Cross product-enhanced, subband block-based harmonic transposition

FIELD: physics, acoustics.

SUBSTANCE: invention relates to systems for encoding audio signal sources. Provided is subband block-based harmonic transposition, where the time block of complex discrete values of subbands is processed by common phase modification. Superposition of multiple modified discrete values yields the resultant effect of limiting undesirable cross products, making it possible to use coarser frequency resolution and/or lesser degree of oversampling. In one embodiment, the invention further includes a window function suitable for use with cross product-enhanced, subband block-based HFR. A hardware embodiment may include an analysing filter unit (101), a control data-configurable subband processing module (102) and a synthesising filter unit (103).

EFFECT: efficient implementation of high-frequency reconstruction (HFR) through enhancement with cross products, where a new component with frequency QΩ+rΩ0 is generated based on existing components with frequencies Ω and Ω+Ω0.

63 cl, 9 dwg

 

AREA of TECHNOLOGY

The present invention relates to systems for encoding the audio signal, applying the harmonic conversion method for high frequency reconstruction (HFR) digital effects processors, such as built-in that generate harmonic distortion to add brightness to the processed signal, and the devices for stretching the timeline, which increase the duration of the signal with preservation of spectral composition.

Background of the INVENTION

In document WO 98/57436 the concept of transformation was established as a way of recreating the high-frequency band of the low-frequency band of the audio signal. By using this concept it is possible to obtain significant savings in bit-rate data transmission by encoding the audio signal. In the coding system of the audio signal based on the HFR signal with low-frequency bandwidth is transmitted in the reference encoder waveform, and the higher frequencies are regenerated with the use of conversion and additional supplementary information transmitted with a very low bit-rate data transmission, which describes the target spectrum shape on the side of the decoder. For low bit-rate, when the bandwidth of the basic coded signal is� narrow, increasingly the importance of the reconstruction of the high-frequency band with a pleasant perception characteristics. Harmonic transform defined in WO No. 98/57436, runs very well for complex musical material in a situation with a low transition frequency. The principle of harmonic conversion is that a sinusoid with frequency ω is mapped to a sinusoid with a frequency of Qφω, where Qφmore 1 - an integer specifying the order of transformation. For comparison, HFR on the basis of the modulation signal with a single side band displays a sine wave with a frequency ω in a sinusoid with frequency ω+Δω, where Δω is a fixed frequency shift. For any given base signal with low bandwidth, the transformation of the SSB will be the artifact dissonant ringing.

To achieve the best possible quality sound, the ways of high-harmonic HFR at the present level of technology used to achieve the desired sound quality blocks of complex modulated filter with a very high resolution in frequency and a high degree of oversampling. High resolution is necessary in order to avoid unwanted intermodulation distortion arising from the nonlinear processing �Umm sine waves. With sufficient narrowness of the sub-bands of high-ways strive to ensure that each sub-band contained no more than a single sinusoid. A high degree of oversampling in time is necessary in order to avoid distortion due to insufficient sampling rate, and a certain degree of oversampling in frequency is necessary in order to avoid anticipatory echo for transient signals. The obvious disadvantage is that the computational complexity becomes very high.

Another well-known disadvantage associated with harmonic transformations, is shown for signals with a pronounced periodic structure. These signals are a superposition of harmonically related sinusoids with frequencies Ω, 2Ω, 3Ω..., where Ω is the fundamental frequency. At harmonic transform of order Qφweekend sinusoids have frequencies QφΩ 2QφΩ, 3QφΩ... that in case of Qφ1 is a strict subset of the desired total harmonic series. In relation to the resulting sound quality as a rule, will be considered as "spurious" the basic tone corresponding to the fundamental frequency converted QφΩ. Often harmonic conversion leads to a "metal" sound character encoded and decoded�wow the audio signal.

In document WO 2010/081892, which by reference is included in the present description, to refer to the solution described above to the problem of "spurious" main tone in a high-quality conversion method was developed cross-pieces. For a given partial or total transmitted information about the value of the fundamental frequency of the predominant harmonic parts of the signal that you are converting with high accuracy, a nonlinear modification of the sub-bands are complemented by nonlinear combinations of at least two different of the analyzed sub-bands, where the distance between the indices of the analyzed sub-bands associated with the fundamental frequency. As a result generated the missing harmonics of the converted output signal, which, however, comes at a significant computational cost.

BRIEF description of the INVENTION

In view of the above-described disadvantages of the available methods HFR the aim of the present invention is to provide a more efficient implementation of HFR, enhanced cross-pieces. In particular, the goal is the creation of this method that allows for the playback of audio signals with high accuracy under the condition of reducing the computational cost compared to the available access methods.

The present invention achieves, according to møn�Shea least one of these goals by creating devices and methods according to independent claims.

In the first aspect of the invention provides a system configured for generating stretched in time and/or converted by the frequency signal from the input signal. The system contains:

- block analyzing filters, configured to receive from the input signal to some amount of Y signals of the analyzed sub-bands, where each signal is analyzed sub-band comprises the number of analyzed complex-valued discrete values, each of which has a phase and amplitude;

- processing module of the sub-bands is configured to determine the signal synthesized sub-band based on the Y signals of the analyzed sub-bands using the conversion coefficient Q sub-bands and the stretch factor S subbands, where at least one of the coefficients Q and S is greater than one, wherein the processing module subbands includes:

extractor units configured for:

- (i) the formation of the Y frames consisting of L discrete input values, where each frame is extracted from the specified number of complex-valued discrete values in the signals of the analyzed sub-bands, and the frame length L is greater than 1; and

ii) applying the values jump� block of discrete values of h up to the specified number of analyzed discrete values before generating the next frame, consisting of L discrete input values, thereby generating a sequence of discrete input values;

module nonlinear frame processing configured to generate, on the basis of Y corresponding frames of input discrete values, formed in the extractor blocks of the frame of the processed discrete values by determining the phase and amplitude of each of the processed discrete values of the frame, where at least one of the processed discrete values:

- i) the phase of the processed discrete values based on the respective phases of the corresponding input digital values in each of the Y frames of discrete input values; and

ii) the amplitude of the processed discrete values based on the amplitude of the corresponding input digital values in each of the Y frames of discrete input values; and

module overlay and addition, is configured to determine the signal synthesized sub-band by superimposing and adding discrete values from a sequence of frames of the processed discrete values;

and

- block synthesizing filters configured to generate stretched in time and/or converted by the frequency signal from the signal synthesized sub-band.

With�system can operate with any positive integer value Y. However, it is valid when at least Y=2.

In a second aspect, the invention provides a method of generating stretched in time and/or converted by the frequency signal from the input signal. The method includes:

- receiving from the input signal to some amount of Y>2 signals of the analyzed sub-bands, where each signal is analyzed sub-band comprises the number of analyzed complex-valued discrete values, each of which has a phase and amplitude;

- formation of Y frames consisting of L discrete input values, where each frame is extracted from the specified number of analyzed complex-valued discrete values in the signals of the analyzed sub-bands, and the frame length L of more than 1;

- the use of magnitude jump from h block of discrete values to a specified number of analyzed discrete values before receiving the next frame consisting of L discrete input values, thereby generating a sequence of discrete input values;

- generation based on Y corresponding frames of input discrete values of the frame of the processed discrete values by determining the phase and amplitude of each of the processed discrete values of the frame, where at least one of the processed discrete values:

- phase milling�Ivanovo discrete values based on the respective phases of the corresponding input digital values in at least one of Y frames of the input of discrete values; and

- the amplitude of the processed discrete values based on the amplitude of the corresponding input digital values in each of the Y frames of the input discrete values;

- definition of signal synthesized sub-band by superimposing and adding discrete values from a sequence of frames of the processed discrete values; and

- generate stretched in time and/or converted by the frequency signal from the signal synthesized sub-band.

Here Y is an arbitrary integer greater than one. The system according to the first aspect acts for implementing the method, at least for Y=2.

The third aspect of the invention provides a computer software product that includes a computer readable storage medium (or storage media), in whose memory stores commands software designed to trigger the execution of a programmable computer, the method according to the second aspect.

The invention is based on the understanding that the General concept HFR, enhanced cross-pieces, will provide improved results when processed data that is organized into blocks, which consist of complex discrete values of sub-bands. Among other things, this makes possible the application to the disk�values of ethno-frame phase shift, that, as observed, in some situations weakens matching components. You may also adjust the amplitude, which can lead to similar beneficial effects. Implementation of enhanced cross-pieces HFR according to the invention includes a harmonic mapping of the unit of sub-bands, which may significantly weaken the matching components. Therefore, despite the high perceived quality can be used the filter unit (such as unit QMF filters) with coarser resolution in frequency and/or a lesser degree of oversampling. At processing block-based temporal subbands unit complex discrete values of sub-bands processed by the well-known modifications of the phases and the superposition of several modified discrete values when generating the output discrete values of sub-band gives the combined effect of suppression of combinational components that would otherwise arise when the input signal sub-band consists of multiple sine waves. The conversion based on sub-band processing block-based, has a much smaller computational complexity than the converters with high resolution, and for many signals reaches almost the same quality�STW.

For the purpose of this disclosure note that in embodiments where Y>2, the non-linear processing module uses as input Y "respective" frames of input discrete values in the sense that the frames are synchronous or nearly synchronous. For example, discrete values in the respective frames may refer to intervals of time, with considerable overlap in time between frames. The term "appropriate" is also used in relation to discrete values to indicate that they are synchronous or approximately are not. In addition, the term "frame" will be used interchangeably with the term "block". Accordingly, the magnitude of the jump of the block may be equal to the length of the frame (perhaps adjusted in relation to the downsampling, if applicable, is or may be less than the length of the frame (perhaps adjusted in relation to the downsampling, if it applies), and in this case, successive frames are blended in the sense that the discrete input value can belong to more than one frame. The system optionally generates each of the processed discrete value in the frame by setting its amplitude and phase based on phase and amplitude Y of the respective frames of the input dis�specific values; without derogating from the invention, the system may generate the phase and/or amplitude of some of the processed discrete values based on a smaller number of relevant input discrete values or on the basis of only one discrete input values.

