Method and apparatus for applying reveberation to multichannel audio signal using spatial label parameters. RU patent 2509442.

FIELD: information technology.

SUBSTANCE: method of applying reverberation to an M-channel reduced input audio signal which indicates X separate audio channels. In response to spatial label parameters, which indicate a spatial image of the reduced input signal, Y discrete signals of the reverberated channel are generated, where each of the signals of the reverberated channel at a time t is a linear combination of a subset of values of X separate audio channels at time t. The Y discrete signals of the reverberated channel are generated using a premixing matrix containing time-variable coefficients which are determined in response to spatial label parameters.

EFFECT: enabling separate determination and generation of different reverberation audio signals for each discrete channel subjected to composite audio signal upmixing.

15 cl, 3 dwg

BACKGROUND OF THE INVENTION

1. The technical field of the invention

The invention relates to methods and systems applied reverb reduced to multi-channel audio signal, indicating a greater number of separate audio channels. In some cases, implementation is achieved by increasing mixing of input signal and use of reverb to at least some of its individual channels in response to at least one parameter of spatial label (indicating, at least one spatial label for input) so that for each channel to which you apply reverb, used different pulse characteristics reverb. Not necessarily, after applying reverb separate channels are downward mixing to generate the N-channel reverberated output signal. In some embodiments, the implementation of the input signal represents MPEG encoded Surround (MPS) the signal in the area QMF (quadrature mirror filter), which increases the mixing and application reverb run within the scope of the QMF in response to the parameters of the spatial labels MPS that include at least some of the parameters of a difference of levels (CLD), parameters of the coefficient prediction channel (CPC) and parameters of the channel coherence (ICC).

2. Background of the invention

Throughout this specification, including the claims of the expression «reverb» (or «system " reverb») is used to denote a system that is configured for use reverb to the sound signal (such as all or some multi-channel audio signal).

Throughout this specification, including the claims of the expression «the system» is used broadly to mean the device, system or subsystem. For example, the subsystem that implements the reverb can be called a system of reverb (or reverb), and the system that includes the specified subsystem reverb (for example, the system decoder, which generates X+Y output signals in response to Q+R input signals, in which the subsystem reverb generates X output signals in response to Q input signals, and other output signals are generated by another subsystem decoder), also called system of reverb (or reverb).

Throughout this specification, including the claims of the expression «play» acoustic systems means creating conditions to generate sound speaker systems in response to signals, which consists of performing any required the strengthening and/or other processing of signals.

Throughout this specification, including the claims of the expression «the linear combination of the variables v 1 , v 2 , ..., v n (for example, n elements subsets of X signals separate audio channel occurring at time t, where n is less than or equal to X) denotes the value of a 1 v 1 +a 2 v 2 +...+a n v n , where a 1 , a 2 , ..., a n ratios. In General, for the values of the coefficients are no restrictions (for example, each coefficient can be positive, negative, or zero). This disclosure statement is used in a broad sense, as well, including the case when one of coefficients equal to 1, and the remainder is zero (for example, in the case when the linear combination of a 1 v 1 +a 2 v 2 +...+a n v n is v-1 (or v 2 , ..., v or n )).

Throughout this specification, including the claims of the expression «option spatial label» multichannel sound signal means any parameter that points to at least one spatial tag for the audio signal, where each specified «spatial tag is pointing to (for example, describing the spatial image multichannel signal. Examples of spatial labels is the difference of levels (or intensity) between (or correlations between pairs of channels of audio signal, phase difference between these pairs of channels and criteria correlation between these pairs of channels. Examples of parameters spatial labels are the parameters of the difference in levels of channels (CLD) and factor parameters prediction channel (CPC), which form part of the bitstream traditional MPEG Surround ("MPS") and used in encoding MPEG Surround.

In accordance with the well-known MPEG Surround ("MPS") of multiple channels of audio data can be encoded by down-mixing in fewer channels (for example, M channels, where M, as a rule, is equal to 2) and compression specified M-reduced channel audio signal can be decoded by the decompression and processing (Overdrive mix) with the purpose of generating a N decoded audio channels (for example, M = 2, N = 5).

Typical traditional decoder MPS acts, performing increase the mixing to generate N decoded audio signals (where N is greater than two) in response to the dual reduced sound signal in the time domain (and the parameters of the spatial labels MPS, including parameters on the difference between the levels of channels (CLD) and factor parameters prediction channel (CPC). Typical traditional decoder MPS operates in binaural mode, generating binaural signal in response to the dual reduced sound signal in the time domain and the parameters of the spatial labels, and at least one mode, performing increasing mixing with the purpose of generating decoded audio channels 5.0 (where the callout channels "x.y" means "x" full band channels and «u» channel subwoofer), 5.1 7.0 or 7.1 in response reduced to 2-channel audio signal in the time domain and the parameters of the spatial labels. The input signal is converted from the time domain into the frequency domain QMF (region quadrature mirror filter), forming two channels frequency components region QMF. These frequency components are decoded in the field QMF, and the resulting components, as a rule, then converted back into the time domain with the purpose of generating sound is output to the decoder.

Figure 1 is a simplified block diagram of the elements of traditional decoder MPS configured to generate N decoded audio channels (where N is greater than two and N, as a rule, is equal to 5, or 7) in response to the dual reduced sound (L' and R') and the parameters of the spatial labels MPS (including the options of a difference of levels and factor parameters prediction channel). Reduced input (L' and R') refers to the "X" separate audio channels, where X is greater than 2. Reduced input, usually points on five separate channels (for example, left front, right front, center, left and surrounding right surrounding channels).

Each of the input signals, the «left» of the input signal L' and right input signal R', is a sequence of frequency components region QMF generated by converting a dual-channel encoded signal MPS in the time domain (not shown ) in the transition phase from the time domain into the area QMF (not shown).

