═ Rt | (1-gr) (equation 18)

In addition to this, the present invention is,

(equation 19)

It is clear that in equation 1, L_tAnd R_tIs irrelevant. The result depends only on their ratio L_t/R_t(ii) a This ratio is referred to as X. Will be g in equation 2_rSubstituting the 1 st equation to obtain the equation for g_lThe solution of (the other root of the quadratic does not represent a real system):

(equation 20)

Drawing g_lAnd g_rThe graph with respect to the pan angle α can be obtained as shown in fig. 7. As expected, when the input represents only the left (α ═ 90 °), g_lFrom a small value at the rear, it rises to a maximum value of 1 and then falls to a small value at the front (α ═ 180 °). In the right half, g_lA small value is maintained. Similarly and symmetrically, g except in the center of the right half of the roll_rAre all very small, g when α is 270 ° (right only)_rUp to 1.

The above results are for L_t/R_tThe feedback is derived from the control system. The sum/difference feedback derivation control system functions in exactly the same way, producing the sum gain g_cAnd a difference gain g_sAs shown in fig. 8. Again, as expected, the sum gain rises to 1 in the middle and drops to a small value everywhere else, and the difference gain rises to 1 in the back.

If the gains of the VCA of the feedback-derived control system depend on the exponent of the control voltage, as in the preferred embodiment, the control voltage depends on the logarithm of the gain. Thus, from the above equation, the equation for L can be derived_t/R_tAnd/or differenceThe control voltage, i.e. the feedback, derives an expression of the output of the comparator of the control system, i.e. the comparator 30 of fig. 3. FIG. 9 shows the left/right and sum/difference control voltages in an embodiment where the maximum and minimum values of the control signal are +/-15V, with the latter term being inverted (i.e., in effect, difference/sum). Obviously, other ratios are possible.

The curves in fig. 9 intersect at two points, one of which is that each signal represents a reflection somewhere to the left and back of the listener, and the other one represents a reflection somewhere in the front half. Due to the symmetry inherent in the curves, these intersections lie exactly between the respective alpha values corresponding to adjacent cardinal directions. In fig. 9, they appear at 45 ° and 225 °.

Although the prior art (e.g. U.S. patent No. 5,644,640 to James w. fosgate, the present inventor) derives the master control signals in different ways and uses the resulting control signals in different ways, the prior art shows that it is possible to derive one more control signal from the two master control signals, which is the larger (larger positive value) or the smaller (smaller positive value) of the two. Fig. 10 shows a signal equal to the smaller of the curves of fig. 9. When α is 45 °, the derived control signal reaches a maximum value, that is, the value at which the original two curves intersect.

It is undesirable that the derived control signal rises to its maximum value only precisely at α ═ 45 °. In a practical embodiment, the derived cardinal direction representing the left back is rather closer to the back, that is, has a value of less than 45 °. By offsetting (adding or subtracting a constant) or ratioing one or both of the left/right and sum/difference control signals so that their curves intersect at the preferred value of alpha, before taking a more or less positive function, the precise location where the maximum occurs can be moved. For example, fig. 11 shows the same operation as fig. 10, except that the sum/difference voltage has been multiplied by a scaling factor of 0.8, with the result that the maximum value now occurs at α -31 °.

In exactly the same way, comparing the inverted left/right control with the inverted sum/difference control and using similar offset or scaling methods, the 2 nd new control signal can be derived, the maximum of which occurs at a position with the desired and predetermined α corresponding to the right rear of the listener (e.g., 360 ° -31 ° or 329 °, 31 ° on the other side of 0 °, symmetrical to the left rear). It is a left/right inversion of fig. 11.

Fig. 12 shows the effect of applying these derived control signals to the VCAs such that the most positive value gives a gain of 1. Just as the gains given by the left and right VCAs in the left and right cardinal directions rise to 1, when a signal is placed at a predetermined position (in this example, at 31 ° on both sides of 0 °), the gains of these derived left and right rear VCAs rise to 1, but still remain small for all other positions.

Similar results can be obtained with various linear controlled VCAs. The curve of the main control voltage versus the pan parameter a will change but will intersect at points selected by scaling or shifting appropriately, so that additional control voltages for a particular map location (rather than the first 4 cardinal directions) can be derived by a one-way small operation. It is clear that it is also possible to invert the control signals and to derive a new result by taking the larger (more positive) instead of the smaller (more negative).

Modifications intended to move their intersection points before taking out larger or smaller ones may optionally also include non-linear operations, instead of, or in addition to, offset or scaling operations. It is evident that this modification allows the generation of a further control voltage, the maximum of which is at L_tAnd R_tThe amplitude ratio (of the input signals) is more or less any desired and the relative polarity.

Adaptive matrix with more than 4 outputs

Fig. 2 and 4 show that a passive matrix may have adaptive cancellation terms added to cancel unwanted crosstalk. In these cases, there are 4 possible cancellation terms derived by 4 VCAs, and each VCA achieves a maximum gain, typically 1, for a signal source in one of the 4 cardinal directions and corresponding to one main output from one of the 4 outputs (left, center, right and rear). The system is complete in the sense that a signal that is panned between two adjacent cardinal directions produces little or no output from directions other than that corresponding to the two adjacent cardinal outputs.

