JP2000350300A - Directivity decoding means and system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To decode an audio signal for a plurality of output channels by normalizing channels according to 1st and 2nd normalizing modes in response to the discrimination as of whether or not two channels are in correlation. SOLUTION: Input channels Lt 12 and Rt 13 are given to an RMS responding detector, correlation and phase analyzer 40, where |Lt+Rt |, |Lt=Rt |, |Lt |, and |Rt | are produced. These signal quantities are given to a logic 42, which provides outputs of normalizing coefficients A1-A4 in response to a degree of correlation. A mean time interval is applicable to contents of an input signal decided by the correlation and phase analyzer 40. In the case that the input signals are not in correlation, the mean time interval is comparatively long, but when the input signals are in correlation, that is, the input signals have similar waveforms, the mean time interval is short. When the correlation of the signals is high, a differential mode normalizing coefficient is derived as result, and when the correlation of the signals is low, a common mode normalizing coefficient, is derived as result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

This application is a continuation of "encoding and decoding" of U.S. patent application Ser. No. 08 / 796,285, filed Feb. 7, 1997, now abandoned. Is a continuation-in-part of the named US patent application.

[0002] The present invention relates to the decoding of audio signals into directional channels. More particularly, it relates to a novel apparatus and method for decoding an input channel to a cardinal output channel. For background, see the above-referenced U.S. patent application and its background description.

[0003]

It is an important object of the present invention to provide an improved method and apparatus for decoding an audio signal into a plurality of output channels.

[0004]

According to the present invention, a method of processing a multi-channel audio signal comprises the steps of determining the degree of correlation between the two channels, and determining whether the two channels are correlated and uncorrelated. And normalizing the channels according to the first and second normalization modes, respectively, in response to the determination.

In another aspect of the invention, a method of processing a multi-channel audio signal includes the steps of determining the degree of correlation between the two channels, wherein the two channels are partially correlated and partially correlated. In response to the determination that they are not, the first normalization mode and the second
And processing according to the combination with the normalization mode.

In another aspect of the present invention, a method for decoding an encoded multi-channel audio signal comprises the steps of:
Determining the correlation between the first channel and the second channel; processing the first channel and the second channel to generate a third channel and a fourth channel;
including.

In yet another aspect of the invention, an apparatus for processing a multi-channel audio signal includes an input characterization determiner for determining a degree of correlation between two channels, and an input characterization determiner coupled to the input characterization determiner. A first normalizing multiplier for providing a first normalization coefficient responsive to the degree to one of the two channels, and a second normalization coefficient coupled to the input characteristic determiner and responsive to the degree of correlation. A second normalizing multiplier for applying the second signal to the second signal.

[0008]

DETAILED DESCRIPTION OF THE INVENTION Referring now to FIG. 1, a wideband directional decoded audio signal processing system 1 having two input channels and eight output channels according to the present invention.
It is shown. The input channel characteristic determiner 10 includes input channels 12 and 13 (left input channel Lt12) connected to a signal source such as a receiver, a VCR or a DVD player.
, And the right input channel Rt13). The input channel characterization unit 10 sends the input on channels 12, 13 (by signal lines 17, 19) and other signals to be described later in FIG. 3 to output channel synthesizer 14. . The output channel synthesizer 14 outputs the output signal to the output channels 50, 56, 6
Synthesize on 2, 66, 68, 70, 72, 74.

[0009] The term "channel" as used in this application means that decoding and / or processing and playback are possible at a position relative to the listener so that the listener can move sound from one direction in space. Audio information that is encoded so that it can be perceived as occurring. An input channel can be decoded into a plurality of output channels, so that the total number of output channels can be encoded to be greater than the total number of input channels. The output channel is typically
Directional designators, such as “left”, “right”, “center”, “surround”, “left surround”, and “right surround”, depending on the direction in which sound is intended to arrive Is specified. For illustration purposes, input channels 12, 13 and output channel 5
0, 56, 62, 66, 68, 70, 72, 74
It is shown as a separate element. The number of input channels is
It is not necessarily the same as the number of physical signal lines transmitting information in the channel. Digital signal transmission system
Typically, it has one signal line for transmitting a plurality of input channels. Input channels are typically encoded as analog electrical signals or as digital bit streams.

[0010] "presentation channe"
l) means a channel available for decoding or playback, and a "reproduction channel"
Refers to a channel that has been decoded and is intended for playback by a device such as a loudspeaker.