Reduced input L' and R' is decoded in decoder 1 of figure 1 in N signals for the different channels S1, S2, ..., SN in response to the parameters of the spatial labels MPS who are sent together with the input signals to the system of figure 1. N sequences output frequency components region QMF, S1, S2, ..., SN, as a rule, are transformed back to the staging area in the transition phase from the area of the QMF, the temporary area (not shown) and can be sent as output from the system without post-processing. Optional signals S1, S2, ..., SN are postprocessing in the field QMF) in the postprocessor with the purpose of generating an N-channel sound is output, including channels OUT1, OUT2, ..., OUTN. N sequences output frequency components region QMF, OUT1, OUT2, ..., OUTN, as a rule, is converted back to the staging area in the transition phase from the area of the QMF, the temporary area (not shown) and can be sent as output from the system.

Traditional decoder MPS of figure 1, functioning in binaural mode generates dual channel binaural sound signal output S1 and S2, and, optionally, also dual channel binaural sound signal output OUT1 and OUT2 outputs in response to the dual reduced sound (L' and R') and the parameters of the spatial labels (including the options of a difference of levels and factor parameters prediction channel). When playing a pair of headphones-channel audio output signal S1 and S2 is perceived tympanic membranes the listener like the sound of «X» speakers (where X > 2 and X, as a rule, equal to 5 or 7)located in any of the many possible positions (determined by factors decoder 1), include provisions to the listener and behind the listener. In binaural mode postprocessor can apply reverb to two-channel audio output signal (S1, S2) decoder 1 (in this case, the postprocessor 5 implements artificial reverb). The system of figure 1 can be implemented (by the way, which will be described below), so that dual channel output signal of the postprocessor (OUT1 and OUT2) was a binaural sound signal output, applied reverb and that when playing headphones perceived tympanic membranes as the sound of «X» speakers (where X > 2 and X, as a rule, equal to 5), in any of a variety of provisions, including provisions to the listener and behind the listener.

Playback signals S1 and S2 (or OUT1 and OUT2), generated during the operation in binaural mode decoder of figure 1, can give the listener the feeling of the sound that comes from more than two (for example, five) «others» sources. At least some of these sources are virtual. More generally, for virtual surround sound, traditionally, is the use of modeling capabilities perceptions of sound (HRTF) to generate beeps (sometimes called virtual surround signals), which, when playing a pair of physical acoustic systems (e.g., speakers, located in front of the listener, or headphones) are perceived tympanic membranes of the listener as the sound of more than two sources (for example, speakers)located in any of a wide selection of provisions (as a rule, includes provisions behind the listener).

As noted above, the decoder MPS of figure 1, the acting binaural mode can be implemented for use reverb using artificial reverb, implemented by the postprocessor. Reverb can be configured to generate reverb in response to two-channel output signal (S1, S2) decoder 1 and apply reverb to the signals S1 and S2 with the purpose of generating reverberated two channel audio signal OUT1 and UT2. Reverb can be used as a post process reverb «stereo stereo to dual-channel signal S1, S2 of the decoder 1 so that all discrete channels defined in one of two flattened audio channels binaural sound signal output of the decoder 1 (for example, the left front and left to the surrounding channels, determined reduced channel S1), was used by the same impulse response reverb, and the same impulse response reverb applied to all discrete channels defined in the second of the two reduced binaural audio channels audio signal (for example, to the right front and right outside the channels defined reduced channel S2).

One of the traditional types of reverb contains a design known as the design is based on the scheme of feedback delay (FDN). In the course of work specified reverb applies reverb to the signal by creating a feedback signal delayed version of the same signal. The advantage of this design relative to other designs reverb is its ability to generate and apply multiple uncorrelated signals reverb to several input signals. This feature is used in mass-produced headphone, Dolby Mobile, which includes reverb, containing a design based on the FDN, and is suitable for use reverb on each channel 5 channel sound (containing the left front, right front, center, left and surrounding right surrounding channels) and filtering each reverberated channel using different pairs of filters from a set of five pairs of filters based modeling capabilities perceptions of sound (HRTF). This generates for each audio channel unique impulse response reverb.

headphone Dolby Mobile also operates in response to 2-channel audio input signal, generating two-channel «reverberated» audio output signal (two-channel virtual output signal ambient sound applied reverb). When reverberating sound signal output is playing a pair of headphones, he perceived the tympanic membranes listener as filtered HRTF reverberating sound of the five speakers, located in the left front, right front, center, left rear (surrounding) and right rear (surrounding) provisions. performs increasing mixing reduced dual-channel sound input (without using any parameters spatial labels, taken together with the input audio signal), generating five audio channels, subject to the step-mixing, applies reverb subjected to increasing mixing channels and performs stereo downmix signal five reverberated channels, generating dual reverberated output . Reverb for each channel, subjected to enhance the mixing which is different from other channels pair HRTF filter.