In one of the embodiments of the invention, the unit of the analyzing filters is a block quadrature mirror filters (QMF), or block pseudo-QMF, with any number of links and points. For example, it can represent a 64-point block QMF. The unit of the analyzing filters can also be out of the classroom window discrete Fourier transforms or wavelet transforms. Mainly, the block synthesizing filters is consistent with the unit of the analyzing filters, being, respectively, the block of the inverse QMF, unit return pseudo-QMF, etc. Knows that such filter units can have a sufficiently coarse resolution in frequency and/or a relatively low degree of oversampling. In contrast to current prior art, the invention can be carried out with the use of these relatively simple components do not necessarily suffering from lowering the quality of the output; thus, these embodiments of the invention has an economic advantage over the current level of technical�I.

In one of the embodiments of the invention, the unit for analyzing filters surely one or more of the following statements:

- step analysis to a time period ΔtA;

- a rating of the analyzed frequency ΔƒA;

- block analyzing filters includes N>1 of the analyzed sub-bands, indexed by the index of the analyzed sub-band n=0, ..., N-1;

- analyzed sub-band is associated with one of the frequency bands of the input signal.

In one of the embodiments of the invention, for the block synthesizing filters surely one or more of the following statements:

- step synthesis of time - Δts;

- the spacing of the synthesized frequency Δƒs;

- block synthesizing filters includes M>1 of synthesized subbands indexed by index of synthesized sub-band m=0, ..., M-1;

- synthesized sub-band is associated with one of the frequency bands of the signal, stretched in time and/or converted by frequency.

In one of the embodiments of the invention, the non-linear processing module frame adapted to enter two frames (Y=2) for the purpose of generating one frame of the processed discrete values, and the processing module sub-bands includes a control module cross-processing is used to generate control data cross-processing. About�of redesa quantitative and/or qualitative characteristics of the processing of the subbands thus, the invention achieves flexibility and adaptability. The control data may define sub-bands (e.g., identified by indices), which differ in frequency from the fundamental frequency of the input signal. In other words, the indices identifying the sub-bands may differ by an integer, which is a private approximation of a specified fundamental frequency divided by the spacing of the analyzed frequencies. This will lead to pleasant from the psychoacoustic point of view of output, because the new spectral components are generated by harmonic conversion will be compatible with a number of natural harmonics.

In the further development of the preceding embodiment of the invention, the indices (input) and analyzed (output) of synthesized sub-bands are selected to satisfy the following equation (16). Appearing in this equation the parameter σ makes it applicable to both unevenly and evenly arranged blocks of filters. If the indices of the sub-bands are obtained as an approximation (e.g. least-squares) solution of the equation (16), a new spectral component obtained by harmonic conversion is likely to be compatible with a number of natural harmonics. Thus, HFR is likely to provide dost�faithful reconstruction of the original signal, from where it was resolved high-frequency content.

Further development of the preceding embodiment of the invention provides for a method of choosing the parameter r appearing in equation (16), and represent the transformation with cross pieces. For a given index m of the output sub-band each value of r transformation procedure will determine two indices n, n2the analyzed sub-bands. This further development of estimates of the amplitude of the two specified ranges for a certain number r of options and chooses the value that gives the maximization of the minimum of the two amplitudes of the analyzed sub-bands. This method of selection indexes can afford to avoid having to recreate sufficient magnitude amplitude by amplifying weak components of the input signal and the output can lead to low quality. In this regard, the amplitude of sub-bands can be calculated by a method which is known in itself, such as the square root of the squares of the input of discrete values, comprising a frame (block) or part of the frame. The sub-band amplitude can also be calculated as the amplitude of the Central, or near Central, discrete values in the frame. This calculation can create a simple but at the same time, an adequate number of�enny indicator of amplitude.

In the further development of the preceding embodiment of the invention synthesized subband can accept deposits from the events of harmonic conversion according as direct processing and processing based on the cross-pieces. In this regard, to determine whether to use a special ability to restore missing harmonics by processing based on the cross-pieces, can be used criteria for decision making. For example, this further development can adapt to abstain from the use of a single module cross-processing sub-bands in the case when either of the following conditions:

(a) the ratio of the amplitude of the Msmember of analyzed sub-band direct from the source, resulting in the synthesized sub-band, and at least the amplitude of the Mcin the optimal pair of members from a cross of the source, resulting in the synthesized sub-band, more a pre-defined constant;

(b) synthesized sub-band is already receiving a significant contribution from the direct processing module;

c) the fundamental frequency Ω0less than the frequency spacing unit of the analyzing filters Δƒ

In one of the embodiments of the invention, the invention includes lowering of discretiza�Oia (thinning) of the input signal. Moreover, one or more frames of input discrete values can be determined by downsampling complex-valued discrete values in the sub-band which may be performed by the extractor units.

In the further development of the preceding embodiment of the invention applicable downsampling coefficients satisfy the following equation (15). Equality to zero of both coefficients downsampling is not allowed since it corresponds to the trivial case. Equation (15) determines correlation coefficients downsampling D1, D2with the stretching factor S subbands and the conversion coefficient Q sub-bands, as well as with phase coefficients T1, T2appearing in the expression (13) for determining the phase of the processed discrete values. This ensures the consistency of the phase of the processed discrete values with other components of the input signal, adding a subject to which the processed discrete values.

In one of the embodiments of the invention, the frames of the processed discrete values prior to their imposition and undergo addition of window treatment. The module window treatment can be adapted to apply to the treated discrete�nd the values of the window function of finite length. Suitable window functions are listed in the attached claims.

The inventor realized that the cross-ways of the compositions disclosed in WO No. 2010/08892 initially not completely compatible with the methods of processing on the basis of the block of sub-bands. Although this method can satisfactorily be applied to one of the discrete values in the block, it can lead to artifacts in the overlay of the spectra, if it is directly extended to other discrete values of the block. To this end one of the embodiments of the invention applies a window function comprising a windowed discrete values, which, when weighted by complex weights and the bias on the magnitude of the surge are, in large measure, a constant sequence. The magnitude of the jump can be a product of size of the jump of h block at a stretching ratio of subbands S. the Use of these window functions mitigates the impact of artifacts overlay of the spectra. Alternatively or in addition, these window functions can also allow and such other measures to reduce artifacts, the phase sequence of the processed discrete values.

Preferably, the successive complex weighting coefficients, notoriamente to windowed discrete values to assess their condition, differ only by a fixed alternating phases. It is also preferred that the specified fixed alternating phase was proportional to the fundamental frequency of the input signal. The alternation of phases can also be proportionately subject to the application of the procedure of converting the cross-pieces and/or physical parameter conversion, and/or difference of the coefficients of the downsampling and/or step of the analysis time. The alternation of phases can be in the form of equation (21) at least in an approximate sense.

In one of the embodiments of the invention, the present invention makes it possible harmonic conversion, increased cross-pieces, by modifying the synthesizing window treatment in response to a parameter of the fundamental frequency.

In one of the embodiments of the invention, successive frames of the processed discrete values are added with a certain overlap. To perform a suitable overlay frames processed discrete values properly shifted by the value of the jump, which is a measure of the jump h block, multiplied by the stretch factor S subbands. Thus, if the imposition of consecutive frames of input discrete values is L-h, the imposition of consecutive kad�s processed discrete values can be S(L-h).

In one of the embodiments of the invention, the system according to the invention applies not only to generate the processed discrete values based on Y=2 input discrete values, but also on the basis of only Y=1 discrete values. Thus, the system can recover the missing harmonics not only through an approach based on cross works (as, for example, by equation (13)) but also through direct approach based on sub-bands (as, for example, by equation (5) or (11)). Preferably, the control module configured to control action of the system, including those which approach should be used to restore the particular missing harmonics.

In the further development of the preceding embodiment of the invention, the system also is adapted to generate the processed discrete values on the basis of more than three discrete values, i.e., for Y≥3. For example, the processed discrete value can be obtained by contributing to the processed discrete value from a few events harmonic conversion based on cross-sectional works, through several events direct processing of subbands, or by a combination of conversion with cross pieces and direct conversion. �demonstrated the possibility of adapting the method of conversion ensures efficient and multifunctional HFR. Accordingly, this embodiment of the invention operates to perform the method according to the second aspect of the invention, for Y=3, 4, 5, etc.

One of the embodiments of the invention are configured to determine the processed discrete values as a complex number having an amplitude which is an average of the respective amplitudes of respective input of discrete values. The specified average value can be a (weighted) arithmetic, the (weighted) geometric or the (weighted) harmonic mean of two or more discrete input values. In the case Y=2, the average is based on two complex input discrete values. Preferably, the amplitude of the processed discrete values is a geometric weighted average. More preferably, the geometric value, as shown in equation (13), is weighted using the parameters ρ and 1-ρ. Here the parameter ρ geometric weighting of the amplitudes is a real number that is inversely proportional to the ratio Q of the transform subbands. The parameter ρ can also be inversely proportional to the stretching factor S.

In one of the embodiments of the invention the system is adapted to detect�tion of the processed discrete values as a complex number, having a phase that is a linear combination of the respective phases of the corresponding input digital values in the frames of the input discrete values. In particular, the linear combination may include phase belonging to two discrete input values (Y=2). A linear combination of the two phases can apply a non-zero integer coefficients whose sum is equal to the stretching factor S multiplied by the conversion coefficient Q subbands. Optionally, the phase obtained by the specified linear combinations, further guided through a fixed parameter of phase correction. The phase of the processed discrete values can be in the form of equation (13).