[0011] Information in an output channel can be "cardinal" if it is exclusively and uniquely associated with the direction associated with that output channel. In other words, if the information in the output channel includes only information in the direction associated with the output channel and no other information includes information in the direction associated with the output channel, then the output channel is And the associated direction is an azimuthal direction. Thus, for example, a left surround channel is said to be in an azimuthal state if the left surround channel contains only left surround signal content and none of the other channels contain left surround signal content. , Left surround direction is referred to as an azimuth direction, and the position of the azimuth direction with respect to the listener is referred to as an azimuth position.

Referring now to FIG. 2, there is shown an example of input channel information. In FIG. 2, the input channel information is encoded as a signal level, typically measured as volts v versus time t. For ease of explanation, some channels (eg, input channel L
The signal level at t) is referred to in the formula as Lt. Similarly, for example, the time average of the signal level in a certain channel (for example, the input channel Rt) is | Rt
(Note: Here, the symbol ず れ is shifted from Rt, but has the same meaning as the overline seen in the following Equation 1 and the like. The same applies hereinafter.) The channels Lt and Rt
, The difference between the signal levels is expressed as Lt−Rt, the time average of the sum of the signal levels is expressed as Lt + Rt￣, and the channel L
The absolute value of the time average of the difference between the signal levels at t and the channel Rt is expressed as | Lt−Rt￣ |, and the same applies to other signals. Typical time average intervals are from about 5 ms to about 1000 ms. The length of the time averaging interval will be discussed later in conjunction with FIG. The input channel information is
It can also be digitally encoded as a bit stream of signal levels measured at time intervals.

Referring now to FIG. 3, the input channel characterization unit 10 is shown in more detail. Input channel L
t12 and Rt13 are input to the RMS response level detector and the correlation and phase analyzer 40, and the next time average signal amount is generated from the RMS response level detector and the correlation and phase analyzer 40.

[0014]

(Equation 1)

These quantities are provided to logic 42, which determines the quantity X, which is the greater of (1) and (2).
And the quantity Y which is the larger of (3) and (4).
The signal quantities (1) and (2) are combined with the signal quantities (3) and (4) and together with the quantities Y and X are normalized coefficients A1, A
2, A3 and A4. Quantities (1), (2),
(3) and (4) and the quantities X and Y used to construct A1, A2, A3 and A4 are the correlation and phase determined by the RMS response level detector and correlation and phase analyzer 40. Depends on relationship information. Input channel Lt
And Rt are correlated (hereinafter referred to as "panned mono"), A1, A2,
The values of A3 and A4 are as follows.

[0016]

(Equation 2)

The domain of all normalization coefficients is from 0 to 1 including both ends. Therefore, for conditions where the sum signal is dominant (sum signal dominance), the normalization coefficients (A2) and (A3) are evaluated as zero. Similarly, for conditions where the difference signal is dominant, the normalization coefficients (A1) and (A4) are evaluated as zero.

The normalization factors applied to the signals in channels Lt and Rt are different. In the case of the normalization coefficients A1 and A2, the normalization coefficient responds to the sum or difference of the two signals and the magnitude of Lt, whereas in the case of the normalization coefficients A3 and A4, the normalization coefficient is two signals. , And the magnitude of Rt. This type of normalization mode applies a different normalization factor to the input signal and calls it "differential (di
fferential mode).

In one embodiment, the time averaging interval may be adapted to the content of the input signal as determined by the correlation and phase analyzer 40. If the input signals are uncorrelated, the averaging interval is relatively long (eg, about 1000 m).
s). If the input signals are correlated, ie have similar waveforms, the time interval is short (eg, about 5 ms). If the magnitude of the signal is relatively small, the time averaging interval is short. When both input signals are close to zero, the time averaging interval is short. When the difference between the signal magnitudes is large (for example, | Lt−Rt | ≧ 20 dB), the time average interval is short. A common way to achieve a time averaging interval is to measure the signal periodically and weight each measurement exponentially smaller than the preceding measurement. Using this measurement, the average interval is typically, for example, 1 /
It is expressed as the time period required for the weighting of the measurement to decrease to a part of the weighting of the previous measurement, such as 3.

Referring now to FIG. 4, a first portion of the circuitry of the output channel synthesizer 14 is shown. L
t input channel 12 is provided to multipliers 22 and 24;
There, the normalization coefficients A1 and A2 are respectively multiplied to form the post-normalized channels Lc'30 and Ls'32. Similarly, the Rt input channel 13 is provided to multipliers 34 and 36, where the normalization factor A
3 and A4 are each multiplied to form post-normalized channels Rs'30 and Rc'32, respectively.