In published application in the U.S. patent No. 2008/0071549 A1, published on 20 March 2008, describes another traditional system for the use of reverb certain form reduced to the input sound during decoding reduced signal to generate the signals for the different channels. In this reference describes the decoder, which converts the reduced input signal in the time domain into the region of the QMF, applies to reduced signal M(t,f) in the field QMF reverb some form, regulates phase reverb, generating an option reverb to improve the mixing of each channel is defined from the final signal (for example, to generate the reverberation parameter L reverb (t,f) to enhance the mixing of the left channel and reverberation parameter R reverb (t,f) - to improve the mixing of the right channel, certain of reduced signal M(t,f)). Reduced signal together with the parameters of the spatial labels (for example, with ICC parameter that points to a correlation between the left and right components reduced signal and the parameters of the phase difference between the channels IPDL IPD R ). The parameters of the spatial label is used to generate reverb parameters (for example, L reverb (t,f) and R reverb (t,f)). If the label ICC indicates a greater correlation between the left and right components channels reduced signal for reduced signal M(t,f) is generated reverb smaller size, and reverb greater value is generated from reduced signal, if the ICC indicates a lower correlation between the left and right components channels reduced signal, and, obviously, the phase of each of the parameters correlation is governed (in block 206 and 208) in response phase, specify the corresponding label IPD. However, reverb is used only as in parametric stereo decoder (synthesis of mono-stereo), where for the reconstruction of correlation between the left and right channels used signal (which is orthogonal M(t,f)), the link does not offer separate definitions (or generate) characterized the reverb to apply to each discrete channel subjected to improve the mix mixed audio signal is determined from the consolidated sound signal M(t,f), or to each linear combination of many linear combinations of the values of the individual channels subjected to the step-mixing mixed audio signal is determined from the consolidated sound signal, for each discrete channel subjected to improve the mix mixed sound or each of these linear combinations.

The author of the invention took note that you may need a separate definition (and generate) different signals reverb for each of the discrete channels subjected to improve the mix mixed beep determined from reduced the audio signal from each of the discrete channels mixed output signal subjected to the step-mixing, or the definition and generate different signals reverb to (and from) every linear combination of the many combinations of values in the specified discrete channels. The author of the invention also took into account that at the specified separate definition signals reverb for individual channels mixed output signal subjected to increasing mixing (or linear combinations of the values specified channels), reverb, having different pulse response reverb, may apply to the channel mixed signal output subjected to increasing mixing (or linear combinations).

Up to the present invention, the parameters of the spatial labels, taken together with flattened by a sound signal is not used for generation of discrete channels mixed output signal subjected to improve the mixing of reduced sound signal (for example, in the area of the QMF, where reduced sound signal is a sound signal, encoded MPS) or linear combinations of its values, and to generate reverb from each of the specified channel mixed signal subjected to increasing mixing (or their linear combinations), separately for the purpose of application to the specified channel mixed sound signal subjected to increasing mixing (or their linear combination). Also existed reverberated channel mixed signal subjected to improve the mix that would be generated in this way and was , generating kept reverberating sound signal of reduced input audio signal.

SHORT DESCRIPTION OF THE INVENTION

In one of the classes of the embodiments of the invention, the invention is a method of application of reverb to M-channel reduced input sound indicating X separate audio channels, where X is a number greater than M In these cases the invention, the method includes the following steps:

For example, in one case, where M=2, X=5, Y=4, the input signal is a sequence of values of L(t), R(t)identifying signals of five separate channels L front R front , C, L sur R sur . Each signal five separate channels is a sequence of values

where W - matrix with increasing mixing MPEG Surround, having the form:

and four signals reverberated channels represent signals (g lf w 11 )L+(g lf w 12 )R (g, w rf 21 )L+(g rf 22 w )R (g ls w 11 )L+(g ls w 12 )R and (g rs 21 w +w 31 )L+(g rs w 22 +32 w )R can be represented as follows:

In some embodiments of the invention, where the input signal is an M-channel mixed signal MPEG Surround ("MPS"), phases (a) and (b) are carried out in the field QMF, and the parameters of the spatial labels are taken together with the input signal. For example, the parameters of the spatial labels may constitute or include the parameters of the difference in levels of channels (CLD) and/or factor parameters prediction channel (CPC)relating to the type, which is a part of traditional bitstream MPS. If the input signal is a reduced signal MPS in the time domain, the invention, as a rule, includes the stage of transformation of this signal from the staging area into the area QMF with the purpose of generating a frequency components region QMF and implementation phases (a) and (b) in the field QMF on these frequency components.

Optional method also includes the stage of generation of the N-channel summarized version Y signal reverberated channel (including each signal of the channel that was previously reverb, and, if they occur, each signal channel to which reverb was not used), for example, by coding reverberated channel in the form of the N-channel reduced signal MPS.

In a typical variants of the method of invention reduced input signal is a two-channel mixed signal MPEG Surround ("MPS"), pointing to five separate audio channels (left front, right front, center, left and surrounding right surrounding channels), and reverb is determined by different pulse characteristics reverb applied to at least some of these five channels, resulting in improved quality surround sound.

Preferably, the way the invention also includes the application phase to the signals reverberated channel respective functions modeling perceptions of sound (HRTF) by filtering signals reverb channels in the HRTF filter. Functions HRTF are applied in order to create conditions for perception of the listener reverb used in accordance with the invention, as sounding more natural.

Other features of the invention represent reverb configured (for example, programmed to perform any variant of the method of invention, , including specified reverb, decoder (for example, the decoder MPS), including specified reverb, and computer software media (disk)that contains the software code used to implement any of the variants of the method of invention.

SHORT DESCRIPTION OF GRAPHIC MATERIALS

Figure 1 - block diagram of a system of traditional decoder MPEG Surround.

Figure 2 - block diagram of reverb (100) based on the feedback delay (FDN) with multiple inputs and multiple outputs, which can be implemented in accordance with one of the embodiments of the present invention.

Figure 3 is a block diagram of the system of reverb which includes reverb 100 of figure 2, the traditional processor 102 MPS, filter 99 conversion from the staging area into the area QMF that converts multichannel input in the area QMF for the purpose of processing in reverb and 100 processor, 102, and filter 101 conversion from the field of the QMF, the temporary area dedicated to converting the combined output signal reverb 100 processor, 102, the temporary area.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS OF THE INVENTION

Technologically feasible many embodiments of the present invention. From this disclosure middle specialists in this field will be clear how to implement them. Options for implementation of the system of the invention, the way of the invention and the media will be described with reference to figure 2 and 3.