In one of the embodiments of the invention, the extractor blocks (or similar stage in the method according to the invention) adapted to interpolate two or more of the analyzed discrete values of signals of the analyzed sub-band for the purpose of receiving one of the input discrete value, which will be included in the frame (block). This interpolation can make possible a lowering discretization of the input signal by a noninteger factor. Analyze of discrete values to be interpolated may be or may not be following each other.

� one of the embodiments of the invention, handling configuration of sub-bands may be controlled by control data delivered from outside of the module that performs the processing. The control data can relate to the instantaneous acoustic properties of the input signal. For example, the system may contain a section adapted to determine such instant the acoustic properties of the signal (dominant) frequency of the signal. Knowledge of the fundamental frequency provides leadership for the selection of the analyzed sub-bands, from which to retrieve the processed discrete values. Accordingly, the spacing of the analyzed sub-bands is proportional to the specified fundamental frequency of the input signal. Alternatively, the control data may also be delivered from outside the system, preferably by including an encoding format suitable for transmission as bitstream via digital communication network. In addition to the control data specified encoding format may contain information relating to the low-frequency signal components (e.g., components in Fig.701 in Fig.7). However, in the interest of saving bandwidth format, preferably, does not include complete information relating to high-frequency components (POS.702), which may be recovered according to the invention. The invention m�Jette, in particular, provide for a system of decoding module receiving the control data configured for receiving these control data, or included in a received bit stream, which also encodes the input signal, or the received as a separate signal or bit stream.

One of the embodiments of the invention provides a method for the effective implementation of the calculations, due to the method according to the invention. With this purpose, a hardware implementation may include prenormalization designed to change the scale of the amplitudes of respective discrete values in some of Y frames on which to base the frame of the processed discrete values. After this change the scale of the processed discrete value can be calculated as the (weighted) complex product of the input of discrete values, subject to scale, and may not be revised in the scale. The input discrete value that occurs in the novel as the ratio of the modified scale, usually re-appears as a coefficient with the same scale. With the exception, perhaps, of the parameter in the correction phase can estimate equation (13) as the product of the complex input of discrete values (possibly with a modified scale). It s�ing the computational advantage compared with processing the amplitude and phase of the processed discrete values separately.

In one of the embodiments of the invention the system is configured for the case Y=2, contains two extractor units adapted for forming each of them one frame of input digital values in parallel.

In the further development of the embodiments of the invention, representing Y≥3, the system may contain a number of processing blocks of subbands, each of which is configured to determine the intermediate signal synthesized sub-band using a different conversion factor of sub-bands and/or different stretching ratio of sub-bands, and/or conversion method, characterized in that it is based on cross or direct. For parallel operation of the processing modules of the sub-bands can be parallel. In this embodiment, the system also may include a merge module, located behind the processing modules of the sub-bands and before the block synthesizing filters. The merge module can be adapted to merge (for example, by mixing) the corresponding signals of an intermediate synthesized subbands to obtain a signal synthesized sub-band. As already noted, the intermediate synthesized sub-band, which exposes�I to the merger, can be obtained both by direct harmonic conversion, by the conversion based on the cross-pieces. The system according to this embodiment of the invention may also contain a basic decoder designed to decode the bit stream in the input signal. It can also enable the module HFR-processing, adapted for application information spectral bands, in particular, by performing the formation of the spectrum. The action module HFR-processing can be controlled by information encoded in the bit stream.

One of the embodiments of the invention provides for HFR multidimensional signals, for example, in a system designed to play audio in stereo format containing Z channels, such as left, right, center, surrounds, etc. In one possible implementation of the input signal with multiple channels of the processed discrete values of each channel are based on the same number of discrete input values, although the stretching factor S and the conversion coefficient Q for each band may vary between channels. To this end, the implementation may include a block analyzing filters used to generate Y signals of the analyzed sub-bands of each channel, fashion�ü processing of subbands, used to generate Z signals of sub-bands, and the block synthesizing filters used to generate Z stretched in time and/or converted by the frequency signals, which form the output signal.

In the modifications of the preceding embodiment of the invention, the output signal may include the output channels, which are based on different amounts of signals of the analyzed sub-bands. For example, it may be appropriate to transfer a larger amount of computing resources to HFR for acoustically more pronounced channels; for example, channels intended for playback of sound sources located in front of the listener, may be given preference over others or rear channels.

It should be emphasized that the invention relates to all combinations of the above characteristic features, even if they are presented in different claims.

BRIEF DESCRIPTION of GRAPHIC MATERIALS

The present invention will be described below by means of illustrative examples, not limiting the scope or essence of the invention, with reference to the accompanying graphic materials.

Fig.1 illustrates the principle of harmonic conversion on the basis of the block of sub-bands.

Fig.2 illustrates the effect of nonlinear obra�ODI block of subbands to one input sub-band.

Fig.3 illustrates the effect of non-linear processing unit of the sub-bands with two input ranges.

Fig.4 illustrates the action of the harmonic conversion on the basis of the block of sub-bands, reinforced cross-pieces.

Fig.5 illustrates an example scenario of the use of transform-based block of subbands using several orders of magnitude conversion in the audio codec, an enhanced HFR.

Fig.6 illustrates an example scenario steps convert multiple orders on the basis of the block of sub-bands using a 64-band analyzing unit QMF filters.

Fig.7 and 8 illustrate experimental results of the described method for the conversion of block-based sub-bands.

Fig.9 shows details of the non-linear processing unit according to Fig.2, including prenormalization and multiplier.

DESCRIPTION of PREFERRED embodiments of the INVENTION

Described below are embodiments of the invention are only illustrations of the principles of the present invention "HARMONIC CONVERSION ON the BASIS of the BLOCK of sub-BANDS, the REINFORCED CROSS-PIECES". It should be understood that specialists in this field will be obvious modifications and changes described in this description schemes and details. Therefore, the intention is to �Ohm, so that the invention is limited only by the scope of the attached claims and not the specific details presented in this description for the purpose of description and explanation of embodiments of the invention.

Fig.1 illustrates the principle of transformation, of spacing, or combinations of conversion stretch in time-based block of subbands. The input signal in the time domain is fed to the block 101 of the analysing filter, which creates multiple complex-valued signals of sub-bands. These signals are fed into the module 102 processing of the subbands, which can be influenced by the control data 104. Each output subband can be obtained by processing or one, or two input sub-bands, or even as a result of superposition of several of these processed subbands. Many complex-valued output subbands is fed into the block 103 synthesizing filters, which in turn outputs the modified signal in the time domain. Optional control data 104 describe the configuration and processing parameters of the sub-bands that can be adapted to the signal to be converted. In the case of conversion, enhanced cross-pieces, these data may carry information, pertinently�call to the predominant fundamental frequency.

Fig.2 illustrates the effect of non-linear processing unit of the sub-bands to one input sub-band. For the data of target values of the physical spacing and conversion, as well as the physical parameters of the blocks 101 and 103 of the analyzing and synthesizing filters displays the parameters of spacing and conversion of sub-bands and sub-band index of the source for each index of the target sub-band. Then the purpose of processing the block of sub-bands is to implement the corresponding transformation, of spacing, or combinations of conversion stretch in time complex-valued signal source sub-band for the purpose of generating signals of the target sub-band.

Extractor 201 blocks makes complex-valued input signal sample of the final frame, consisting of discrete values. The frame is determined by the position of the input pointer and the conversion factor ranges. This frame undergoes nonlinear processing section 202 processing and then subjected to window processing Windows ultimate and possibly variable length in section 203 of the window treatment. The resulting discrete values are added to the previous output discrete values in the module 204 overlay and addition, where the position of the output frame is determined based on�eating an output pointer. The input pointer is incremented by a prescribed value, and the output pointer is incremented by the same amount, multiplied by stretching sub-bands. The repetition of this chain of operations will result in an output signal with a duration that is a duration of the input sub-band signal multiplied by the factor of stretching, up to the length of the synthesis window, and with an integrated frequency converted by the transform coefficient sub-bands. The control signal 104 may have an impact on each of the three sections 201, 202, 203.

Fig.3 illustrates the effect of non-linear processing unit of the sub-bands with two input signals of sub-bands. For given target values of the physical spacing and transformation and physical parameters of the blocks 101 and 103 of the analyzing and synthesizing filters displays the parameters of spacing and transform sub-bands, as well as two indexes of the source sub-bands for each index of the target sub-band. In the case where the nonlinear processing unit of the sub-bands to be used to create the missing harmonics by adding cross pieces, the configuration of the sections 301-1, 301-2, 302, 303, and also the values of the two indices of the source sub-bands may depend on the output� signal module 403 404 management of cross-processing. The aim of the processing unit of the sub-bands is to implement the corresponding transformation, of spacing, or combinations of conversion stretch in time for the combination of two complex-valued source signals of sub-bands for the purpose of generating signals of the target sub-band. The first extractor units 301-1 makes from the first complex-valued initial sub-band sampling a finite span of discrete values, and the second extractor 301-2 blocks selected from a finite span of discrete values of the second complex-valued initial sub-band. Frames are determined by the position of the input pointer and the conversion factor ranges. Both frames undergo non-linear processing section 302 and then subjected to window processing window of finite length in section 303 of window treatment. Module 204 overlay and addition may have a construction similar or identical to the module shown in Fig.2. The repetition of this chain of operations will result in an output signal with a duration equal to the duration of the two input signals of sub-bands multiplied by stretching sub-bands (up to the length of the synthesis window). In the case when both input signals have the same frequency, the output signal will have a comp�complete frequency transformed by the transform coefficient sub-bands. In the case where two input signals are of different frequencies, the present invention specifies that the window treatment 303 may be adapted to generate an output signal that has a target frequency, suitable for the generation of missing harmonics in the transformed signal.