The input signals at Lt and Rt are correlated, and furthermore, are in phase or have a relative phase difference of 180
If they are phase shifted so as to be equal to one another, Lt to Lc 'independent of the relative amplitude difference (if any) applied to input terminals Lt and Rt. (Or Ls '), the contribution from Rt to Rc'
(Or Rs'). Furthermore,
From Lt to Lc '(or Ls') and from Rt to Rc'
(Or Rs') is equal to the smaller of the amplitudes of the two input signals at Lt and Rt. The resulting normalized output signals in (A1) to (A4) are equal amplitude monaural contributions from Lt and Rt that are directionally identified as center channel or center surround channel components. If the input conditions at Lt and Rt include both the center channel signal and the surround channel signal, but are considered to produce either a sum signal dominant or a difference signal dominant condition, then a normalization function (normalizat
ion function) responds solely to the dominant condition. Thus, the sum-dominant normalized signal that occurs at the outputs of (A1) and (A4) may include a non-dominant surround channel signal. Similarly, (A2) and (A3)
The differentially dominant normalized signal that occurs at the output of the may include a non-dominant surround channel signal. The surround channel signal present at the outputs of (A1) and (A4) during the sum signal dominant condition will output the output of (A4) to (A
It is obtained by subtracting from 1). Surround
The channel signal is 180 at the input terminals Lt and Rt.
It is identified as having a relative phase difference of degrees. Similarly, during the difference signal dominant condition (A2) and (A3)
Is obtained by adding the output of (A3) to the output of (A2). The center channel signal is identified as an in-phase signal appearing at input terminals Lt and Rt.

The normalization function illustrated in FIG. 4 has important properties. (A1) and (A4) if the input signals at Lt and Rt include the dominant center channel signal and at the same time include uncorrelated and unequal amplitude signals (Lt or Rt is dominant). The contribution of the normalized Lt and Rt signals at the output of
Normalized L instead of equal contribution of the amplitude of the input signal
Include equal contributions of the magnitudes of the t and Rt input signals. When there is a sum signal dominant condition and an Lt or Rt dominant condition, the output of (A4) is divided from (A1) to obtain a surround channel signal, so that a part of the center channel signal is reduced. Will lead to a surround channel. By adding the output of (A2) with (A3) to obtain the center channel signal when there is a difference signal dominant condition in Lt and Rt between the Lt or Rt dominant conditions, A portion of the signal will be directed to the central channel. Therefore, when the input conditions at Lt and Rt are phantom, a differential based normalization function is particularly desirable. However, it is desirable to adapt the normalization function to the input signal conditions whenever the input is not panned mono.

As another feature of the present invention, the present invention relates to Lt
And the method of providing an improved normalization mode when the content of Rt and Rt is other than a panned thing. Referring again to FIG. 3, if the RMS response level detector and the correlation and phase analyzer 40 detect that the signals at Rt and Lt are uncorrelated, the logic 42 proceeds to A1, A2, A3 and A4. Outputs the following values for

[0024]

(Equation 3)

These normalization coefficients are formed by combining the signal quantities (1) and (2) with the Y variable,
It does not include the signal amounts | Lt￣ | and | Rt￣ |.

These normalization coefficients are formed by combining the signal quantities (1) and (2) with the Y variable. The Y variable is common to the normalization coefficients applied to both Lt and Rt, and does not include the signal amounts | Lt￣ | and | Rt￣ |. This type of normalization mode applies a common normalization factor to the input signal and is referred to as "common mode".

[0027] The time averaging interval can vary as discussed above.

When the Y variables for the signal amounts (3) and (4) are substituted into the normalization coefficients (A1) to (A4), the normalization coefficients (A1) to (A4) change from the differential mode to the common mode. I do. When the signals in the input channels Lt and Rt are uncorrelated, any assumed Lt
And the value of A1 for the Rt input condition is equal to the value of A4. Similarly, the value of A2 is equal to the value of A3.

Referring now to FIG. 4, using the new values of A1 through A4, the previous input signal condition at Lt and Rt is that the central channel signal dominated by Lt or Rt and at the same time If so, it produces equal center channel signal contributions from Lt and Rt at the outputs of A1 and A4. Even if the output of A4 is subtracted from A1,
It does not guide the center channel signal in the surround signal. Furthermore, even if the output of A2 is added to A3, Lt and R
If the input signal at t does not include a dominant surround channel signal with an Lt or Rt dominant signal, the surround channel signal will not be steered into the center channel. Therefore, when the input signals at Lt and Rt are uncorrelated, a common mode based normalization function is desirable. The signals in the input channels Lt and Rt are not correlated and form a transform coefficient (A7), and the input channels Lt and Rt are (A1) to (A1).
When it is correlated with the value of 4), the normalization coefficient (A1)
(A4) can be linked, and the conversion coefficient A7 and the operator A8 are defined as follows.