This principle can be generalized to active systems with more than 4 outputs. In such a case, the system is not "complete", but the unwanted signals can still be sufficiently cancelled so that, acoustically, the result is not impaired by crosstalk. See, for example, the 6 output matrices of fig. 13. Fig. 13 is a schematic diagram of the principle of operation of a portion of an active matrix according to the present invention, useful in explaining how to obtain more than 4 outputs. Fig. 14 shows the derivation of the 6 cancellation signals for fig. 13.

Referring first to fig. 13, there are 6 outputs: left front (L)_out) Right front (R)_out) Middle and back (or surround) (S)_out) Right Rear (RB)_out) And left rear (LB)_out). The initial passive matrix is the same as the 4-output system described above for the 3 front and surround outputs (a direct L-output system)_tInputting, will L_t+R_tIs multiplied by a scaling factor of 1/2 and applied to a linear combiner 80 to produce a mid-front output, L_t-R_tIs multiplied by a scaling factor of 1/2 and applied to a linear combiner 82 to produce a medium post output, and a direct R_tInput). With two additional rear outputs, left rear and right rear, by coupling L_tMultiplying by a scaling factor of 1, and_tmultiplied by a scaling factor-b and then applied to a linear combiner 84, and L is applied_tMultiplying by a scaling factor-b, and R_tMultiplied by a scaling factor of 1 and then applied to a linear combiner 86, corresponding to the equation LB_out＝L_t-b*R_tAnd RB_out＝R_t-b*L_tDifferent combinations of inputs. Here, b is a positive coefficient, typically less than 1, e.g., 0.25. It is noted that symmetry is not important to the present invention, but is desirable to achieve symmetry in any practical system.

In fig. 13, in addition to the passive matrix entries, each output linear combiner (88, 90, 92, 94, 96, and 98) receives a plurality of active cancellation entries (on lines 100, 102, 104, 106, 108, 110, 112, 114, 116, 118, 120, and 122) as needed to cancel each output of the passive matrix. These terms include various inputs and/or combinations of inputs multiplied by the gain of the VCAs (not shown), or various combinations of inputs and inputs multiplied by the gain of the VCAs. As described above, the VCAs are controlled such that their gains rise to 1 for basic input conditions and are substantially smaller for other conditions.

The configuration of fig. 13 has 6 cardinal directions, defined by inputs L in defined relative magnitudes and polarities_tAnd R_tIt is provided that each of them should derive the signals only from the appropriate output, while the signals from the other 5 paths are substantially cancelled. For input conditions representing signals that are panned between two cardinal directions, the outputs corresponding to these cardinal directions should provide various signals, while the remaining outputs should provide no or little signals. It is therefore expected that for each output, in addition to the passive matrix, there will be several cancellation terms (in practice, more than the two shown in fig. 13), each of which corresponds to an undesired output for one input (which corresponds to each of the other cardinal directions). In practice, the configuration of FIG. 13 may be modified to eliminate the post-mid output S_out(thus, the combiners 82 and 94 are eliminated) so that the mid-rear is only half way between the left-rear and right-rearOne shake instead of the 6 th cardinal direction.

For the 6-output system of fig. 13 or its 5-output variant, there are 6 possible cancellation signals: 4 derived via two pairs of VCAs, which are parts of a left/right and/or difference feedback derivation control system, and also two derived via left and right rear VCAs controlled according to the method described above (see the embodiment of fig. 14 to be described below). The gains of the 6 VCAs are all according to FIG. 7 (g)_lLeft and g_rRight), FIG. 8 (g)_cAnd g_sPoor), and FIG. 12 (g)_lbLeft rear and g_rbRight rear). The cancellation signals are summed with the passive matrix terms using calculated or selected coefficients to minimize unwanted crosstalk, as described below.

By considering the input signals and VCA gains for each other cardinal direction, the desired cancellation mixing coefficients for each base cardinal direction output can be obtained, bearing in mind that only the VCA gains for the signals in the respective cardinal directions are raised to 1 and the gains drop significantly and rapidly from 1 as the image leaves.

Thus, for example, in the case of left output, one needs to consider signal conditions for mid-front, right-only, right-rear, mid-rear (in the case of 5-way output, this is not a true cardinal direction), and left-rear.

Considering now the left output in detail, i.e. for L in the 5-way output modification of FIG. 13_out. It contains entries L from the passive matrix_t. When the input is centered, to cancel the output, if L_t＝R_tAnd g is_c1, one needs the term-1/2 g_c*(L_t+R_t) This is exactly the same as in the 4-way output system of fig. 2 or fig. 4. When the input is located somewhere in the middle or between the middle and front right (thus including the right back), one needs-1/2 × g_s*(L_t-R_t) This is thenLess precisely as in the 4-way output system of fig. 2 or fig. 4. When the input is at the rear left, one needs a signal from the rear left VCA, the gain g of which is the above-mentioned VCA, in order to cancel the output_lbThe variation of (a) is shown in fig. 12. It does provide a significant cancellation signal only when the input is located in the rear left region. Since the left rear can be considered to be between the left front (denoted as L only)_t) And after (denoted as 1/2 (L)_t-R_t) Somewhere in between), one expects that the rear left VCA should work on the combination of these signals.