Fig.4 illustrates the principle of enhanced cross-conversion works on the basis of the block of sub-bands, the spacing, or combinations of conversion stretch in time. Module 401 direct processing of the sub-bands may be of the type already described with reference to Fig.2 (section 202) or Fig.3. In the module 402 cross-processing sub-bands also serves a variety of complex-valued signals of sub-bands, and its action is influenced by the data 403 management of cross-processing. The module 402 cross-processing sub-bands performs the processing of blocks of sub-bands related to the type of processing two input signals of sub-bands described in Fig.3, and the output of the target sub-bands are added to the subbands from direct processing 401 of sub-bands in the adder 405. Data 403 management of cross-processing can be changed for each position of the input pointer and consist of, for less�th least following data:

- a selected list of target indices of the subbands;

- a pair of indices of the source sub-bands for each selected index of the target sub-band; and

Windows synthesis of finite length.

Module 404 management of cross processing gives these data 403 management of cross processing for a given part of the control data 104, describing the fundamental frequency, and a multitude of complex-valued output signals of sub-bands from the block 101 of the analysing filter. Control data 104 may also incur other dependent signal configuration parameters, which affect the processing of cross-pieces.

In the following text with reference to Fig.1-4 and by adding appropriate mathematical terminology, a description will be given of the principles of enhanced cross-pieces of spacing and convert the block-based sub-bands.

There are two main parameters configuration total harmonic Converter and/or the device of spacing in General are:

- Sφ, is the ratio required physical spacing, and

- Qφ- the ratio required physical transformation.

Blocks 101, 103 filters can relate to any type of modulated filters with integrated e�shonentai, such as QMF or windowed DFT or wavelet transform. Unit 101 analyzes the filter and the block 103 synthesizing filters may be evenly or unevenly arranged in the modulation and decide from a wide range of filter prototypes and/or Windows. Despite the fact that all of these options second order effect to such details in a subsequent design, as the correction phase, and control the display of sub-bands, the main design parameters of the system for processing the sub-bands, as a rule, are obtained from two private: ΔtS/ΔtAand Δƒs/ΔƒAthe following four parameters of filter, where all parameters are measured in physical units. In the above private:

- ΔtA- step or a time shift of discrete values of sub-band unit 101 analyzes filters (e.g., measured in seconds [s]);

- ΔƒA- a rating of the frequency sub-band unit 101 analyzes filters (e.g., measured in Hertz, [1/p]);

- ΔtS- step or a time shift of discrete values of sub-band block 103 synthesizing filters (e.g., measured in seconds [s]); and

- Δƒs- a rating of the frequency sub-bands block 103 synthesizing filters (e.g., measured in Hertz, [1/s]).

To configure the module 102 processing the sub-bands should R�to scitatj the following options:

- S - ratio stretching sub-bands, i.e., the stretching ratio applied in the module 102 processing of the subbands as the ratio of input and output discrete values with the aim of achieving General physical spacing of the signal in the time domain by means of the coefficient Sφ;

- Q - conversion coefficient of the sub-bands, i.e. the conversion factor that is applied in the module 102 processing the subbands to achieve overall physical transformation of the signal in the time domain by a factor Qφ; and

- the correspondence between the indices of the source and target subbands, where n denotes the index of the analyzed sub-bands included in the module 102 processing of subbands, and m is the index of the corresponding synthesized sub-band as the output signal of the module 102 processing sub-bands.

To determine the stretching factor S of subbands made the observation that the input signal in block 101 analyzing filters with a physical length D corresponds to the number of D/ΔtAdiscrete values of the analyzed sub-bands at the entrance to the module 102 processing of the subbands. These D/ΔtAdiscrete values will be stretched to S·D/ΔtAdiscrete values of the module 102 processing paddy�of Puzanov, which applies a stretching factor S subbands. The output of block 103 synthesizing these filters S·D/ΔtAdiscrete values result in an output signal having a physical duration ΔtA·S·D/ΔtA. Since this last the duration must match a specified value of Sφ·D, i.e. as the duration of the output signal in the time domain must be stretched in time compared with the input signal in the time domain by means of the coefficient Sφ, physical spacing, obtained the following design rule:

S=(ΔtA/ΔtS)Sϕ(1).

To determine the conversion coefficient Q subbands, which is used in the module 102 processing the subbands to achieve the physical transformation of Qφmade the observation that the input sine wave in block 101 analyzing filters with the physical frequency Ω will lead to complex signals of the analyzed sub-bands with discrete p� time angular frequency ω=2πΩ·Δt Aand the main contribution is made from the analyzed sub-band with index n≈Ω/ΔƒA- Sine wave output at the output of block 103 synthesizing filters with desired converted physical frequency Qφ-Ω will be a result of the filing of the synthesized subband with index m≈Qφ·Ω/ΔƒSthe complex sub-band signal with discrete angular frequency 2πQφ·Ω·ΔtS. In this context, care should be taken in order to avoid the synthesis of frequencies with overlapping spectra that differ from Qφ·Ω. Usually this can be avoided by making the appropriate versions of the second order, as discussed above, for example, by checking the appropriate blocks analyzing and/or synthesizing filters. Discrete frequency 2πQφ·Ω·ΔtSthe output of the module 102 processing the sub-bands should correspond to the discrete time frequency ω=2πΩ·ΔtAat the entrance to the block 102 of the processing sub-bands multiplied by the conversion coefficient Q subbands. I.e. equating 2πQΩΔtAto 2πQφ·Ω·ΔtSyou can define the following relation between factor Qφphysical transformation and the conversion coefficient Q sub-bands:

Q=(ΔtS/ΔtA )Qϕ(2).

Similarly, the corresponding index n of the source, or decomposed, sub-band module 102 processing of subbands for a given index m of the target, or synthesized, sub-band must satisfy the following condition:

n(ΔfS/ΔfA)(1/Qϕ)m(3).

In one of the embodiments of the invention, it is true that ΔƒS/ΔƒA=Qφ, i.e. the frequency spacing unit 103 synthesizing filter corresponds to the frequency spacing unit 101 analyzes filters, multiplied by the coefficient of physical transformation, and can be applied one-to-one mapping of the index of the analyzed sub-band in index of synthesized sub-band n=m. In other embodiments, the display indices podiamos�new may depend on the details of the parameters of the filter unit. In particular, if the quotient of the spacing between the frequencies of the block 103 synthesizing filter and the block 101 of the analyzing filters is different from the coefficient of Qφphysical transformation for the given target subband can be assigned one or two original sub-band. In the case of two source subbands may be preferable to use two adjacent source subbands with indices n, n+1, respectively. That is, the first and second source sub-bands have the appearance of or (n(m) n(m)+1), or (n(m)+1, n(m)).

Processing of the subbands shown in Fig.2, with the only source subband will now be described as a function of the parameters S and Q processing of the subbands. Let x(k) is input to the extractor 201 blocks, and let h be the input step block. I.e. x(k) is complex-valued signals of the analyzed sub-band with index n. The block extracted by the extractor 201 blocks, we can without loss of generality to consider as defined by L=R1+R2discrete values:

xi(k)=x(Qk=hl),k=R1...R21(4 ).

where the integer l is the index counting blocks, L is the block length, and R1, R2non - negative integers. Note that for Q=1 the block is extracted from successive discrete values, but for Q greater than 1 - is decreasing discretization so that the input address was stretched by a factor of Q. If Q is an integer, this operation usually is performed directly, while for non-integer values of Q may require interpolation. This statement is also true for non-integer values of the increment h, i.e. for the input step block. In one of the embodiments of the invention, the complex-valued sub-band signal can be applied short of the interpolating filters, e.g., filters having two links of the filter. For example, if you want a discrete value with a fractional temporal index k+0,5 sufficient quality may be ensured of a two-level interpolation of the form x(k+0,5)≈Ah(k)+bx(k+1), where the coefficients a, b can be a constant or may depend on the subband index (see, for example, documents WO No. 2004/097794 and WO No. 2007/085275).

An interesting special case of formula (4) is Ri=0, R2=1, where the extracted block consists of only discrete values ie the block length is L=1.

In the polar representation of a complex number z=|z|exp(i∠z) where |z| is the amplitude of a complex number, and ∠z - phase of a complex number, the module 202 non-linear processing, generating the output frame y1from the input frame ximainly determined by the modification coefficient of the phases T=SQ via:

{y1(k)=(T1)x1(k)+θ|y1(k)|=|x1(0)|ρ|x1(k)|1ρ},k=R1,KR21(5),

where ρ∈[0,1] is a geometric weighting parameter amplitude. The case p=0 corresponds to a pure modification of the phases of the extracted block. Especially attractive�th weighing value of the amplitude is ρ=1-1/T for which the removal of some computational complexity is obtained regardless of the length of the block L, and the resulting transient response improves slightly relative to the case ρ=0. The parameter θ of phase correction depends on the details of the filter unit, as well as indices of the source and target subbands. In one of the embodiments of the invention, the parameter in the correction phase can be determined experimentally by the sweep of the set of input sinusoids. In addition, the parameter in the correction phase can be obtained by examining the phase difference related complex sinusoids of the task or sub-band by optimizing the performance on the input signal type of the Dirac impulse. Finally, with appropriate design of the blocks 101 and 103 of the analyzing and synthesizing filters option in the correction phase can be set to zero or omitted. The coefficient T of a modification of the phases needs to be an integer so that the coefficients of T-1 and 1 be integers in linear combinations of the phases in the first line of the formula (5). With this assumption, i.e. assuming that the coefficient T of a modification of the phase is an integer, the result of nonlinear processing is well defined even though the phases are ambiguous because of the identification module 2π.

In words, the formula (5) determines that the phase of the discrete value output�th frame is determined by the phase shift of the corresponding discrete values of the input frame by a constant offset. The constant offset value may depend on the ratio T of a modification, which itself depends on the ratio of stretching sub-bands and/or from the transform coefficient sub-bands. In addition, a constant offset value may depend on the particular phase of the discrete values of the input frame from the input frame. Specified particular discrete value of the input frame is stored unchanged for determining the phases of all discrete values of the output frame for the current block. In the case of formula (5) as a particular phase of the discrete values of the input frame is the Central phase of the discrete values of the input frame.