[0030]

(Equation 4)

Where ε is any number much smaller than any other quantity inserted so that the circuit does not include division by zero even if the other term in the denominator evaluates to zero. It is.

The normalization coefficients (A1) to (A4) can be generated as follows.

[0033]

(Equation 5)

The generalized forms of A1, A2, A3 and A4 are applicable to all degrees of correlation and phase. If the signal is highly correlated, these generalized equations result in differential mode normalization coefficients. If the degree of signal correlation is low, these generalized equations result in common mode normalization coefficients. In the case of a partially correlated signal, the generalized equation will result in having some differential content and some common content. This type of normalization is called "complex mode".

Referring now to FIG. 5, a second portion of the circuitry of the output channel synthesizer 14 is shown. The post-normalized channels of FIG. 4 are combined to form the following interim channels Lc50, Ls''5
2, yielding Rs ″ 54 and Rc56.

[0036]

(Equation 6)

The intermediate channels are divided into normalization coefficients A1 to A4.
The following equation is obtained by writing using.

[0038]

(Equation 7)

Referring now to FIG. 6, added to the circuit of FIG. 5 are intermediate channels Lo'60 and Ro'62 which produce the following Lo 'and Ro' at the output of the combiner.

[0040]

(Equation 8)

The normalization factor responds solely (and thus exclusively) to the dominant input signal condition in Lt and Rt. If the input signals at Lt and Rt are dominant in the sum signal and the input signals at Lt and Rt are correlated and in phase, then only normalization multipliers (A1) and (A4) are active. If the input signals at Lt and Rt are dominant in the difference signal and the input signals at Lt and Rt are correlated and there is a relative 180 degree phase shift, the normalizing multipliers (A2) and (A3) Only) is active. When the input signals at Lt and Rt are not correlated (or have a quadrature relationship), the magnitude of the sum signal and the magnitude of the difference signal are equal, and all the normalizing multipliers (A1) to (A4) Are activated by the same number.

The result of subtracting the correlated Rc signal from the Lt input results in a correlated in-phase (or
(Center channel) The decrease in signal magnitude. This does not reduce the magnitude of the uniquely left channel signal component, since Rc has no uniquely left channel signal component at all.
The amount of Rc signal removed from the Lt input is Lt and Rt
It has a linear relationship with the degree of correlation between input signals. Lc
The same result is obtained when the signal component is subtracted from the Rt input. The magnitude of the correlated in-phase signal component at Rt decreases in proportion to the degree of correlation between the Lt and Rt input signals.

The result of adding the correlated (but out of phase) Rs ″ signal to the Lt input is:
A reduction in the magnitude of the correlated but out-of-phase (or surround) channel signal in Lt. This does not reduce the magnitude of the uniquely left channel signal component, since Rs ″ has no uniquely left channel signal component. The amount of Rs ″ signal removed from the Lt input has a linear relationship with the correlation between the Lt and Rt input signals and the degree of phase shift.
The same result is obtained when a signal component having a phase shift and a correlation in Ls '' is added to Rt. Rt
The magnitude of the correlated, but out-of-phase, signal component at decreases in proportion to the degree of correlation between the Lt and Rt input signals.

When there is no correlation between the input signal conditions at Lt and Rt, the matrix of terms Rs "-Ls and Ls" -Lc decreases (respectively) to the next value.

[0045]

(Equation 9)

Thus, the Lt and Rt input signals are reduced by subtracting the normalized magnitude of Lt from Lt and subtracting the normalized magnitude of Rt from Rt, respectively. Thereby, the magnitudes of Lo 'and Ro' are correspondingly reduced.

When considering the properties of the signals Lc, Rc, Ls ″ and Rs ″, it is necessary to recall the following.