Various fixed combinations may be used, but by using already passed left and difference VCAs (i.e., g)_l*L_tAnd 1/2 g_s*(L_t-R_t) The sum of the signals) that changes according to the position of the signals that are panned (but not exactly located) in the left back region, thereby providing better cancellation of those panned and the underlying left back itself (the output of the source). It is noted that in this rear left position, which can be considered as being between left and rear, g_lAnd g_sBoth having a finite value less than 1. Thus desired for L_outThe equation of (a) would be:

L_out＝[L_t]-**g_c*(L_t+R_t)-**g_s*(L_t-R_t)-x*g_lb*((g_l*L_t+g_s***(L_t-R_t) Equation 21

The coefficient x may be derived empirically or from consideration of the exact VCA gain when the sound source is in the region of the left rear cardinal direction. Term [ L ]_t]Is an entry in the passive matrix. Each item 1/2 g_c*(L_t+R_t)，-1/2*g_s*(L_t-R_t) And 1/2 x g_lb*(g_l*L_t+g_s*1/2*(L_t-R_t) Is shown in linear combiner 88 (FIG. 13), which may be followed by L_tAre combined together to derive an output audio signal L_outEach of (1) to (2)Item (see fig. 14). As described above, there may be more than two crosstalk cancellation term inputs, not just the two (100 and 102) shown in FIG. 13.

Similarly, or with the aid of symmetry, R is derived_outThe equation of (2):

R_out＝[R_t]-**g_c*(L_t+R_t)+**g_s*(L_t-R_t)-**x*g_rb*((g_r*R_t-g_s*(L_t-R_t) Equation 22

Term [ R ]_t]Is an entry in the passive matrix. Each item-1/2 × g_c*(L_t+R_t)，1/2*g_s*(L_t-R_t) And-1/2 x g_rb*(g_r*R_t-g_s*(L_t-R_t) Is shown in linear combiner 98 (fig. 13), which may be aligned with R_tAre combined together to derive an output audio signal R_outThe respective cancellation terms of (c) are shown in fig. 14. As described above, there may be more than two crosstalk cancellation term inputs, rather than just the two (120 and 122) shown in FIG. 13.

Middle front output C_outContaining passive matrix entries 1/2 (L)_t+R_t) Plus left and right cancellation terms, -1/2 × g for a 4-output system_l*L_tAnd-1/2 × g_r*R_t：

C_out＝[*(L_t+R_t)]-**g_l*L_t*-**g_r*R_tExpression 23

For left rear, middle rear or right rear, there is no need for significant cancellation terms for them since they effectively pan and have already been cancelled via the rear (wrap around in a 4-way output system), between the left front and right front. Term [1/2 (L)_t+R_t)]Is an entry in the passive matrix. Each item-1/2 × g_l*L_tAnd-1/2 × g_r*R_tThe representation may be applied to the inputs 100 and 102, andand in the linear combiner 90 (fig. 13), may be associated with L_tAnd R_tTo derive an output audio signal C_outThe respective cancellation terms of (c) are shown in fig. 14.

For the left rear output, the starting passive matrix is L as described above_t-b*R_t. For the case of only left input, when g is_lWhen 1, it is apparent that the desired offset is-g_l*L_t. For the case of only right input, when g_rWhen 1, the term of offset is + g_r*R_t. For medium-front input, L_t＝R_tAnd g is_c1, can be represented by (1-b) × g_c*1/2*(L_t+R_t) To cancel out the undesired output L from the passive terms_t-b*R_t. The right posterior term of offset is-g_rb*(g_r*R_t-1/2*g_s*(L_t-R_t) ) with R_outThe same term is used with an optimized coefficient y, which can again be obtained empirically or calculated from the VCA gains under left-hand or right-hand conditions. Therefore, the temperature of the molten metal is controlled,

LB_out＝[Lt-b*R_t]-g_l*L_t+b*gr*R_t-(1-b)*g_c***(Lt+Rt)-y*g_rb*(g_r*R_t-g_s***(L_t-R_t) (equation 24)

In a similar manner to that described above,

RB_out＝[R_t-b*L_t]-g_r*R_t+b*g_l*L_t-(1-b)*g_c***(L_t+R_t)-y*g_lb*(g_l*L_t+g_s***(L_t-R_t) (equation 25)

For equation 24, the term [ L ]_t-b*R_t]Is an entry in a passive matrix, and each entry-g_l*L_t，+b*g_r*R_t，-1/2*(1-b)*g_c*(L_t+R_t) And-y x g_rb*(g_r*R_t-g_s*1/2*(L_t-R_t) Is shown in linear combiner 92 (fig. 13), which may be followed by L_t-b*R_tAre combined together to derive an output audio signal LB_outThe respective cancellation terms of (c) are shown in fig. 14. As described above, there may be more than two cancellation term inputs, not just the two (108 and 110) shown in FIG. 13.