The second line of formula (5) determines that the amplitude of the discrete values of the output frame may depend on the amplitude of the corresponding discrete values of the input frame. In addition, the amplitude of the discrete values of the output frame may depend on the particular amplitude of the discrete values of the input frame. This particular discrete value of the input frame may be used to determine the amplitudes of all discrete values of the output frame. In the case of formula (5), as a special discrete values of the input frame using the center of the discrete value of the input frame. In one of the embodiments of the invention, the amplitude of the discrete values�of the output frame may correspond to the geometric mean of the amplitude of the corresponding discrete values of the input frame and a special discrete values of the input frame.

In the module window 203 of the processing to the output frame is applied, a window w of length L, which leads to the window output frame:

z1(k)=w(k)y1(k),k=R1...R21(6)

Finally, it is assumed that all frames are padded with zeros, and the operation 204 overlay and addition is defined as

z(k)=lzl(kshl),(7)

where it should be noted that the module 204 overlay and addition applies the step block Sh, i.e. the time step, which is S times larger than the step h of the input block. Because of this difference in time steps according to the formulas (4) and (7) the duration of the output signal z(k) to S times the duration of the input signal x(k), i.e., the signal synthesis�trolled sub-band is stretched by the stretch factor S subbands compared with the signals of the analyzed sub-bands. It should be noted that this observation is generally applicable when the length L of the window is negligible compared to the duration of the signal.

In the case when the input signal processing 102 of the sub-bands uses a complex sinusoid, i.e. signals of the analyzed sub-band corresponds to a complex sinusoid:

x(k)=Cexp(iωk)(8),

using formulas (4) to(7), we can determine that the output signal processing 102 of the sub-bands, i.e., a signal synthesized sub-band, is:

z(k)=|c|exp[i(TC+θ+Qωk)]lw(kShl)(9)

regardless of ρ. Thus, the complex sinusoid disk�ethno-time frequency ω is converted into a complex sinusoid with a time discrete frequency Qω provided the synthesis window is shifted with a step of Sh, that the sum for all k leads to the same constant:

lw(kShl)=k(10)

To illustrate, consider the special case of a pure transformation, where S=1 and T=Q. If the step input block h=1 and R1=0, R2=1, all of the above, i.e.

formula (5), reduces to the pointwise or based on discrete values of rule modification of phases:

{z(k)=Tx(k)+θ|z(k)|=|x(k)|}. (11)

To set a particular processing parameters, i.e. the length of the block in the extractors blocks, the module 102 processing p�of diapazonov can use the control data 104.

Below is a description of the processing of the subbands will be expanded to the case shown in Fig.3, two input signals of sub-bands. Let x(1)(k) to the signal input sub-band in the first extractor units 301-1, and let x(2)(k) is the input signal of the second subband extractor 301-2 blocks. Each extractor may use a different downsampling factor, which leads to the recoverable blocks:

{xl(1)(k)=x(1)(Dlk+hl)xl(2)(k)=x(2)(D2k+hl)}k=-R1,KR21(12)

Nonlinear processing 302 generates an output ka�R land can be defined as

{yl(k)=Tlxl(1)(k)+T2xl(2)(k)+θ|yl(k)|=|xl(1)(k)|1ρ|xl(2)(k)|ρ}and (13)

processing 303 is again described by formulas (6) and (7), 204 and processing identical to the processing of the overlay and addition described in the context of the case with a single input signal.

The definition of non-negative real parameters D1, D2, ρ, a non-negative integer pairs�spectra of T 1, T2and the synthesis window w now depends on the required operating mode. Note that if both inputs to the same sub-band, x(1)(k)=x(2}(k) and D1=Q, D2=Q, T1=1, T2=T-1, then the operation according to the formulas (12) and (13) are reduced to operations on formulas (3) and (4) in the case of a single input signal.

In one of the embodiments of the invention, where the ratio of the spacing ΔƒS, frequency unit 103 synthesizes filters and spacing ΔƒAfrequency unit 101 analyzes filters differs from the desired ratio of Qφphysical transformation, it may be helpful to determine discrete values of the synthesized subband with index m of the two analyzed sub-bands, respectively, with the indices n, n+1. For a given index m corresponds to the index n can be an integer value obtained by truncating the values of n of the analyzed index, having a form according to the formula (3). One of the signals of the analyzed sub-bands, for example, signals of the analyzed sub-band corresponding to the index n, is injected into the first extractor units 301-1 and the other signals of the analyzed sub-bands, for example, a signal corresponding to the index n+1, is fed into the second extractor 301-2 blocks. On the basis of these two signals of the analyzed sub-bands by �written above processing is determined by the signal synthesized sub-band, corresponding to the index m. The appointment of adjacent signals of the analyzed sub-bands of the two extractors 301-1 and 302-1 blocks may be based on the residue that is obtained by truncating the values of the index according to the formula (3), i.e. the difference between the exact value of the index, having a form according to the formula (3), and a truncated integer value of n, obtained by the formula (3). If the remainder is greater than 0.5, the signal of the analyzed sub-band corresponding to the index n, can be assigned to the second extractor 301-2 blocks, otherwise, the signals of the analyzed sub-bands can be assigned to the first extractor units 301-1. In this operating mode the parameters can be designed so that the input signals of sub-bands share the same complex frequency ω:

{x(1)(k)=C1exp(iωk)x(2)(k)=C2exp(iωk)}, � (14)

which leads to the signal output sub-band, which is a complex sinusoid with a time discrete frequency Qω. It turns out that this behavior occurs when the following relations are true:

{Q=T1D1+T2D2SQ=T1+T2}. (15)

For the operating mode of generation of missing harmonics by means of cross-pieces design criteria are different. Returning to the parameter Qφphysical conversion, for the purpose of adding works is the generation of the output signal at frequencies QφΩ+rΩ0where r=1, ..., Qφ-1, for given input signals at frequencies Ω and Ω+Ω0where Ω0- the fundamental frequency, which is the predominant component of the pitch of the input signal. As described in document WO 2010/01892, selective addition of these members will result in the filling of the harmonic series and a significant weakening of spurious artifact of the pitch.

Below will be described the constructive algorithm of the control 404 cross-processing. For a given index t target sub-band output, the parameter r=1, ..., Qφ-1 and fundamental frequency Ω corresponding index n1and n2the original sub-bands can be obtained by solving in approximate sense of the following system of equations:

{m+σ=QϕΩ+rΩ0ΔfSn1+σ=ΩΔfAn2+σ=Ω+Ω0ΔfA}and (16)

where σ=1/2 for unevenly arranged modulation filter unit (normally used for QMF blocks and DCT filters) and σ=0 for uniformly arranged modulation filter unit (normally used for blocks of FFT filters).

For definitions

- ρ=Ω0/ΔƒA- the fundamental frequency, measured in units of frequency distribution unit of the analyzing filters;

- F=Δƒs/ΔƒA- the ratio of the spacing between the frequencies synthesized and analyzed sub-bands; and

- nf=((m+σ) (F-rp)·Qφ-σ - dejstvitelnostyu index target to lower the original index with an integer value

the predominant example of the approximate solution of the system of equations (16) has the form of selection and/ as integer nearest to nfand n2- as the integer nearest to nf+ρ.

If the fundamental frequency less than the frequency spacing unit of the analyzing filters, i.e., if p is less than 1, it may be preferential to cancel the adding of works.

As indicated in document WO 2010/081892, cross-product should not be added to the output sub-band, which already contains a significant main contribution from the conversion without crossover works. Moreover, the contribution to the cross-product must pay, at most, one of the cases r=1, ..., Qφ-1. In this description, these rules can be implemented by performing the following three steps for each index t target sub-band output:

1. To calculate the maximum amplitude of the Mwithfor all options r=1, ..., Qφ-1 minimum and�of plitude original sub-applicants |x (1)| and |x(2)|, evaluated at the Central time slice k=hl (or its surroundings), where the original sub-bands of x(1)and x(2)may be in the form of index n1and n2as in equation (16);

2. To calculate the corresponding amplitude of the Msto direct the source member |x| obtained from the initial sub-band with index n=(F/Qφ)m (cf. equation (3));

3. Powering the cross member from the winning option for Mwithin the above stage 1 only if Mwithmore qMs, where q is a pre-defined threshold.

Depending on the specific system configuration parameters may be desirable modifications of this procedure. One such change is the replacement of the hard threshold at stage 3 with a more lenient rules that depend on private Mwith/Ms. Another change is to spread the maximization in step 1 by more than Qφ- 1 variants, for example, defined by a finite list of values-candidates for the fundamental frequency, measured in units p of the spacing of the analyzed frequencies. Another change is the use of other quantitative indicators of the amplitudes of the sub-bands, such as the amplitude of the fixed discrete values, maximum amplitude, mean amplitude, amplitude in the sense of�e l φ-rules, etc.

The list of target source bands m selected for the adding works, together with values of n1and n2is main data 403 management of cross-processing. It remains to describe the configuration parameters D1, D1, ρ, a non-negative integer parameters T1, T2appearing with the alternation of phases (13), and a window into the synthesis, intended for use in cross-processing sub-bands 402. Insert sinusoidal model with cross-product leads to the following source signals of sub-bands:

{x(1)(k)=C1exp(iωk)x(2)(k)=C2exp(i(ω+ω0)k)}and (17)

where ω=2πΩΔtA and ω0=2πΩ0ΔtA. Similarly, the desired output sub-band is of the form

z(k)=C3exp[iQ(ω+rω0/Qϕ)k](18).