[0048]

(Equation 10)

Since Lc ′ and Ls ′ are components of the normalized Lt input and Rc ′ and Rs ′ are components of the normalized Rt input, the Lc ′ signal is cumulatively different from the Ls ′ signal. ), And the Rc ′ signal is Rs ′
Combined with signals cumulatively. (A1) through (A
The normalization coefficient variables in 4) are numerically identical when the input signal conditions at Lt and Rt have no correlation in nature. For this condition, Lc and Ls ''
The Lt contribution to Lc and Ls ″ is dominant by a factor of 3, ie, about 10 dB, when compared to the contribution of Rt to Lc and Ls ″. R to Rc 'and Rs''
The contribution of t is dominant by a factor of 3, ie 10 dB, when compared to the contribution of Lt to Rc ′ and Rs ″. Thus, the Lc ′ and Ls ″ signals are substantial components of the normalized Lt input, and Rc ′ and Rs ″
The signal is a substantial component of the normalized Rt input. L
If the c ′ and Rc ′ signals are reproduced by separate loudspeakers located to the left and right of the center respectively, L
The stereophonic content of the uncorrelated signal at t and Rt is substantially preserved. The signal processing system according to the present invention uses a separate Lc and Rc separate Lc and Rc center contributions whenever Lt and Ct contributions can be practically used in a reproduction system. Play as a signal. This is because Lt and Rt are matrix-encoded by adding some (or all) of the component signals in Lt.
It is superior to audio signal processing systems that derive the center channel from the t stereophonic signal. The values of the normalization coefficients of the input normalization multipliers (A1) and (A4) are
It should be recalled that the input signal at Lt and Rt is almost zero whenever the difference signal is dominant. Lt and Rt during the condition that the difference signal is dominant
The center channel signal that can be present at t is Lc
And Rc are defined as follows.

[0050]

[Equation 11]

Assuming such conditions for the Lt and Rt input signals, Lc and Rc are the same. When the signals Ls ′ and Rs ′ are added in Lc and Rc, L
This forces c and Rc to be monaural in nature. Adding the Lt and Rt component signals in Ls 'and Rs' to produce Lc and Rc ensures that the Lc and Rc signals do not contain the dominant surround channel signal. The contents of Lc and Rc are
When the input conditions at Lt and Rt are uncorrelated and stereophonic in nature, they are very stereophonic, whereas when the input signal properties at Lt and Rt are dominant in the difference signal (or surround channel). Always monaural. Channels Lc and Rc
Is very monaural whenever the input signal conditions at Lt and Rt are substantially correlated.

The intermediate (in) in Ls ″ and Rs ″
The terim) signal likewise decreases to the next value whenever the input signal conditions at Lt and Rt are substantially sum signal (center channel) dominant.

[0053]

(Equation 12)

The values of the normalization coefficients in the input normalizing multipliers (A2) and (A3) are almost zero whenever the input signal condition in Lt is dominant in the sum signal (or center channel). The surround channel signal that may be present at Lt and Rt during the sum signal dominant condition is R
subtracting the signal component of c ′ from Lc ′ to produce Ls ″,
Similarly, the signal component of Lc ′ is subtracted from Rc ′ to obtain R
s ''. Rc 'to Lc'
From Lc ′ to generate Rs ″ and Lc ′ from Rc ′ to generate Rs ″, whenever Lt and Rt are substantially sum signal (or center channel) dominant. It is guaranteed that '' and Rs '' do not contain any central signal component. Intermediate signals Ls ″ and R
The components of s '' are very stereophonic in nature whenever the input signal conditions at Lt and Rt are uncorrelated or substantially stereophonic in nature. The intermediate signals Ls '' and Rs '' are Lt and Rs
Whenever the input signal condition at t is substantially sum signal (center channel) dominant, it is substantially monaural in nature. The stereophonic nature of the intermediate signals Ls ″ and Rs ″ for uncorrelated input signals at Lt and Rt is such that some (or all) of the Rt input signal is
It is superior to audio signal processing systems that derive a mono surround channel signal from matrix-encoded Lt and Rt signals by subtracting from the t input signal.

The intermediate signal at Ls ″ and Rs ″ is very stereophonic in nature when the input signal conditions at Lt and Rt are uncorrelated,
It does not show exclusive cardinal states. With the encoded Lt and Rt signals, an exclusive left surround channel signal or an exclusive right surround channel signal appears at Lt and Rt, respectively, as follows. In the equation (13), the upper part is for an exclusive left surround channel signal input, and the lower part is for an exclusive right surround channel signal input.

[0056]

(Equation 13)

Exclusive left only or right only surround
For the channel signal, the Lt or Rt coded signal encodes the difference signal dominant condition using the Lt or Rt dominant condition. Furthermore, the encoded Lt and Rt signals are phantom mono. The advantage of the audio signal processing system according to the invention is that given the encoded Lt and Rt signal conditions, it can be decoded as an exclusive left only or right only surround channel signal.