For equation 25, the term [ R ]_t-b*L_t]Is an entry in a passive matrix and each component-g_r*R_t，b*L_t*g_l，-1/2*(1-b)*g_c*(L_t+R_t) And-y x g_lb*(g_l*L_t+g_s*1/2*(L_t-R_t) Is shown in linear combiner 96 (FIG. 13), which may be aligned with R_t-b*L_tAre combined together to derive an output audio signal RB_outThe respective cancellation terms of (c) are shown in fig. 14. As described above, there may be more than two cancellation term inputs, not just the two (116 and 118) shown in FIG. 13.

In practice, all of the coefficients may need to be adjusted to compensate for the finite loop gain of the feedback-derived control system (which cannot provide exactly equal signal levels) and other disadvantages, and other combinations of 6 cancellation signals may be used.

Of course, these principles can be generalized to embodiments having 5 or more than 6 outputs. By further applying scaling, offsetting or non-linear processing of the two main control signals from the left/right and sum/difference feedback sections of the feedback-derived control system, additional control signals may also be derived, allowing additional cancellation signals to be generated via the VCAs whose gains are raised to a maximum value when a is at other desired predetermined values. In the case where the signals are present in each of the other cardinal directions, the synthesis process taking into account each of the outputs will in turn produce the appropriate terms and coefficients to produce the additional outputs.

Referring now to FIG. 14, each input signal L_tAnd R_tIs applied to a passive matrix 130 which produces a signal from an input L_tFrom the input R_tFrom a mid output of the linear combiner 132, at L_tAnd R_tIs an input with a scaling factor of +1/2, and a surrounding output from the linear combiner 134, is represented by L_tAnd R_tIs an input and carries scale factors of +1/2 and-1/2, respectively. The basic directions of the passive matrix are designated as "left", "center", "right", and "surround". Adjacent cardinal directions lie on mutually orthogonal axes such that, for such directional indicia, the left is adjacent mid-way and surrounding; around adjacent left and right, and so on.

The left and right passive matrix signals are applied to the 1 st pair of variable gain circuits 136 and 138 and the associated feedback derived control system 140. The mid and surround passive matrix signals are applied to the 2 nd pair of variable gain circuits 142 and 144 and the associated feedback derived control system 146.

The "left" variable gain circuit 136 includes a Voltage Controlled Amplifier (VCA)148 having a gain g_lAnd a linear combiner 150. In combiner 150, the VCA output is subtracted from the left passive matrix signal so that the total gain of the variable gain circuit is (1-g)_l) And the output of the variable gain circuit constituting an intermediate signal at the output of the combiner is (1-g)_l)*L_t. The output signal of VCA 148, which constitutes a cancellation signal, is g_l*L_t。

The "right" variable gain circuit 138 includes a Voltage Controlled Amplifier (VCA)152 having a gain g_rAnd a linear combiner 154. In combiner 154, the VCA output is subtracted from the right passive matrix signal so that the overall gain of the variable gain circuit is (1-g)_r) And the output of the variable gain circuit constituting an intermediate signal at the output of the combiner is (1-g)_r)*R_t. Form aThe output signal of VCA 152 for one cancellation signal is g_r*R_t. Each intermediate signal (1-g)_r)*R_tAnd (1-g)_l)*L_tConstituting the 1 st pair of intermediate signals. It is desirable that the relative amplitudes of the 1 st pair of intermediate signals tend to be equal. This is accomplished by the associated feedback-derived control system 140 as described below.

The "center" variable gain circuit 142 includes a Voltage Controlled Amplifier (VCA)156 having a gain g_cAnd a linear combiner 158. In combiner 158, the VCA output is subtracted from the intermediate passive matrix signal so that the overall gain of the variable gain circuit is (1-g)_c) And the output of the variable gain circuit constituting an intermediate signal at the combiner output is 1/2 x (1-g)_c)*(L_t+R_t). Output signal 1/2 xg of VCA 156_c*(L_t+R_t) A cancellation signal is formed.

The "surround" variable gain circuit 144 includes a Voltage Controlled Amplifier (VCA)160 having a gain g_sAnd a linear combiner 162. In combiner 162, the VCA output is subtracted from the surround passive matrix signal so that the overall gain of the variable gain circuit is (1-g)_s) And the output of the variable gain circuit constituting an intermediate signal at the combiner output is 1/2 x (1-g)_s)*(L_t-R_t). Output signal 1/2 xg of VCA 160_s*(L_t-R_t) A cancellation signal is formed. Each intermediate signal 1/2 x (1-g)_c)*(L_t+R_t) And 1/2 (1-g)_s)*(L_t-R_t) Constituting the 2 nd pair of intermediate signals. It is desirable that the relative amplitudes of the 2 nd pair of intermediate signals tend to be equal. This is accomplished by the associated feedback-derived control system 146 as described below.

The feedback derivative control system 140 associated with the 1 st pair of intermediate signals includes filters 164 and 166 that receive the outputs of the combiners 150 and 154, respectively. The respective filter outputs are applied to logarithmic rectifiers 168 and 170 which rectify the inputs and produce logarithms of their inputs. The rectified and logarithmized outputs are applied with opposite polarity to a linear combiner 172 whose output, constituting the subtraction of its input (the translator: should be the difference), is applied to a non-inverting amplifier 174 (devices 172 and 174 correspond to the amplitude comparator 30 of fig. 3). Subtracting each logarithmic signal provides a comparison function. As mentioned above, this is a practical way to implement the comparison function in the analog domain. In this case, each VCA 148 and 152 is of a type that inherently takes the inverse logarithm of their control inputs, thereby taking the inverse logarithm of the control outputs of the log-based comparators. The output of the amplifier 174 constitutes the control signal for each of the VCAs 148 and 152.