In one of the classes of options for the implementation of the invention is a method of application of reverb to M-channel reduced incoming audio signal, indicating X separate audio channels, where X is a number greater than M, and the system is configured to run method. In these cases the implementation of the method includes the following steps:

(b) individual use of reverb to each of at least two signals reverberated channel (for example, in the field QMF) by creating a feedback delayed version of the corresponding signal reverberated channel for each of the signals reverberated channel, and, therefore, the generation Y signal reverberated channel. Preferably, reverb, applied to at least one of the signals reverberated channel, has the impulse response reverb, which differs from the impulse response reverb applied to at least one signal reverberated channel. In some variants of implementation of the X=Y, but other variants of realization of X is not equal to y In some embodiments, the implementation of Y greater than M, and input at the stage of (a) is subject to the step-mixing in response to the parameters of the spatial labels with the purpose of generating Y signal reverberated channel. In other variants of realization of Y is equal to M or Y less M.

Figure 2 is a block diagram of reverb 100 based on the feedback delay (FDN) with multiple inputs and multiple outputs, namely subsystem apply reverb, representing a delay circuit feedback, which includes Y branches, and each branch configured for specific uses of reverb to differ one from signals reverberated channel. This reverb can be implemented described below for the implementation of this method. Reverb 100 Fig. 2 includes:

matrix 30 preliminary mixing (matrix «In»), which is a matrix 4 x M, connected and configured to receive and generate four discrete signals U1, U2, U3, U4 reverberated channel (corresponding to the branches of filing 1', 2', 3', 4' respectively) in response to M-channel reduced input audio signal, including channels IN1, IN2, ..., INM, which indicate five (X=5) separate audio channels, mixed signal, subjected to increasing mixing. Each signal reverberated channel at time t is a linear combination of a subset of values of X separate audio channels, mixed signal, subjected to increasing mixing, at the moment of time t. In the case when M is less than four, the matrix does increase mixing the input signal to generate signals reverberated channel. In a typical implementation, M is 2. Matrix 30 is also connected to receiving spatial labels that indicate (for example, describe the spatial image is reduced M-channel input and configured to generate four (Y=4) discrete signals of channels, mixed signal, subjected to increasing mixing, i.e. discrete signals U1, U2, U3, U4 reverberated channel, in response to the parameters of the spatial labels;

items 40, 41, 42, 43 summation related outputs matrix 30, to which are directed the signals U1, U2, U3, U4 reverberated channel. Item 40 configured for summation of the output element of strengthening g1 (i.e. for use feedback from the output element g1 gain) signal U1 reverberated channel. Item 41 configured for summation of the output element g2 amplification signal U2 reverberated channel. Item 42 configured for summation of the output element g3 amplification signal U3 reverberated channel. Item 43 configured for summation of the output element g4 amplification signal U4 reverberated channel;

the 32 matrix (matrix «And») scattering, which is connected to reception of output signals items 40, 41, 42, 43 summation. Matrix 32 preferably is a unitary matrix 4 x 4 configured for directions filtered version of the output signal of each item 40, 41, 42, 43 summation to the appropriate one of the delay lines, where 0 < = k-1 < 3, and, preferably, to ensure maximum , is completely filled matrix. Line z-M1 , z-M2 , z-M3 and z-M4 delay in figure 2 are marked, respectively, as line 50, 51, 52, 53 delay;

gain elements, gk, where 0 < = k-1 < 3 that apply the gain to output delay lines and thus provide damping factors, designed to control the decay time reverb applied to each channel of the mixed sound signal subjected to increasing mixing. Each element gk gain as a rule, is combined with the low pass filter. In some embodiments, the implementation of elements of the increasing use different preset gain of different bands QMF. Signals R1, R2, R3, R4 reverberated channel are directed, respectively, to the outputs of the elements g1, g2, g3, g4 gain; and

matrix 34 (matrix «With») , which is the matrix N x 4 connected and configured for down-mixing and/or increasing mixing (and, optionally, to perform other operations filtering) signals reverberated channel R1, R2, R3, R4, directed to the outputs of the elements gk strengthening, in response to at least a subset (for example, all or some) of the parameters of the spatial labels aimed at matrix, 30, and, thus, to generate the N-channel kept reverberating sound signal output in the area of the QMF, which includes channels S1, S2, ..., SN. Some variants of option implementation of figure 2, the matrix 34 is a constant matrix, factors which do not change over time in response to any of the parameters of the spatial labels.

In some variations of the variants of the invention of figure 2, according to the invention, contains Y reverberated channels (where Y is less than or more than four), the matrix 30 preliminary mixing configured to generate Y discrete signals reverberated channel in response to reduced M-channel input and the parameters of the spatial labels, matrix 32 scattering is replaced by the matrix Y x Y, and the invention includes Y delay lines.

For example, in the case when Y=M=2, reduced input points to five channels mixed sound signal subjected to increasing mixing (X=5): the left front, right front, center, left and surrounding right surrounding channels. According to the present invention, in response to the parameters of the spatial labels indicating the spatial image of reduced input signal, the matrix pre-fader (variation matrix 30 figure 2) generates two discrete signal reverberated channel (for example, in the field of quadrature mirror filter, or «QMF»), one for mixed audio signal to front channels, the second - for the mixed sound signal of the surrounding canals. Reverb, having briefly the characteristic decay, generated from (and used to) one beep reverberating channel, and reverb, with a long characteristics of attenuation, generated from (and used to) a second signal reverberated channel (for example, to simulate premises with acoustics type LEDE ).

Coming back to figure 2, the postprocessor 36, not necessarily connected to the outputs matrix 34 and acts, performing post-kept reverberating output signal S1, S2, ..., SN matrix 34, with the purpose of generating an N-channel audio output mode subjected postprocessing and containing channels OUT1, UT2, ..., OUTN. As a rule, N=2, then the system of figure 2 displays binaural kept reverberating sound signal S1, S2 and/or binaural kept reverberating sound signal output OUT, UT2 subjected to post-processing.