Referring now to FIG. 7, another portion of the channel synthesizer 14 is shown. Intermediate channel L
The s "52 and Rs" 54 signals are combined and the left front channel Lo64, the right front channel Ro66, the left center surround channel Lcs68, the right center surround
Channel Rcs70, left surround channel Ls72
And the right surround channel Rs 74 occurs according to the following equation:

[0059]

[Equation 14]

However, A5 and A6 are given by the following equations.

[0061]

(Equation 15)

The effect of the circuit of FIG.
s '' and Rs '' are re-matrixed with normalization coefficients A5 and A6. Out-of-phase (or surround channel) signals are cumulatively combined, while in-phase (or center channel) signals are differentially combined. By re-matrixing Ls "and Rs", all central channel signals that may be present in Ls "or Rs" while there are differential dominant and uncorrelated input signal conditions in Lt and Rt The amplitude of the component is correspondingly reduced. The process of re-matrixing the Ls "and Rs" signals further reduces the stereophonic content of Ls "and Rs", but the Lt of Lt
The contribution to s ″ is still dominant as compared to the contribution of Rt to Ls ″. Similarly, the contribution of Rt to Rs ″ is still dominant as compared to the contribution of Lt to Rs ″. Thus, when the signal conditions at Lt and Rt are substantially uncorrelated, the re-matrixed Ls "and Rs" signals still retain stereophonic properties. Considering the panned mono and correlated out-of-phase input conditions at Lt and Rt, the signals Ls ″, Rs ″, Lo ′ and R ′
It is helpful to examine the nature of o 'again. The normalized contribution of Lt and Rt in Ls ″ and Rs ″ is
The input signal conditions at Lt and Rt are the signals Lt and Rt
Is substantially monaural when correlated and out of phase irrespective of the relative amplitude of. Ls ''
And the normalized contribution of Lt and Rt in Rs ″ is equal to the smaller of the two input signals Lt and Rt whenever their relative amplitudes are different. Therefore,
The differentially dominant and Lt dominant input signal condition with correlation in Lt and Rt results in Lt and Rt to L
causes a contribution equal to the Rt input signal amplitude to s ″ and Rs ″.

Since these signals are removed from Lt and Rt to produce intermediate signals Lo ′ and Ro ′ (shown in FIG. 6), the Lo ′ intermediate signal is the differential surround dominant in Lt. -Includes channel signals. Lt and Rt that are correlated, out of phase and dominant in Rt
The same observation can be made for the intermediate signal Ro ′ corresponding to the input signal in. The output of the multiplier (A5) or (A6) is Ls '' or Rs '' which is a component of the originating encoded Ls or Rs input signal condition.
Is equal to the contribution to Lt or Rt at. Therefore, the input signal condition dominated by Lt and dominated by the difference signal is Ls
Defined as the dominant output signal condition. Similarly, Rt
The dominant and difference signal dominant input signal condition is defined as the Rs dominant output signal condition. Directionally directional Ls
Or Rs encoded signal condition is azimuthal Ls or Rs
Decoded as s output signal condition. At this point, the decoder is the complement of the encoded signal condition.
The output signals at Ls and Rs are almost zero whenever the encoded signals at Lt and Rt are equal amplitude signals. For this condition, the encoded signal is decoded at the Lcs and Rcs output terminals. At this point, the decoded output signal condition is the directional complement of the encoded signal condition.

Referring now to FIG. 8, according to the nature of the decoding method disclosed herein, the signal is, as shown, placed on a Lo6 located relative to a listener 78.
2, Lc50, Rc70, Ro66, Ls72, Lcs
68, Rcs56, and Rs74.

It is also possible to decode the matrix-encoded Lt and Rt signals into six directional positions in a 360 degree space. Intermediate directional positions are "phantom" sources based on the presence of decoded signals in multiple channels. For example, a signal is complementarily encoded and then decoded,
It can occur at any point between the left channel output and the left surround channel output. Similarly, a signal is complementarily encoded and then decoded,
It can occur anywhere between the right output channel and the right surround output channel. Therefore,
A signal may be encoded and then decoded to occur at any point within a 360 degree spatial angle.

Rendering of the source adjacent to the left or right side of the listener is more easily perceived when the physical playback channel is at the foreseen spatial angle. The availability of even more presentation channels, especially in larger commercial locations such as cinemas using multiple playback channels, is a particular advantage of this aspect of the invention.