As mentioned above, if implemented digitally. It would be more convenient to implement the division of the two amplitudes and use the result as a direct multiplier to the VCA function. As previously noted, each of filters 164 and 166 may be empirically derived, having a frequency response that attenuates low and very high frequencies and provides a slightly elevated response in the middle of the audible frequency range. These filters do not change the frequency response of the output signals, they only change the control signals and VCA gains in the feedback derived control system.

The feedback derivative control system 146 associated with the 2 nd pair intermediate signal includes filters 176 and 178 that receive the outputs of the combiners 158 and 162, respectively. The respective filter outputs are applied to logarithmic rectifiers 180 and 182 which rectify the inputs and produce logarithms of their inputs. The rectified and logarithmized outputs are applied with opposite polarity to a linear combiner 184 whose output, constituting the subtraction of its input (the translator: should be the difference), is applied to a non-inverting amplifier 186 (devices 184 and 186 correspond to the amplitude comparator 30 of fig. 3). The feedback derived control system 146 operates in the same manner as the control system 140. The output of the amplifier 186 constitutes the control signal for each of the VCAs 158 and 162.

Additional control signals are derived from feedback-derived control systems 140 and 146. Control signals of the control system 140 areTo the 1 st and 2 nd scaling, offset, phase inversion, etc. blocks 188 and 190. The control signals of the control system 146 are applied to the 1 st and 2 nd scaling, offset, phase inversion, etc. blocks 192 and 194. Each of the functional blocks 188, 190, 192, and 194 may include one or more of the polarity inversions, amplitude offsets, amplitude scaling, and/or non-linear processing described above. Also as described above, by reducing or increasing the functional blocks 196 and 198, the outputs of the functional blocks 188 and 192 and the functional blocks 190 and 194, respectively, will be reduced or increased to generate additional control signals and applied to the left rear VCA200 and the right rear VCA 202, respectively. In this case, additional control signals will be derived in the manner described above to provide control signals adapted to produce a left rear cancellation signal and a right rear cancellation signal. The input to the rear left VCA200 is obtained by additionally combining the left and surround cancellation signals in a linear combiner 204. The input to the rear right VCA 202 is obtained by subtractively combining the right and surround cancellation signals in a linear combiner 204. Alternatively and sub-optimally, the inputs to the VCAs 200 and 202 may be derived from the outputs of the left and surrounding passive matrices and from the outputs of the right and surrounding passive matrices, respectively. The output of the rear left VCA200 is a rear left cancellation signal g_lb*1/2*(g_l*L_t+g_s*(L_t-R_t)). The output of the rear right VCA 202 is a rear right cancellation signal g_rb*1/2*(g_r*R_t+g_s*(L_t-R_t))。

Fig. 15 is a schematic circuit diagram representing a utility circuit embodying aspects of the present invention. The resistance value is expressed in units of Ω. Capacitor values not noted are in units of μ F.

In fig. 15, "TL 074" is a texas instruments dual low noise jfet input (high input impedance) general purpose operational amplifier for high fidelity and audio preamplifiers. Details of the device are readily available in published literature. Web site on internethttp://www.ti.com/sc/docs/products/analog/tl074.htmlThe associated data table can be found.

In FIG. 15, "SSM-2120" is a monolithic integrated circuit for audio applications. It comprises two VCAs and two level detectors allowing logarithmic control of the gain or attenuation of the signals to the level detectors according to their amplitude. Details of the device are readily available in published literature. Web site on internethttp://www.analog.com/pdf/1788 c.pdfThe associated data table can be found.

The following table relates the entries used in this document to the labels at the VCA outputs and the labels of the vertical bus of fig. 15.

Items used in the above description	The markings on the output of the VCA of FIG. 15	Labels on the vertical bus of FIG. 15
			gl*Lt	Left VCA	LVCA
gr*Rt	Right sideVCA	RVCA
			**gc*(Lt+Rt)	Front VCA	FVCA
**gs*(Lt-Rt)	Rear VCA	BVCA
			glb*((gl*Lt+gs***(Lt-Rt)	Rear left VCA	LBVCA
grb*((gr*Rt-gs***(Lt-Rt))	Rear right VCA	RBVCA

In fig. 15, the labels on the wires connecting the resistors of the output matrix are intended to convey the function of the signals, not their source. Thus, for example, the upper table connected to the left front outputThe root lead is represented as follows:

reference numerals in FIG. 15	Means of
		LT	From input LtContribution of (1)
CF cancellation	Signal for canceling unwanted output of center front sound source
		LB offset	Signal for canceling unwanted output of left rear sound source
BK cancellation	Signal for canceling unwanted output of rear sound source
		RB cancellation	Signal for canceling undesired output of right rear sound source
LFGR	Left front gain control-panning along the front, giving more constant loudness

Note that in fig. 15, the matrix itself can provide the inversion of any term (U2C, etc.) regardless of the polarity of the VCA terms. Further, "servo" in fig. 15 refers to the feedback-derived control system described herein.