For example, the output matrix 34 in some implementations, the system of figure 2 is a binaural virtual signal ambient sound when playing headphones is perceived by the listener as the sound emitted from the left ("L"), Central ("C") and right (R) front sources (e.g., left, center and right physical acoustic systems located in front of the listener) and from the left surround (LS) and right surround (RS) surround sources (for example, the left and right physical speakers placed behind the listener).

Some variants of the system in Fig. 2 matrix 34 is skipped, and reverb, according to the invention displays Y-channel reverberating sound signal (for example, reverberating sound signal subjected to increasing mixing) in response to M-channel reduced input audio signal. In other variations matrix 34 is an identity matrix. In other embodiments, the system contains Y channels mixed sound signal subjected to increasing mixing (where Y is the number four), and matrix 34 is the matrix N X Y (for example, Y=7).

Despite the fact that the system of figure 2 contains four reverberated channel and four lines of delay, variations in the system (and other options for reverb, according to the invention) implement a more or less four the number of reverberated channels. As a rule, reverb, according to the invention, includes one line delay of one channel reverb.

In realization of figure 2, where the input signal is an M-channel mixed signal MPEG Surround ("MPS"), input is directed to the input matrix 30, contains signals IN1(t,f), IN2(t,f), ..., INM(t,f) in the field QMF and the system of figure 2 performs processing (for example, in the matrix 30) and the application to them of reverb in the field QMF. In such implementations parameters spatial labels sent to the matrix 30, as a rule, are the parameters of the difference in levels of channels (CLD), and/or factor parameters prediction channel (CPC), and/or parameters interchannel coherence (ICC)relating to the type, which is a part of traditional bitstream MPS.

For example, because the input signal system on figure 3 is a reduced sound signal MPS in the time domain, including M channels I1(t), I2(t), ..., IM(t), the system on 3 includes filter 99 dedicated to converting the specified time domain signal in signal in the area QMF. More precisely, the system in figure 3 contains reverb 100 (appropriate and possible, identical to the reverb 100 of figure 2), traditional processor MPS 102, filter 99 conversion from the staging area into the area QMF connected and configured to convert each of the input channels I1(t), I2(t), ..., IM(t) in the time domain in the region QMF (i.e. the sequence of frequency components region QMF)intended for processing in reverb and 100 traditional processing of the processor 102. The system in figure 3 also includes filter 101 conversion from the field QMF the temporary area is connected and configured to convert N-channel combined output signal reverb 100 processor, 102, the temporary area.

More precisely, the filter 99 converts signals I1(t), I2(t), ..., IM(t) in the time domain, respectively, in the signals IN1(t,f), IN2(t,f), ..., INM(t,f) in the field QMF, which are sent to the reverb 100 and processor 102. Each of the N-channel output signal processor 102 combined (in summer) with a corresponding output signal reverberated channel reverb 100 (one of the channels S1, S2,..., SN, shown in figure 2, or one of the channels OUT1, OUT2, ..., OUTN, shown in figure 2, if the reverb 100 figure 3 also includes the postprocessor 36, shown in figure 2). Filter 101 figure 3 converts combined (reverberated) output signal reverb 100 processor, 102 (N sequence of frequency components S1'(t,f), S2'(t,f), ..., SN'(t,f)) in the area QMF in signals S1'(t), s'(t), ..., SN'(t) in the time domain.

In a typical variants of implementation of the present invention reduced input signal is a two-channel mixed signal MPS, pointing to five separate audio channels (left front, right front, center, left surrounding, right surrounding channels), and reverb, as defined different pulse characteristics reverb is applied to each of these five channels that lead to better quality surround sound.

If the coefficients of the matrix 30 preliminary mixing (Y x M matrix, which in the case of Y=4 and M=2 is an matrix 4 x 2) are constant coefficients (unchanging in time coefficients, which are determined in response to the parameters of the spatial labels), and the coefficients of the matrix 34 (N X Y matrix C, which in the case of Y=4 and N=2 is a matrix 2 x 4) are constant coefficients, the system of figure 2 is not able to generate and apply individual reverb with separate pulse characteristics for different channels in reduced consumption sound signal, designated M-channel, reduced, coded MPS input reverb (for example, in response to coded MPS, M-channel, reduced signal IN1(t,f), IN2(t,f),..., INM(t,f)) in the area QMF. Consider an example where M=2, Y=4 and N=2, and the matrix b and C of figure 2 (also noted in figure 2 as matrix 32 and 34), substituted, respectively, constant matrices 4 x 2 and 2 x 4 with the following constant coefficients:

In this example, the coefficients are constant matrices b and C may not be changed depending on time and in response to the parameters of the spatial labels indicating reduced input sound signal and the system of figure 2, modified in this way, can function in the traditional mode reverb «stereo - stereo». This traditional mode reverb reverb has the same impulse response reverb applied to each separate channel in reduced consumption sound signal (i.e. the contents of the front left channel in reduced consumption sound signal takes reverberation, which has the same impulse response, and the contents of the right front channel in reduced consumption of sound signal).

However, applying the process of reverb in the field QMF in response to the parameters of the difference in levels of channels (CLD), the coefficient prediction channel (CPC) and/or parameters interchannel coherence (ICC), which is available as part of the bitstream MPS (and/or in response to other parameters of the spatial labels), in accordance with the invention, the system of figure 2 can generate and apply to each reverberated channel, determined reduced input signal system, reverb, with separate reverb characteristics for each of reverberated channels. In a typical application, a smaller reverb, according to the invention applied to the Central channel (for clearer voice playback/dialogue), than at least one of reverberated to the channel so that the impulse response reverb, applied to each of these reverberated channels, differ. In this application (and other applications) switching characteristics reverb applied to different reverberated channels, not based on different traces of channels to the matrix 30, but instead are simply different scaling factors used by the matrix 30 preliminary mixing or matrix 34 (and/or, at least, another element of the system to different reverberatory channels.