Calculations reveal that this target output signal can be obtained to satisfy the condition (15) together with

T2/(T1+T2)=r/Qϕ(19)

Conditions (15) and (19) is equivalent to the following expression:

{T1=(Qϕr)SϕT2=rSϕ (Qϕr)D1+rD2=Qϕ/S}and (20)

which determines the integer coefficients of T1, T2to modify the phases in (13) and provides some freedom in the design by specifying values of the coefficients for downsampling (D1, D2. The weighting parameter amplitudes mainly can be chosen as ρ=r/Qφ. As you can see, these configuration parameters depend only on the fundamental frequency Ω0through the choice of r. However, in order to correct the equation (18), there is a new condition for the window w of the synthesis, namely:

The window w of synthesis, that is either exactly or approximately satisfies the condition (21) must be provided in the last data item 403 management of cross-processing.

Note that the above algorithm to calculate data 403 management of cross processing on the basis of input parameters such as the index t target sub-band output and the fundamental frequency Ω0has pure and�ilustrativo essence, and as such, does not limit the scope of the invention. Changes to this disclosure within the knowledge and everyday experience of experts in this field, for example, additional processing method based on the block of subbands, generating the signal (18) as an output signal in response to input signals (17), wholly fall within the scope of the present invention.

Fig.5 illustrates an example scenario for the application of transform-based block of subbands using several orders of magnitude conversion in the audio codec, an enhanced HFR. The transmitted bit stream is accepted by the basic decoder 501, which generates the decoded base signal with low-frequency bandwidth at a sampling rate ƒS. The decoded signal with low-frequency bandwidth re is sampled to the output of the sampling frequency 2ƒSusing a 32-band unit 502 of complex modulated QMF analyzing, followed by a 64-band block 505 QMF synthesizing (inverse QMF). Both units 502 and 505 of the filters share the same physical parameters Δts=ΔtAand ΔƒS=ΔƒAand module 504 HFR-processing just misses the unmodified low frequency sub-bands corresponding to the baseline signal with low-frequency bandwidth. High frequency content in�of the output signal is obtained by feeding the higher frequency sub-bands in a 64-band block 505 QMF synthesizing the output bands from the module 503 multiple Converter, exposure to the formation of the spectrum and modifications carried out by the module 504 HFR-processing. Multiple Converter 503 receives as input the decoded base signal and outputs a variety of signals of sub-bands that represent the 64-band analysis of the superposition, or a combination of several components of the transformed signal. The aim is to, if HFR-processing costs, each component corresponded to physically transform integer without stretching the base signal in time (Qφ=2, 3 ...,Sφ=1). In the scenario according to the invention, the signal 104 to the Converter control contains data that describes the fundamental frequency. These data can either be transmitted through the bitstream from the corresponding audio codecs, or be output by detecting the main tone decoder, or be obtained from a combination of transmitted and detected information.

Fig.6 illustrates an example scenario of the transformation of several orders of magnitude on the basis of the block of sub-bands using a single 64-band analyzing unit QMF filters. Here the generation and delivery in the area of the 64-band QMF operating at a sampling frequency 2ƒsare subject to three orders of magnitude transform of Qφ=2, 3, 4. Module 603 merge simply selects and combines Zn�chimo subbands of the branches of each order conversion into a single lot QMF-subbands, be fed into the module HFR-processing. The aim, in particular, is that the processing chain, consisting of a 64-band QMF analysis 601, module 602-Qφprocessing of sub-bands and a 64-band QMF-605 synthesis, leads to physical transform with a factor Qφand Sφ=1 (i.e., without stretching). When you identify the three blocks through 101, 102 and 103 of Fig.1 found that ΔtA=64ƒSand ΔƒAS=128, so that ΔtS/ΔtA=1/2 and F=Δƒs/ΔƒA=2. Designing the specific configuration settings for 602-Qφwill be described for each of the cases Qφ=2, 3, 4 separately. For all cases, the analysis step is chosen as h=1, and assume that the well-known normalized parameter fundamental frequency ρ=Ω0/ΔƒA=128Ω0S.

First, consider the case of Qφ=2. In this case, 602-2 should perform stretching sub-band by a factor of S=2 and conversion of sub-band by a factor Q=1 (i.e., not to convert), the correspondence between source n and target m sub-bands for direct processing of sub-bands has the form n=m. In the scenario of adding cross pieces according to the invention there is only one type of cross-pieces for consideration, namely: r=1 (see the discussion after equation (5)), and equations (20) reduce to T1=T2=1 and D1+D2=1. The sample solution consists of choosing D1=0 and D2=1. As the synthesis window direct processing can be used a rectangular window of odd length L=10 with R1=R2=5, since it satises (10). For window synthesis in cross-processing can be used short box with L=2 units and R1=R2=1 in order to support the additional complexity of adding works to a minimum. However, the favorable effect of a long unit for processing the sub-bands is the most significant in the case of complex audio signal, where the suppressed unwanted matching members; in the case of dominant fundamental tone, the appearance of these artifacts is less likely. A window with L=2 links is the shortest of those which can satisfy the condition (10) since h=1 and S=2. However, according to the present invention, the window is mostly satises (21). With the available options this is equivalent to the following condition:

{w(1)=w(0) w(v)=w(v)exp(iav),a=πρ/2},

which is performed by choosing w(0)=1 and w(-1)=exp(iα)=exp(iπρ/2).

For the case of Qφ=3 specification for 602-3, having the form of conditions (1) to(3), such that it must perform stretching subband S=2, the conversion of sub-band Q=3/2, and the correspondence between source n and target m sub-bands for processing direct members has the form n≈2m/3. There are two types of members with cross-product r=1, 2, and equations (20) reduce to

{T1=3rT2=r(3r)D1+rD2=3/2}

The sample solution consists of choosing the parameters of downsampling as

- D1=0 and D2=3/2 for r=1;

- D1=3/2 and D2=0 for r=1.

As the synthesis window direct processing can be a rectangular window of odd length L=8 with R 1=R2=4. As a window treatment cross-pieces can be used short box with L=2 units and R1=R2=1 satisfying the condition

{w(1)=w(0)w(v)=w(v)exp(iαv)α=πρr(3r)3(D2D1)}and

which is performed by choosing w(0)=l and w (l)=exp(iα).

In the case of Qφ=4 specification for 602-4, having the form of conditions (1) to(3), such that it must perform stretching sub-band by a factor of S=2, conversion of the sub-band by a factor Q=2, and the correspondence between source n and target m sub-bands for processing direct members has the form n=2m. There are three types of members with cross product, r=1, 2, 3, and equations (20) reduce to

{T1=4rT2=r(4r)D1+rD2=2}

The sample solution consists of choosing

- D1=0 and D2=2 for r=1;

- D1=0 and D2=1 for r=2;

- D1=2 and D2=0 for r=3.

As the synthesis window direct processing, you can use a rectangular window of even length L=6 R1=R2=3. As a window treatment cross-pieces can be used short box with L=2 units and R1=R2=1 satisfying the condition

{w(1)=w(0)w(v)=w(v)exp(iav),a=πρr(4r)4(D2 D1)}and

which is performed by choosing w(0=1) and w(-1)=exp(iα).

In each of these cases, where applicable more than one value of r, will be choice, for example, similarly to the three-step procedure described before equation (17).

Fig.7 depicts the amplitude spectrum of the harmonic signal with the fundamental frequency Ω0=564,7 Hz. The low-frequency part of the signal 701 is to be used as input for multiple Converter. The purpose of the Converter is to generate a signal, as much as possible close to the high-frequency part 702 of the input signal, so the transmission of the high-frequency part 702 becomes optional, and available bit rate data transmission can be used sparingly.

Fig.8 depicts the amplitude spectrum of the output signals from the Converter, which contains as input the low-frequency portion 701 of the signal shown in Fig.7. Multiple Converter is constructed by using a 64-band QMF blocks filters and input sampling rates ƒs=14400 Hz in accordance with the description of Fig.5. However, for clarity only considered two of the transformation Qφ =2,3. Three different panels 801-803 represent the final output signal obtained by using different units of management data of cross-processing.

The top panel 801 depicts the output spectrum obtained when cancelled all processing cross-pieces, and is only active direct processing 401 of sub-bands. This will be the case when the control 404 cross processing does not accept the basic tone, or p=0. Transformation through a Qφ=2 generates the output signal in the range 4-8 kHz, and transformation through a Qφ=3 generates the output signal in the range of 8-12 kHz. As can be seen, the generated harmonics are a large and growing distance from each other, and the output signal deviates significantly from the target high-frequency signal 702. In the resulting output audio signal will be audible artifacts doubled and trebled "parasitic" of the pitch.

Middle panel 802 depicts the output spectrum obtained when the cross processing works active that uses the parameter of the pitch p=5 (which is an approximation 128Ω0/ƒs=5,0196), but for cross-processing sub-bands uses a simple two-tier box, the synthesis of w(0)=w(-1)=1, satisfying the condition (10). This is equivalent to the combined direct�th processing block-based sub-band and harmonic conversion, reinforced crisscross works. As shown, additional components of the output signal compared to the 801 did not coincide exactly with the desired harmonic series. This shows that when using the cross processing works procedures inherited from the direct processing of the subbands, the treatment leads to insufficient for use as sound quality.

Bottom panel 803 depicts the output spectrum obtained under the same scenario as for the middle panel 802, but, in this case, with Windows of synthesis when cross-processed sub-bands having the form of the formulas described in the case of Qφ=2,3 according to Fig.5, that is: a two tiered box of synthesis in the form w(0)=1 and w(-1)=exp(iα), satisfying the condition (21), and pointed to the present invention the characteristic feature of which is that it depends on R. As can be seen, the combined output signal coincides very well with the desired harmonic series 702.