The invention may be implemented using analog, mixed analog/digital, and/or digital signal processing, in which case the functions are implemented in software and/or hardware. Analog terms such as VCA, rectifiers, etc. are intended to include their digital equivalents. For example, in a digital embodiment, a VCA is implemented by multiplication or division.

Claims (38)

1. Method for deriving at least 3 audio output signals from two input audio signals, comprising:

deriving 4 audio signals from said two input audio signals, wherein said 4 audio signals are derived using a passive matrix, said passive matrix being responsive to the two input audio signals for generating two pairs of audio signals, a 1 st pair of derived audio signals representing 1 st and 2 nd directions in a 1 st axis, a2 nd pair of derived audio signals representing 3 rd and 4 th directions in a2 nd axis, said 1 st and 2 nd axes being substantially orthogonal to each other,

processing said each pair of derived audio signals to produce a 1 st and a2 nd pair of intermediate audio signals respectively, wherein the relative amplitudes of the audio signals in each pair of intermediate audio signals are intended to be equalized by deriving each pair of intermediate audio signals from a respective pair of derived audio signals and applying a relative scale to each pair of derived signals to generate a pair of intermediate signals whose amplitudes are closer to being equalized than the derived signals from which the intermediate signals were derived,

generating a 1 st output signal representing a 1 st direction on an axis of the derived audio signal pair, generating a 1 st pair of intermediate signals from the pair of derived audio signals, said 1 st output signal being generated at least by combining at least one component of each of said 2 nd pair of intermediate audio signals with the same polarity,

generating a2 nd output signal representing a2 nd direction on an axis of the derived audio signal pair, generating a 1 st pair of intermediate signals from the pair of derived audio signals, generating said 2 nd output signal at least by combining at least one component of each of said 2 nd pair of intermediate audio signals with opposite polarities, and

generating a3 rd output signal representing a 1 st direction on an axis of the derived audio signal pair, generating a2 nd pair of intermediate signals from the pair of derived audio signals, said 3 rd output signal being generated at least by combining at least one component of each of said 1 st pair of intermediate audio signals with the same or opposite polarity.

2. The method of claim 1, further comprising:

generating a 4 nd output signal representing a2 nd direction on the axis of the pair of derived audio signals, generating a2 nd pair of intermediate signals from the pair of derived audio signals, if the 3 rd output signal is generated by a combination of like polarities, generating the 4 th output signal at least by combining at least one component of each of the 1 st pair of intermediate audio signals with opposite polarities, if the 3 rd output signal is generated by a combination of opposite polarities, generating the 4 th output signal at least by combining at least one component of each of the 1 st pair of intermediate audio signals with like polarities.

3. A method according to claim 1 or 2, wherein a pair of variable gain circuits apply the relative scaling to a respective pair of derived signals.

4. The method of claim 2, wherein:

generating a 1 st output signal comprises combining a component of each of said 2 nd pair of intermediate audio signals with a passive matrix audio signal representing said 1 st direction, said components constituting a cancellation signal that is in anti-phase with said passive matrix audio signal,

generating a2 nd output signal comprises combining a component of each of said 2 nd pair of intermediate audio signals with a passive matrix audio signal representing said 2 nd direction, said components constituting a cancellation signal that is in anti-phase with said passive matrix audio signal,

generating a3 rd output signal includes combining a component of each of the 1 st pair of intermediate audio signals with a passive matrix audio signal representing the 3 rd direction, the component constituting a cancellation signal that is in anti-phase with the passive matrix audio signal, and,

generating the 4 th output signal includes combining a component of each of the 1 st pair of intermediate audio signals with a passive matrix audio signal representing the 4 th direction, the component forming a cancellation signal that is in anti-phase with the passive matrix audio signal.

5. The method of claim 4, wherein matrix audio signals representing the 1 st, 2 nd, 3 rd and 4 th directions, respectively, are generated by the passive matrix.

6. The method of claim 4, wherein passive matrix audio signals representing the 1 st, 2 nd, 3 rd and 4 th directions, respectively, are generated in a plurality of linear combiners, which also combine the passive matrix audio signals with the signal components.

7. The method of claim 1, wherein:

generating a 1 st output signal comprises combining a component of each of said 2 nd pair of intermediate audio signals with a passive matrix audio signal representing said 1 st direction, said components constituting a cancellation signal that is in anti-phase with said passive matrix audio signal,

generating a2 nd output signal includes combining a component of each of the 2 nd pair of intermediate audio signals with a passive matrix audio signal representing the 2 nd direction, the component forming a cancellation signal that is in anti-phase with the passive matrix audio signal, and

generating the 3 rd output signal includes combining a component of each of the 1 st pair of intermediate audio signals with a passive matrix audio signal representing the 3 rd direction, the component forming a cancellation signal that is in anti-phase with the passive matrix audio signal.

8. The method of claim 7, wherein matrix audio signals representing the 1 st, 2 nd, and 3 rd directions, respectively, are generated by the passive matrix.