For example, one implementation of a system of figure 2 configured for applications reverb to stereo encoded MPS, mixed and mixed the audio signal in the area QMF of the five channels mixed sound signal subjected to enhance the mixing matrix 30 is a matrix 4 x 2, containing time-varying coefficients that depend on the current values of the coefficients w ij , where i ranges from 1 to 3, and j is in the range from 1 to 2.

In this illustrative implementation of M=2, X=5 & Y=4 input signal is a sequence of pairs IN1(t,f)=L(t) and IN2(t,f)=R(t) values in the field QMF indicating the sequence of the values of signals L front R front , C, L sur R sur five separate channels. Each signal five separate channels is a sequence of values

where W - matrix with increasing mixing MPEG Surround in the form:

In this example, the coefficients w ij will be updated in response to the current values of traditional CPC-settings CPC_1 and CPC_2 and traditional ICC parameter ICC_TTT (parameter interchannel coherence increase mixer «two to three», or «TTT», alleged during encoding reduced input):

In addition, when using the traditional parameters CLD for the front left/surround channels (CLD lf_ls ) and right front/surround channels (CLD rf_rs ) time-varying coefficients matrix 30 will also depend on the following four variables over time values of the coefficients of amplification, where CLD lf_ls - current value of the front left/surround setting CLD, and CLD rf_rs - the current value of the right-front/surround parameter CLD:

Then the time-varying coefficients matrix 30:

Thus, for illustrative implementation of the output signals of the four reverberatory channels matrix are 30 U1=(g lf w 11 )L+(g lf w 12 )R, U2=(g, w rf 21 )L+g(w rf 22 )R, U3=(g ls w 11 )L+(g ls w 12 )R, U4=(g rs 21 w +w 31 )L+(g rs 22 w +w 32 )R. Thus, matrix multiplication is performed by the matrix 30 (with the ratios shown in equation 3), can be represented as follows:

This matrix multiplication is equivalent to raising the mixing to the five signals for the different channels (via matrix increasing mixing MPEG Surround W, defined above) with further lowering to mix these five signals up to four signals reverberated channel through matrix In 0 .

In one of the variations of implementation matrix 30, with the ratios shown in equation 3, matrix 30 is implemented with the following factors:

where K LF , K RF

, K C , LS K , K RS - fixed gain values reverb for different channels, and g lf , g rf

, g c , g ls , g rs and w 11 w 32 , respectively, similar to the coefficients in equations 2 and 1A. As a rule, four fixed gain values reverb largely equal to each other except for the coefficient K c , which, as a rule, has a somewhat lower value than other factors (size, a few decibels less than the value of other factors) in order to the Central channel was used less reverb (for example, for more dry speech/dialogue).

Matrix 30, implemented with coefficients from equation 4, equivalent to the product of the matrix increases mixing MPEG Surround W, defined above, and the following matrix To 0 :

Alternatively, the coefficients of the matrix 30 are defined differently in response to the available options spatial labels. For example, some versions of the invention, the coefficients of the matrix 30 are determined in response to the available options spatial labels MPS, leading to the implementation of the matrix 30 increase mixer TTT, which operates in the mode of beyond predictions (for example, in energy, in the presence or in the absence of the subtraction of the centre). This approach can be implemented in a way that will be understandable to the average experts is aware of this description, when using the well-known formula increases for mixing of cases that are described in the MPEG standard (ISO/IEC 23003-1:2007).

In one of the implementations of the system of figure 2 configured for applications reverb to encoded MPS, single-channel (), mixed and mixed the audio signal in the area QMF of four channels mixed sound signal subjected to enhance the mixing matrix 30 is a matrix 4 x 1, containing time-varying coefficients:

where the coefficients, which gains derived from CLD-settings CLD lf_ls , CLD rf_rs , CLD c_lf , CLD l_r , which is available as part of the traditional bitstream MPS.

In variations of the system of figure 2 and other variants of implementation of reverb, according to the invention, discrete reverberated channels (for example, channels mixed sound signal, subjected to the step-up mix) extracted from the reduced input and traced to individual branches delay reverb any of the many different ways. In different variants of implementation of reverb, according to the invention to increase mixing reduced input signal there are other options spatial labels (for example, involves management weighing channels). For example, in some embodiments of the invention to determine the coefficients of the matrix a preliminary mixing and, thus, control levels reverb parameters are used as the ICC (available as part of the traditional bitstream MPS)that describe the diffusion of the front and rear channels.

Preferably, the way the invention also includes the application phase to the signals reverberated channels corresponding functions modeling perceptions of sound (HRTF) by filtering signals reverberated channels, HRTF filter. For example, the matrix 34 system of figure 2, preferably implemented as a filter HRTF that uses these functions HRTF to reverberated channels R1, R2, R3 and R4, and performs the same algorithm down-mixing on reverberated channels R1, R2, R3, R4. This implementation matrix 34 can usually perform the same filtering as matrix 5 x 4 and subsequent matrix 2 x 5 where the matrix 5 x 4 generates five virtual signals reverberated channel (left front, right front, center, left surround and right surround channels) in response to four output signals reverberated channels R1 to R4 of reinforcing elements g1, g2, g3 and g4, and matrix 2 x 5 applies appropriate function HRTF to each designated signal virtual reverberated channel and performs stereo downmix obtained in the result of five signals of channels, generating dual kept reverberating output signal. However, as a rule, the matrix 34 can be implemented as a single matrix 2 x 4, which performs the functions described individual matrices 5 x 4 and 2 x 5. Functions HRTF are applied in order to create conditions for perception of the listener reverb used in accordance with the invention, as sounding more natural. The HRTF filter can, as a rule, to perform matrix multiplication for each individual band QMF by matrices with elements.