Fig.9 shows the non-linear processing module 202 of frame processing, including section configured to receive two input discrete values of u1and u2and to generate them on the basis of the processed discrete values of w, the amplitude of which has the form of the geometric mean of the amplitudes of the input values, and the phase of which is a fun� linear combination of the phases of input of discrete values, ie:

{|w|=|u1|ρ|u2|1ρargw=T1argu1+T2argu2(22)

According to this description, the value of w to be processed can be obtained by first normalizing each of the discrete values of u1u2in the corresponding prenormalization 901, 902 and pre-multiplying the normalized input discrete values of v1=u1/|u1|av2=u2/|u2|bin the weighted multiplier 910, which displays w=v1αv2β. It is clear that the action of prenormalization 901, 902, and a weighted multiplier 910 is determined by the input parameters a, b, α and β. It is easy to verify that equation (22) will be performed, if α=T1, β=T2a=1-ρ/T1b=1-(1-ρ)/T. Experts will easily generalize e�the scheme on an arbitrary number of N 0the input of discrete values, where the multiplier is fed N0the input of discrete values, some or all of which were subjected to normalization. Then, you will discover that the total pre-normalization (a=b under the assumption that prenormalization 901, 902 generate the same results) is possible if ρ to be equal to ρ=T1/(T1+T2). This leads to a computational advantage then, when one considers the large number of sub-bands, as in all sub-bands of the candidate before the multiplication can be global stage prior to normalization. In preemptive hardware implementation of a number of identically functioning of prenormalization is replaced by a single module, which alternates between discrete values from different ranges with time division.

Further embodiments of the present invention will become apparent to experts in this field after reading the above description. Although the present description and the drawings disclose variants and embodiments of the invention, the invention is not limited to these specific examples. Numerous modifications and changes may be made without departing from the scope of the present invention, which is defined by the accompanying forms�Loy invention.

Disclosed above in this specification, the system and method may be implemented as software, firmware, hardware or their combination. Some components or all the components may be implemented as software executable by a processor for digital signal processing or microprocessors, or they may be implemented as hardware or as a integrated circuit special purpose. This software may be distributed on computer-readable media may include computer storage media (or timeless media) and communication tools (transitory media). As is well known to specialists in this field, computer storage media include volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable commands, data structures, program modules or other data. Computer storage media include as non-limiting examples of RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, CD-DVD, or other optical disk media, magnetic cassettes, magnetic tape, magnetic disk storage media or other magnetic hard�subject the device or any other medium which can be used to store the desired information and for access to her computer. As is also well known to specialists in this field, communications, generally include computer-readable commands, data structures, program modules or other data in the modulated data signal such as a carrier wave, or other transmission mechanism, and includes any medium of information delivery.

1. The system is configured to generate stretched in time and/or converted by the frequency signal from the input signal, and the system contains:
unit (101) of the analysing filter, configured to obtain a certain number Y≥1 signals of the analyzed sub-bands of the input signal, where each signal is analyzed sub-band comprises the number of analyzed complex-valued discrete values, each of which has a phase and amplitude;
module (102) processing of sub-bands configured to generate a signal synthesized sub-band of the Y signals of the analyzed sub-bands using the conversion coefficient Q sub-bands and the stretch factor S sub-bands, wherein at least one of the coefficients Q and S is greater than one, where the module (102) processing of sub-bands contains:
ekstrak�or (201) blocks, configured for:
(i) the formation of the Y frames consisting of L discrete input values, and each frame is extracted from the specified number of analyzed complex-valued discrete values in the signals of the analyzed sub-bands, and the frame length is L>1; and
ii) applying the values jump from h block of discrete values to a specified number of analyzed discrete values before generating the next frame consisting of L discrete input values, thereby generating a sequence of discrete input values;
module (202) nonlinear frame processing configured to generate, on the basis of Y corresponding frames of input discrete values generated by the extractor blocks of the frame of the processed discrete values by determining the phase and amplitude of each of the processed discrete values of the frame, where at least one of the processed discrete values:
(i) the phase of the processed discrete values based on the respective phases of the corresponding input digital values in each of the Y frames of discrete input values; and
(ii) the amplitude of the processed discrete values based on the amplitude of the corresponding input digital values in each of the Y frames of the input discrete value�rd; and
module (204) addition and overlay configured to determine the signal synthesized sub-band by superimposing and adding discrete values from a sequence of frames of the processed discrete values;
and
unit (103) synthesizing filter configured to generate stretched in time and/or converted by the frequency signal from the signal synthesized sub-band,
where the system operates, at least when Y=2.

2. A system according to claim 1, characterized in that
unit (101) of the analyzing filters is one of the following: block quadrature mirror filters, windowed discrete Fourier transform or a wavelet transform; and
block (103) of the synthesizing filter is a block of the corresponding inverse filters or reverse conversion.

3. A system according to claim 2, characterized in that the block (101) of the analysing filter is a 64-point block quadrature mirror filters, and block (103) of the synthesizing filter is a 64-point block backward quadrature mirror filters.

4. System according to any one of the preceding claims, characterized in that
unit (101) analysing applies filters to the input signal step in the analysis of time - ΔtA;
the unit of the analyzing filters is the spacing of the analyzed frequency - �f A;
the unit of the analyzing filters contains the number N of analyzed sub-bands, wherein N>1, where n is the index of the analyzed sub-bands, n=0...N-1;
the analyzed sub-band from among N of the analyzed sub-bands is associated with one of the frequency bands of the input signal;
unit (103) synthesizing filter applies to the signal synthesized sub-band step synthesis of time - ΔtS;
block synthesizing filters has a spacing of synthesized frequencies ΔfS;
block synthesizing filters contains the number M of synthesized subbands, and M>1, where m is the index of the synthesized sub-band, and m=0, ..., M-1; and
the synthesized sub-band from among M of synthesized sub-bands is associated with one of the frequency bands stretched in time and/or converted by the frequency of the signal.

5. A system according to claim 4, characterized in that the module (102) processing of sub-bands configured for Y=2 and further comprises a module (404) management of cross-processing, configured to generate the data (403) manage cross-processing, determining the indices of the n1n2sub-bands associated with the signals of the analyzed sub-bands so that the indices of the subbands differ by an integer p, the approximation relation of the fundamental frequency Ω0input signals the spacing Δf Athe analyzed frequencies.

6. A system according to claim 4, characterized in that the module (102) processing of sub-bands configured for Y=2 and further comprises a module (404) management of cross-processing, configured to generate the data (403) manage cross-processing, determining the indices of the n1n2sub-bands associated with the signals of the analyzed sub-bands and with the index m of the synthesized sub-band, where the specified indices are correlated as approximate integer solutions of a system of equations

where Ω0- the fundamental frequency of the input signal;
σ=0 or;
Q=(Δts/ΔtA)·Qφand
r is an integer satisfying the inequality 1≤r≤Qφ-1.

7. A system according to claim 6, characterized in that the module (404) management of cross-processing is configured to generate control data processing so that the indices of the n1n2sub-bands based on the value of r that maximizes the minimum of the amplitudes of the two sub-bands of frames formed by extracting the analyzed discrete values of the signals of the analyzed sub-bands.

8. A system according to claim 7, characterized in that the amplitude of the sub-band for each frame consisting of L discrete input values, ameri� an amplitude Central or nearest to the Central discrete values.

9. A system according to any one of claims.1-3, characterized in that the extractor (201) block configured to receive at least one frame of input digital values by downsampling analyzed complex-valued discrete values in the signals of the analyzed sub-bands.

10. A system according to claim 9, characterized in that is configured for Y=2, where the extractor blocks configured to receive the first and second frames of input discrete values by downsampling analyzed complex-valued discrete values, respectively, in the first and second signals of the analyzed sub-bands by means of the coefficients of D1and D2downsampling satisfying
and the module (202) nonlinear frame processing is configured to determine the phase of the processed discrete values based on a linear combination with nonnegative integer coefficients (T1, T2the respective phases of the corresponding input digital values in the first and second frames of input discrete values.

11. A system according to any one of claims.1-3, characterized in that the module (102) processing of sub-bands further comprises a module (203) of window treatment, which is in front of the module (204) overlay and addition, and konfigurirovanii� to apply to the frame of the processed discrete values of the window function of finite length.

12. A system according to claim 11, characterized in that the window function has a length, which corresponds to the length L of the frame, and the window function is one of:
the Gaussian window,
cosine window,
the raised cosine window,
Hamming window,
box Hannah,
rectangular window,
the Bartlett window, and
window Blackman.

13. A system according to claim 11, wherein the windowing function includes a number of windowed discrete values, and superimposed and folded windowed discrete values of a number of window functions when weighing by complex weighting coefficients and the shift to the value of the jump form Sh is largely consistent sequence.

14. A system according to claim 13, characterized in that the successive complex weighting coefficients differ only by a fixed phase sequence.

15. A system according to claim 14, characterized in that the phase rotation is proportional to the fundamental frequency of the input signal.

16. A system according to any one of claims.1-3, characterized in that the module (204) overlay and addition applies the magnitude of the jump next to each other frames of the processed discrete values, the magnitude of the jump is equal to the h value of the jump of a block, multiplied by a stretching factor S subbands.

17. A system according to any one of claims.1-3, characterized in that made with the possibility of functioning, �ENISA least for Y=1 and Y=2.

18. A system according to claim 17, characterized in that made with the possibility of operation at least for one of the future values of Y > 3.

19. A system according to any one of claims.1-3, characterized in that is configured for Y=2, and the module (202) frame processing is configured to determine the amplitude of the processed discrete values as the average value of the amplitude of the corresponding input discrete values in the first frame of input digital values and the amplitude of the corresponding input digital values in the second frame of input digital values.

20. A system according to claim 19, characterized in that the module (202) nonlinear frame processing is configured to determine the amplitude of the processed discrete values as a geometric weighted average.

21. A system according to claim 20, characterized in that the geometric parameters weighting the amplitudes are ρ and 1-ρ, where ρ is a real number that is inversely proportional to the ratio Q transform sub-bands.

22. A system according to any one of claims.1-3, characterized in that is configured for Y=2, and the module (202) nonlinear frame processing is configured to determine the phase of the processed discrete values based on a linear combination with nonnegative integer coefficients (T1, T2) �sootvetstvujushij phases of the corresponding discrete values in the first and second frames of input discrete values.