Concurrent US Patent Application Serial No. 08/796
By combining the pair-wise decoding technique disclosed in the description of FIG. 17 of FIG. 17 and the description on page 16 with the decoding technique of this application, a greater number of playbacks in a commercial system are possible. An effect can be obtained using the loudspeaker. Thereby, the opposition (opposi) contained in either the matrix-decoded Lt / Rt signal or the resulting discrete medium.
te) Processing the channel information to provide additional azimuthal presentation channels adjacent to the left and right sides of the viewing audience.

In many applications, using as many as eight physical loudspeakers is impractical. Today's home playback systems are more typically equipped with five physical playback loudspeakers. In addition, the introduction of a 5.1 channel discrete media presentation system determines the number of physical playback loudspeakers typically used. For convenience (ie, compatibility with a limited number of physical presentation loudspeakers and discrete media presentation formats), this disclosed algorithm for playback over five physical playback channels. Is preferably down-mixed. This is done by combining the channels as follows.

[0069]

(Equation 16)

By downmixing the decoded output channel, the number of azimuthal directional states is not reduced, but rather the number of ways in which the azimuthal directions can be reproduced. The azimuthal Ls and Rs directional states are still maintained. The stereophonic nature of the signal at Ls and Rs is preserved as well. Exclusive Lcs / Rcs output conditions are reproduced as equal amplitude signals in Ls and Rs. Similarly, the Lc and Rc output signals occur at a single center channel output, thus
Save the direction of the azimuth center only.

Referring to FIG. 9, there is shown a state in which the circuits of FIGS. 3, 4, 5, 6, and 8 are combined. The composite block diagram of FIG. 9 is made up of the individual block diagrams of FIGS. 4, 5, 6, and 7. Boolean switches 80 and 82 are incorporated in FIG. 9 to enable or disable central channel decoding and / or surround channel decoding. With both sets of switches in the off state, the input signals at Lt and Rt are Lo and Ro.
It is given to the output terminal. When the surround channel mode switch is set to the off state, the surround channel signals at Lt and Rt are provided to the Lo and Ro output terminals. Similarly, when the center channel mode switch is turned off, the center channel signals at Lt and Rt are applied to the Lo and Ro output terminals.

In many cases, the number of provided playback channels is less than the number of available presentation channels. In such a case, it is advantageous to process the smaller number of playback channels so that the resulting number of playback channels equals the number of available presentation channels. In addition, today's signal transport formats allow for the transmission of a minimum of one channel and a maximum of five channels while using low frequency effect channels (spectrally limited). Dolby (registered trademark) (Do
lby) In some signal transport formats, such as AC-3, information identifying the intended playback channel format is included in the transport format as complementary data. This complementary data can also be used as a means of reformatting the number of intended playback channels into further processing into the number of available presentation channels. The provided playback channel information is defined by the number of forward and backward (surround) playback channels. The most popular formats are:

(1) Front channel is 3 and rear channel is 2 (stereo surround) (2) Front channel is 2 and rear channel is 2 (no center channel) (3) Front channel is 3 The rear channel is 2 (mono surround). (4) 2 channels (no center or surround channels). (5) 2 channels Lt / Rt, matrix coding. It is also possible that other playback formats can be desired. It should be understood that it is equally possible to process other intended playback formats using the techniques disclosed in this application.

In any case, it is desirable to process only the necessary channels to obtain the desired number of presentation channels.
In the following description, it is assumed that the number of available presentation channels is five. Therefore, the Lcs, Rcs, Lc and Rc outputs of the decoding system shown in FIG. 9 are considered to be downmixed as described above. But,
The available presentation channel signals are not limited to five.

In the case of format (1), the channels are processed discretely.

In the case of format (2), only the provided left and right playback channels are processed as Lt and Rt, resulting in a new left, right and (derived) center presentation channel signal. The surround channel signal resulting from the format (2) is not processed, and the surround channel mode switch 80 in the block diagram of FIG. 9 is set to the off state.

For format (3), a given channel format is first converted to a matrix format for processing. This involves downmixing a given mono surround channel to a given left channel to form Lnew, and further downmixing (given out of phase) a given mono surround channel to a given right channel. Then Rn
This is achieved by forming ew. Lnew and Rnew are then input to the decoder to obtain new left, right, left surround and right surround presentation channels. The center channel mode switch 80 is set to the off state because the resulting center channel signal is not processed and is reproduced as provided.