In some embodiments of the invention, the signals reverberated channel generated from the encoded MPS reduced input in the field QMF, filtered relevant functions HRTF as described below. In these cases the implementation of the HRTF in parametric QMF mainly consist of the values in the left and the right, gain settings and parameter values of the phase difference between channels (IPD), which characterize the reduced input. Parameters IPD, optional, ignored with the aim of reducing complexity. Assuming that the parameters IPD ignored functions HRTF represent the value of the constant gain (four values of gain for each of the left and right channels respectively): g HRTF_lf_L , g HRTF_rf_L , g HRTF_ls_L , g HRTF_rs_L , g HRTF_lf_R , g HRTF_rf_R , g HRTF_ls_R , g HRTF_rs_R . Thus, the functions of the HRTF can be applied to the signal R1, R2, R3, R4 reverberated channel of figure 2, by implementing a matrix 34 containing the following factors:

Some of these preferred implementation of reverb, according to the invention that are variations of the system of figure 2, which can be configured for use fractional delay (in at least one reverberated channel), as well as integer delay sampling. For example, in one such element implementations of fractional delay is connected with each of reverberated channels sequentially delay line, which uses integer delay integer equal to the number of sampling periods (for example, each element of fractional delay consistently placed after or before one of the lines 50, 51, 52, 53 delay of figure 2). Fractional delay can be approximated by a phase shift (by multiplication by a complex number with module unit) in each band QMF, which corresponds to the proportion of the sampling period: f=t/T, where f is the fraction of delay, t - required delay for the band QMF, T is the sampling period for the band QMF. It is well known how to apply fractional delay in the context of the use of reverb in the field QMF (see, for example, report of J. Engdegard and others, "Synthetic Ambience in Parametric Stereo Coding," presented at the 116th Convention society of engineers-acousticians, Berlin, Germany, may 8-11, 2004, 12 S., and also patents of USA №7487097 issued by J. Engdegard and other 3 February 2009).

Some of the mentioned above preferred implementations of reverb, according to the invention represent variations of the system of figure 2, which are configured for use reverb differently to different frequency bands sound data in at least one riverbelleroom channel to reduce the complexity of realization of reverb. For example, in some implementations where the input audio data IN1-INM represent data MPS in the field QMF, and the use of reverb is in the area of the QMF, reverb is applied in various ways to the following four frequency bands sound data in each reverberated channel:

0-3 kHz (or 0-2.4 kHz): in this band reverb is applied in accordance with the above option implementation of figure 2 with the sensor of 30, which is implemented with coefficients in equation 4;

3-8 kHz (or 2.4 to 8 kHz): in this band reverb applies only arithmetic, not containing complex quantities. For example, this can be done using methods of arithmetic, not containing a complex of values described in the published international application № WO 2007/031171 A1, March 22, 2007 This application describes the 64-band filters block QMF, in which complex values of eight most low frequency bands represent the processed audio data and processed values only, not containing complex quantities, the top 56 of the bands sound data. One of these eight lowest frequencies can be used as a buffer strip comprehensive QMF, and, thus, arithmetic calculations for complex values are only for seven of the eight most low frequency bands QMF (thus, reverb used in this relatively low range as in the above is the embodiment of the invention of figure 2 with the matrix 30, implemented with coefficients in equation 4), and for the other 56 of the bands QMF calculations are performed for values that do not contain complex quantities, where the transition region between calculations of complex quantities and values, not containing complex quantities is located on the frequency (7 x 44.1 kHz)/(64 x 2), which is approximately equal to 2.4 kHz. In this exemplary embodiment of the invention reverb used in relatively high frequency range as in the above is the embodiment of the invention of figure 2, but using a simplified implementation of the matrix pre 30 mixing intended only for computing values, not containing complex quantities. Reverb is applied at relatively low frequencies (below 2.4 kHz) as well as in the embodiment of figure 2, for example, with the matrix 30, implemented with coefficients in equation 4;

8-15 kHz: in this band reverb is applied by way of a simple delay. For example, reverb is applied in a way that is similar to the method used in the embodiment of figure 2, but with two reverberated channels delay line and the low pass filter in each channel reverb, with the passage of matrix elements 32 and 34, with the simple implementation matrix 30 preliminary mixing in the form of a matrix of 2 x 2 (for example, to use less reverb to the Central channel than other channels) and in the absence of feedback from sites along the canals of reverb to the outputs matrix preliminary mixing. Two branches of the delay can just lead, respectively, to the left and right outputs or can switch so that the echoes from the left front (Lf) and left surround (Ls) channels reached the right output channel, and the echoes of the right front (Rf) and right surround (Rs) channels reached the left output channel. Matrix pre-mix 2 x 2 can contain the following factors:

where symbols are defined similarly to the characters given by equation (4) above; and 15-22,05 kHz to this band reverb is not applied.

In some variations disclosed in this description of embodiments (for example, options implementation of figure 2), the system according to the invention applies reverb to M-channel reduced input sound indicating X separate audio channels, where X is a number greater than M, which consists in generating Y discrete signals reverberated channel in response to reduced signal, but not in response to the parameters of the spatial labels. In these variations, the system applies separately reverb to each of at least two signals reverberatory channels in response to the parameters of the spatial labels indicating the spatial image of reduced input and, thus, generates Y signal reverberated channels. For example, some of these variations of the coefficients of the matrix a preliminary mixing (for example, variations in the matrix on 30 figure 2) in response to the parameters of the spatial labels are not defined, but at least one of the matrices scattering (for example, a variation matrix 32 of figure 2), stage gain (for example, the variation of the milestone, including elements g1-gk of figure 2) and the matrix (for example, a variation matrix 34 of figure 2) applies to the signals reverberated channel in a way that is determined by the parameters of the spatial labels indicating the spatial image of reduced input signal, in order to apply reverb to at least one of the two signals reverberatory channels.