9. The method of claim 7, wherein passive matrix audio signals representing the 1 st, 2 nd, and 3 rd directions, respectively, are generated in a plurality of linear combiners that also combine the passive matrix audio signals with the signal components.

10. A method according to claim 1 or 2, wherein the output signals are generated separately by combining the pairs of intermediate signals.

11. A method as claimed in claim 1, 2,4 or 7, wherein the processing comprises feeding back each pair of intermediate audio signals for use in controlling the relative amplitudes of the respective pairs of intermediate audio signals respectively.

12. The method of claim 11, wherein the processing comprises applying each derived audio signal to a respective one of the variable gain circuits, wherein the gain of each of the variable gain circuits associated with each pair of derived audio signals is controlled based on the respective output amplitude of the respective variable gain circuit for each pair of derived audio signals.

13. The method of claim 12, wherein each variable gain circuit comprises a Voltage Controlled Amplifier (VCA) having a gain of g, in combination with a subtractive combiner, the resulting gain of the variable gain circuit being (1-g), and the cancellation signals being derived from the outputs of the voltage controlled amplifiers.

14. The method of claim 12, wherein each variable gain circuit comprises a Voltage Controlled Amplifier (VCA) having a gain g, the resulting gain of the variable gain circuit is g, and the cancellation signals are derived from outputs of the voltage controlled amplifiers.

15. The method of claim 12, wherein the gain of each variable gain circuit is such that the signal outputs are substantially the signals produced by the passive matrix under static input signal conditions.

16. The method of claim 12, wherein the gain of each variable gain circuit associated with each pair of derived audio signals is controlled by applying the output of each variable gain circuit in pairs to an amplitude comparator, the amplitude comparator generating a control signal for controlling the gain of each variable gain circuit.

17. The method of claim 16, wherein each amplitude comparator controls the gain of each variable gain circuit associated with each pair of derived audio signals such that, under some input signal conditions, an increase in the amplitude of the output of one variable gain circuit relative to the other variable gain circuit will result in a decrease in the gain of the variable gain circuit whose output is increased.

18. The method of claim 17, wherein the amplitude comparators control the gain of the variable gain circuits associated with each pair of derived audio signals such that, under some input signal conditions, an increase in the amplitude of the output of one variable gain circuit relative to the other variable gain circuit also results in substantially no change in the gain of the variable gain circuit whose output is not increased.

19. The method of claim 17, wherein the amplitude comparators control the gains of the variable gain circuits associated with each pair of derived audio signals such that an increase in the amplitude of the output of one variable gain circuit relative to the other variable gain circuit results in the product of the gains of the two variable gain circuits remaining substantially constant under some input signal conditions.

20. The method of claim 16, wherein each amplitude comparator controls the gain of each variable gain circuit associated with each pair of derived audio signals such that, under some input signal conditions, an increase in the amplitude of the output of one variable gain circuit relative to the other variable gain circuit will result in an increase in the gain of the variable gain circuit whose output is increased.

21. The method of claim 20, wherein the amplitude comparators control the gain of the variable gain circuits associated with each pair of derived audio signals such that, under some input signal conditions, an increase in the amplitude of the output of one variable gain circuit relative to the other variable gain circuit also results in substantially no change in the gain of the variable gain circuit whose output is not increased.

22. The method of claim 20, wherein the amplitude comparators control the gains of the variable gain circuits associated with each pair of derived audio signals such that an increase in the amplitude of the output of one variable gain circuit relative to the other variable gain circuit results in the product of the gains of the two variable gain circuits remaining substantially constant under some input signal conditions.

23. The method of claim 16, wherein the gains of the variable gain circuits, expressed in dB, are linear functions of their control voltages, each amplitude comparator has a finite gain, and the output of each variable gain circuit is applied to one amplitude comparator via a rectifier whose output signal is proportional to the logarithm of its input.

24. The method of claim 23, wherein each rectifier is preceded by a filter having a frequency response of: attenuating frequencies at the two extremes of the audible range and providing a gently rising response over the portion of the audible range between the two extremes.

25. The method as recited in claim 16, further comprising:

one or more additional control signals are derived from two control signals controlling the variable gain circuits associated with each pair of passive matrix audio signals, wherein the one or more additional control signals are derived by modifying one or both control signals and generating one unmodified control signal and one modified control signal or the lesser or greater of the two modified control signals.

26. The method of claim 25, wherein one or both of the control signals are modified by polarity inversion, amplitude offset, amplitude scaling and/or nonlinear processing of the respective signals.

27. The method of claim 25, further comprising one or more additional variable gain circuits for receiving as inputs a combination of two of said plurality of cancellation signals or a combination of two passive matrix signals, wherein said one or more additional control signals respectively control said one or more additional variable gain circuits such that when said respective input signals represent directions other than in said 1 st and 2 nd axes, the gain of the circuit rises to a maximum value, and

generating one or more additional cancellation signals by controlling the one or more additional variable gain circuits with the one or more additional control signals, respectively.

28. The method of claim 27, wherein at least 5 output signals are generated by combining each of at least 5 passive matrix audio signals with two or more of the plurality of cancellation signals and the one or more additional cancellation signals, each cancellation signal being in anti-phase with each passive matrix audio signal such that when the direction represented by the respective input audio signal is different from the direction represented by the passive matrix audio signal, the passive matrix audio signal is substantially cancelled by the respective cancellation signal.