For formats (4) and (5), the applied signals are input to the circuit of FIG. 9 as Lt and Rt.

The preprocessing for various formats is shown in FIG.
0 is given an appearance.

Other embodiments are also included in the scope of the present invention as defined in the appended claims.

[Brief description of the drawings]

FIG. 1 is a block diagram of an audio signal processing system.

FIG. 2 is a representation of an audio signal useful in describing characteristics of the audio signal.

FIG. 3 is a block diagram of an input characteristic determiner according to the present invention.

FIG. 4 is a first part of the circuit of an output channel synthesizer according to the invention.

FIG. 5 is a second part of the circuit of the output channel synthesizer according to the invention.

FIG. 6 is a third part of the circuit of the output channel synthesizer according to the invention.

FIG. 7 is a fourth part of the circuit of the output channel synthesizer according to the invention.

FIG. 8 is an illustration of an arrangement of audio playback speakers coupled to the output of an output channel synthesizer according to the present invention.

FIG. 9 is a diagram combining the circuits of FIGS. 4 to 7;

FIG. 10 is a circuit diagram illustrating signal pre-processing to an audio signal processing system.

Continuation of the front page (72) Inventor Hilmar Rehnat, Massachusetts, USA 01701-9168, Framingham, The Mountain, Bose Corporation

Claims (21)

[Claims] 1. A method of processing a multi-channel audio signal, comprising: determining a degree of correlation between two channels; and responsive to determining that the two channels are correlated, a first method comprising: Normalizing the channels according to a normalization mode of the method, and responsive to a determination that the two channels are uncorrelated, normalizing the channels according to a second normalization mode. 2. The method according to claim 1, wherein said first normalization mode is a differential mode. 3. The method of claim 2, further comprising determining a phase relationship between the two channels. 4. The method of claim 3, wherein said differential mode is differential signal dominant in response to determining that said two channels are substantially out of phase. 5. The method of claim 3, wherein said differential mode is sum signal dominant in response to determining that said two channels are substantially in phase. 6. The method of claim 1, wherein said second normalization mode is a common mode. 7. The method of claim 6, further comprising determining an absolute value of a sum signal of the two channels and an absolute value of a difference signal of the two channels. 8. The method of claim 7, wherein the common mode is sum signal dominant in response to determining that the absolute value of the sum signal is greater than the absolute value of the difference signal. 9. The method of claim 7, wherein the common mode is difference signal dominant in response to determining that the absolute value of the difference signal is greater than the absolute value of the sum signal. 10. A method of processing a multi-channel audio signal, the method comprising: determining a degree of correlation of two channels; and determining that the two channels are partially correlated and not partially correlated. Responsive to the determination, processing the channel according to a combination of a first normalization mode and a second normalization mode. 11. The method of claim 1, wherein said first normalization mode is a differential mode. 12. The method of claim 1, wherein the second normalization mode is a common mode. 13. The combination of claim 1, wherein the combination is a linear weighted combination of the first mode and the second mode.
0. The method of claim 0. 14. The method of claim 13, wherein said first mode is a differential mode and said second mode is a common mode. 15. A method for decoding an encoded multi-channel audio signal, comprising: determining a correlation between a first channel and a second channel; and determining the correlation between the first channel and the second channel. Processing the three channels to produce a third channel and a fourth channel. 16. In response to a determination that the first channel and the second channel are uncorrelated, the third channel
16. The method of claim 15, wherein the fourth channel is substantially uncorrelated with the fourth channel. 17. A method according to claim 17, wherein said third channel and said fourth channel are substantially correlated in response to determining that said first channel and said second channel are substantially correlated. 16. The method of claim 15, wherein the method comprises: 18. The method of claim 15, further comprising determining an absolute value of a sum of the first channel and the second channel. 19. The third channel and the fourth channel are substantially correlated in response to an absolute value of the sum signal being greater than an absolute value of the difference signal. 19. The method according to 18. 20. The third channel and the fourth channel are substantially uncorrelated in response to the absolute value of the difference signal being greater than the absolute value of the sum signal. 19. The method according to 18. 21. An apparatus for processing a multi-channel audio signal, comprising: an input characteristic determiner for determining a degree of correlation between two of the channels; and an input characteristic determiner coupled to the input characteristic determiner for responding to the degree of correlation. A first normalizing multiplier for applying a first normalizing coefficient to one of the two channels, and a second normalizing coefficient coupled to the input characteristic determiner and responsive to a degree of correlation. A second normalizing multiplier applied to the second signal.