In some cases, implementation, reverb, according to the invention constitutes or includes universal processor, connected to the reception or generate input data indicating M-channel reduced input audio signal, and programmed by software (or firmware) and/or otherwise configured (for example, in response to control data) to perform any of many different operations on the input data, including variant of the method of invention.

The specified universal processor, as a rule, can connect to the input device (e.g. mouse and/or keyboard), memory and display device. For example, in figure 3 can be implemented in a generic processor where the input signals I1(t), I2(t), ..., IM(t) are the input data indicating M channels for reduced sound data, and the output signals S1(t), S2(t), ..., SN(t)are the output indicate N channels kept reverberating sound signal. Traditional digital-to-analog Converter (DAC) can act on this output, generating analog version output signals for playback acoustic systems (for example, a pair of headphones).

Although this disclosure describes the specific embodiments of the present invention and application of the invention, the average experts in this field should understand that there are many possible variations are described here, options implementation and applications of the invention without derogating from the scope of the invention described and declared in this disclosure. You should understand that despite the fact that they were shown and described some form of the invention, the invention is not limited to specific cases described the invention, or describe specific ways.

1. Method of application reverb to M-channel reduced input sound indicating X separate audio channels, where X is a number greater than M, while the method includes the following steps: (a) in response to the parameters of the spatial labels indicating the spatial image of reduced input, generate Y discrete signals reverberated channel from M-channel summarizes the input sound. where each signal reverberated channel at time t is a linear combination of at least the subset of values X separate audio channels at time t; where Y discrete signals reverberated channel generated using the matrix (30) pre-mix containing the odds in time determined in response to the spatial parameters of the labels; (b) separately apply reverb to each of signals reverberated channel and such by the way, generate Y signal reverberated channel, where reverb apply separately to each of the signals reverberated channel by creating feedback delayed version of the corresponding signal reverberated channel for each of the signals reverberated channel; and (C) generate N-channel reverberating sound signal from the Y signal reverberated channel using the matrix (34) .

3. The method according to claim 1, characterized in that the input signal is an M-channel mixed signal MPEG Surround, and the parameters of the spatial labels include at least one of the parameters of a difference of levels, parameters coefficient prediction of channel and settings interchannel coherence.

4. The method according to claim 3, wherein the parameters of the spatial labels include the difference between the levels of channels factor parameters prediction of channel and settings interchannel coherence.

5. The method according to claim 1, characterized in that the input signal is a reduced signal MPEG Surround in the field QMF, which includes M sequences frequency components region QMF, and where each of these steps (a) and (b) is in the area QMF.

6. The method according to claim 5, wherein the parameters of the spatial labels include at least some of the parameters of a difference of levels, parameters coefficient prediction of channel and settings interchannel coherence.

7. The method according to claim 5, wherein the parameters of the spatial labels include the difference between the levels of channels factor parameters prediction of channel and settings interchannel coherence.

8. The method according to claim 1, characterized in that the input signal is a reduced signal MPEG Surround in the time domain, and also includes the following stages: stage before (a) to convert the reduced signal MPEG Surround temporary workspaces QMF and, thus, generate M sequences frequency components region QMF; and where each of these steps (a) and (b) perform in the field QMF.

9. The method of claim 8, wherein the matrix (34) performs stereo downmix Y signal reverberated channel.

10. The method according to claim 1, which also includes the application phase to the signals reverberated channel respective functions simulation of sound perception by filtering signals reverberated channels in the filter functions of modeling of sound perception.

11. The method according to claim 1, wherein Y M. more

12. Reverb, configured for use reverb to M-channel reduced input sound indicating X separate audio channels, where X is a number greater than M, where specified reverb contains: the first subsystem that is connected to the receive input signal and parameters spatial labels, which indicate the spatial image of the specified input signal, and is configured to generate Y discrete signals reverberated channel in response to the input signal, that is the application of a matrix (30) pre-mix containing the odds in time determined in response to the parameters of the spatial labels so that each signal reverberated channel at time t was represented by a linear combination of at least the subset of values X separate audio channels at time t; subsystem(40, 41, 42, 43, 32, 50, 51, 52, 53, 54) apply reverb, connected to the first subsystem and configured for use reverb separately to each of the signals reverberated channel and, thus, to generate a number of Y signal reverberated channel, where the subsystem use the reverb is a feedback delay, which includes Y branches, and each branch configured for specific uses of reverb to differ one from signals reverberated channel; and subsystem connected and configured to generate an N-channel kept reverberating sound signal from the Y signal reverberated channel by matrix (34) .

13. Reverb on p.12, wherein subsystem(40, 41, 42, 43, 32, 50, 51, 52, 53, 54) apply reverb configured for use reverb so that reverberation that applies to at least one of the signals reverberated channel, had the impulse response reverb, different from the impulse response reverb that applies to at least one of the signals reverberated channel.

14. Reverb indicated in paragraph 12, wherein the reduced input audio signal represents many M sequences frequency components region QMF, with the specified reverb also includes: filter (99) conversion of temporary workspaces QMF connected to receive reduced signal MPEG Surround in the time domain and is configured to generate the answer, M sequences frequency components region QMF, and where the subsystem increasing mixing enabled and configured to enhance the mixing of these M sequences frequency components region QMF in the field QMF.

15. Reverb indicated in paragraph 12, which also includes the filter features of simulation of sound perception, enabled and configured to use at least one function modeling of sound perception in each of the signals reverberated channel.