29. The method of claim 16, wherein the amplitude of each audio signal in the 1 st pair of intermediate audio signals is expressed as

[(L_t+R_t)*(1-g_c)]Or equivalently, [ (L)_t+R_t)*(h_c)]Of and, an

[(L_t-R_t)*(1-g_s)]Or equivalently, [ (L)_t-R_t)*(h_s)]Is determined by the amplitude of the signal (c),

and the amplitude of each audio signal in the other pair of intermediate audio signals may be expressed as

[L_t*(1-g_l)]Or equivalently, [ L_t*(h_l)]Of and, an

[R_t*(1-g_r)]Or equivalently, [ R ]_t*(h_r)]Is determined by the amplitude of the signal (c),

in the formula, L_tAnd R_tIs a pair of audio signals, L, generated by said passive matrix_t+R_tAnd L_t-R_tIs another pair of audio signals generated by the passive matrix, (1-g)_c) And h_cIs the output L of the passive matrix_t+R_tGain of the associated variable gain circuit, (1-g)_s) And h_sIs the output L of the passive matrix_t-R_tGain of the associated variable gain circuit, (1-g)_l) And h_lIs the output L of the passive matrix_tGain of the associated variable gain circuit, (1-g)_r) And h_rIs the output R of the passive matrix_tThe gain of the associated variable gain circuit.

30. A method for deriving at least 3 audio signals, each associated with a direction, from two input audio signals, comprising:

generating a plurality of passive matrix signals by using a passive matrix in response to the two input audio signals, wherein the passive matrix signals comprise two pairs of passive matrix audio signals, namely a 1 st pair of passive matrix audio signals representing directions on a 1 st axis and a2 nd pair of passive matrix audio signals representing directions on a2 nd axis, and the 1 st axis and the 2 nd axis are basically orthogonal to each other;

processing said each pair of passive matrix audio signals to produce a 1 st and a2 nd pair of intermediate audio signals, respectively, the relative amplitudes of the audio signals in each pair being intended to be equalized by deriving each pair of intermediate audio signals from the respective pair of passive matrix audio signals and applying a relative scale to each pair of derived signals to generate a pair of intermediate signals whose amplitudes are closer to being equalized than the passive matrix signals from which the intermediate signals were derived;

deriving a plurality of cancellation signals from said two pairs of intermediate audio signals, wherein each of said cancellation signals is a component of an intermediate audio signal; and

generating at least 3 output signals by combining each of the at least 3 passive matrix audio signals with two or more of the plurality of cancellation signals, each cancellation signal being in anti-phase with each passive matrix audio signal.

31. The method of claim 30, wherein the processing comprises feeding back each pair of intermediate audio signals to control the relative amplitudes of the pairs of intermediate audio signals.

32. The method of claim 31, wherein said processing comprises applying each of said two pairs of passive matrix audio signals to a respective variable gain circuit, each circuit comprising a Voltage Controlled Amplifier (VCA) having a gain g, in combination with a subtractive combiner, wherein the resulting variable gain circuit has a gain of (1-g), and wherein said cancellation signals are derived from outputs of said voltage controlled amplifiers.

33. The method of claim 32 wherein the gain of the variable gain circuit associated with each pair of passive matrix audio signals is controlled by applying the respective output of the variable gain circuit for each pair of passive matrix audio signals to an amplitude comparator, the amplitude comparator generating a control signal for controlling the gain of the respective variable gain circuit.

34. The method of claim 33, wherein the outputs of the variable gain circuits for each pair of passive matrix audio signals are applied to an amplitude comparator via a rectifier whose output signal is proportional to the logarithm of its input, said comparator having finite gain, and the gain of the VCAs in dB is a linear function of their control voltage.

35. The method of claim 33, further comprising:

one or more additional control signals are derived from two control signals controlling the variable gain circuits associated with each pair of passive matrix audio signals, wherein the one or more additional control signals are derived by modifying one or both control signals and generating one unmodified control signal and one modified control signal or the lesser or greater of the two modified control signals.

36. The method of claim 35, wherein one or both of the control signals are modified by polarity inversion, amplitude offset, amplitude scaling and/or non-linear processing of the respective signals.

37. The method of claim 35, further comprising one or more additional variable gain circuits for receiving as inputs a combination of two of said plurality of cancellation signals or a combination of two passive matrix signals, wherein said one or more additional control signals respectively control said one or more additional variable gain circuits such that when said respective input signals represent directions other than in said 1 st and 2 nd axes, the gain of the circuit rises to a maximum value, and

generating one or more additional cancellation signals by controlling the one or more additional variable gain circuits with the one or more additional control signals, respectively.

38. The method of claim 37, wherein at least 5 output signals are generated by combining each of at least 5 passive matrix audio signals with two or more of the plurality of cancellation signals and the one or more additional cancellation signals, each cancellation signal being in anti-phase with each passive matrix audio signal such that when the direction represented by the respective input audio signal is different from the direction represented by the passive matrix audio signal, the passive matrix audio signal is substantially cancelled by the respective cancellation signal.