23. A system according to claim 22, characterized in that the amount specified integer coefficients is the product Q×S stretching ratio and the conversion factor.

24. A system according to claim 22, characterized in that the phase of the processed discrete values corresponds to a specified linear combination of phases plus the parameter θ of phase correction.

25. A system according to any one of claims.1-3, characterized in that the extractor (201) blocks configured to interpolate two or more of the analyzed discrete values for the purpose of receiving input discrete values.

26. A system according to any one of claims.1-3, characterized in that it further comprises a module receiving the control data configured for receiving control data (104), wherein the module (102) processing of sub-bands is configured to determine the signal synthesized sub-band based management data.

27. A system according to claim 26, characterized in that is configured for Y=2, and the specified control data (104) may include the fundamental frequency Ω0the input signal, where the module (102) processing of sub-bands is configured to determine the analyzed sub-bands, of which must be received processed discrete values, so that their spacing frequency was proportional to the fundamental frequency.

28. System l�the Boma one of claims.1-3, characterized in that the module (102) non-linear processing comprises:
prenormalization (901, 902) configured to scale the amplitudes of the respective input discrete values in at least one of Y frames of the input discrete values (vm=um/|um|βm); and a complex multiplier (910) configured to determine the processed discrete values by calculating a weighted complex worksof coefficients equal to the corresponding input discrete value, at least two of the Y frames of the input of discrete values, wherein at least one of the coefficients (Vm, m∈M≠⌀) obtained from the discrete values of the amplitude, the scale of which was changed by prenormalization.

29. A system according to any one of claims.1-3, characterized in that is configured for Y=2, which contains:
unit (101) of the analysing filter, configured to receive first and second signals of the analyzed sub-bands of the input signal;
module (102) processing of sub-bands is configured to determine the signal synthesized sub-band from the first and second signals of the analyzed sub-bands, wherein the module (102) processing of sub-bands contains:
the first extractor (301-1) units configured for:
(i) forming�ing first frame, consisting of L discrete input values from a specified number of analyzed complex-valued discrete values of the first signals of the analyzed sub-bands, and the frame length L>1; and
ii) applying the values jump from h block of discrete values to a specified number of analyzed discrete values before generating the next frame consisting of L discrete values, whereby is generated a first sequence of frames of the input discrete values;
the second extractor (301-2) units configured for:
(i) forming the second frame consisting of L discrete input values from a specified number of analyzed complex-valued discrete values of the second signals of the analyzed sub-band; and
ii) applying the values jump from h block of discrete values to a specified number of analyzed discrete values before generating the next frame consisting of L discrete values, whereby is generated a second sequence of discrete input values;
module (302) nonlinear frame processing configured to generate, based on the first and second frames of input discrete values of the frame of the processed discrete values; and
module (204) overlay and addition, configured to generate the synthesized signal �of diapason; and
unit (103) synthesizing filter configured to generate stretched in time and/or converted by the frequency signal from the signal synthesized sub-band.

30. A system according to any one of claims.1-3, characterized in that it further comprises:
a number of modules(401, 402; 503; 602-2, 602-3, 602-4) processing of sub-bands, each of which is configured to determine the intermediate signal synthesized sub-band using different values of the coefficient Q transform sub-bands and/or the stretching factor S of subbands; and
module (405; 603) of a merger, conveniently located for the specified number of processing modules of the sub-bands and in front of the block (103) of the synthesizing filter, configured to merge the respective intermediate signals of synthesized subbands to determine the signal synthesized sub-band.

31. A system according to claim 30, characterized in that it further comprises:
basic decoder (501) located before the block (101) of the analysing filter, configured to decode the bit stream in the input signal; and
module (504) for processing high frequency reconstruction, HFR, located behind the module (405; 603) of the merger and before the block (103; 505) synthesizing filters that are configured to use the spectral bands obtained from b�tovo thread to the signal synthesized sub-band, for example, by performing the formation of the spectrum of the signal synthesized sub-band.

32. A system according to claim 30, characterized in that at least one of the processing modules of the sub-bands is a module (401) direct processing of the subbands, which is configured to define one signal synthesized sub-band signal from one of the analyzed sub-band using the conversion coefficient Q sub-bands and the stretch factor S subbands and at least one of the modules is a module (402) cross-processing sub-bands, which is configured to define one signal synthesized sub-band of the two signals of synthesized sub-bands using the conversion coefficient Q sub-bands and the stretch factor S subbands, which is independent of the first two coefficients.

33. A system according to claim 32, configured for Y=2, characterized in that the block (101) analysing applies filters to the input signal step in the analysis of time - ΔtA;
the unit of the analyzing filters is the separation of analyzed frequencies is ΔfA;
the unit of the analyzing filters contains the number N of analyzed sub-bands, wherein N>1, where n is the index of the analyzed sub-bands, n�eacham n=0...N-1; the analyzed sub-band from among N of the analyzed sub-bands associated with the frequency band of the input signal;
unit (103) synthesizing filter applies to the signal synthesized sub-band step synthesis of time - ΔtS;
block synthesizing filters has a spacing of synthesized frequencies ΔfS;
block synthesizing filters contains the number M of synthesized subbands, and M>1, where m is the index of the synthesized sub-band, and m=0...M-1; and
the synthesized sub-band from among M of synthesized sub-bands is associated with a frequency band is stretched in time and/or converted by the signal frequency,
moreover, the system is configured to deactivate at least one module (402), cross-processing, sub-band, if, for a given synthesized sub-band, met one of the following conditions:
(a) the ratio of the amplitude of the Msanalyzed sub-band member from direct source, resulting in the synthesized sub-band, and at least the amplitude of the Mwithin the optimal pair of members from a cross of the source, resulting in the synthesized sub-band, is greater than a pre-defined constant q;
(b) synthesized sub-band contains a significant contribution from the direct processing module;
c) the fundamental frequency Ω0men�e, than the spacing ΔfAfrequency analyzing unit filters.

34. A system according to any one of claims.1-3, characterized in that:
unit (101) of the analyzing filters is configured to generate x Y Z of the analyzed sub-bands of the input signal;
module (102) processing of sub-bands is configured to generate the Z signal of synthesized subbands of Y x Z signals of the analyzed sub-bands using a few values of S and Q for each group of Y signals of the analyzed sub-bands, which is based on one signal synthesized sub-band; and
unit (103) synthesizing filter configured to generate Z stretched in time and/or converted by the frequency signals from the Z signal of synthesized sub-bands.

35. A method of generating stretched in time and/or converted by the frequency signal from the input signal, and the method includes the stages at which:
get the number Y≥2 signals of the analyzed sub-bands from the input signal, and each signal is analyzed sub-band comprises the number of analyzed complex-valued discrete values, each of which has a phase and amplitude;
form Y frames consisting of L discrete input values, and each frame is extracted from the specified number of analyzed complex-valued discrete values in the signal EN�litereature sub-band, and the frame length is L>1;
apply the value of the jump from h block of discrete values to a specified number of analyzed discrete values before receiving the next frame consisting of L discrete input values, thereby generating a sequence of discrete input values;
generate on the basis of Y corresponding frames of input discrete values of the frame of the processed discrete values by determining the phase and amplitude of each of the processed discrete values of the frame, where at least one of the processed discrete values:
(i) the phase of the processed discrete values based on the respective phases of the corresponding input digital values in each of the Y frames of discrete input values; and
(ii) the amplitude of the processed values based on the amplitude of the corresponding input digital values in each of the Y frames of the input discrete values;
determine the signal synthesized sub-band by applying and addition of discrete values of a sequence of frames of the processed discrete values; and
generate stretched in time and/or converted by the frequency signal from the signal synthesized sub-band.

36. A method according to claim 35, characterized in that the frame of the processed discrete value� is based on Y=2 corresponding frames of the input of discrete values, that form by extracting discrete values from the two signals of the analyzed sub-bands representing frequencies that differ by approximately the fundamental frequency Ω0input.

37. Method to claim 35 or 36, characterized in that
the frame of the processed discrete values based on Y=2 corresponding frames of input discrete values that form by extracting discrete values from the two signals of the analyzed sub-bands, approximately representing the frequencies Ω and Ω+Ω0; and
the signal synthesized sub-band is approximately the frequency of Qφ+rΩ0where r is an integer satisfying the inequality 1≤r≤Qφ-1, and Q=(ΔtS/ΔtA)·Qφwhere ΔtAand ΔtSthe steps of analysis and synthesis of time, respectively.

38. A method according to claim 37, characterized in that the frequency Ω is chosen so that it maximizes the smallest amplitude of sub-bands of two frames of the input of discrete values, extracted from signals of the analyzed sub-bands representing the frequencies Ω and Ω+Ω0.

39. A method according to claim 38, characterized in that the amplitude of the sub-band frame input discrete value represents the amplitude of the Central or nearest to the Central discrete values.

40. Method pop.35 or 36, characterized in that said frames forming the input of discrete values includes a lower discretization analyzed complex-valued discrete values in the signals of the analyzed sub-bands.

41. A method according to claim 40, characterized in that
the frame of the processed discrete values based on Y=2 corresponding frames of input discrete values;
the first frame of input digital values extracted from the discrete values of the first signals of the analyzed sub-band along with the use of the coefficient D1downsampling;
the second frame of input digital values extracted from the discrete values of the second signals of the analyzed sub-band along with the use of the coefficient D2downsampling;
the downsampling coefficients satisfyand the inequality D1≥0, D2>0 or D1>0, D2≥0; and the phase of the processed discrete values based on a linear combination with nonnegative integer coefficients (T1, T2the respective phases of the corresponding input digital values in the first and second frames of input discrete values.

42. A method according to claim 35 or 36, characterized in that said determination of the signal synthesized sub-band advanced VC�uchet the application of a window function of finite length to each frame in the sequence of frames of the processed discrete values before overlay and addition.

43. A method according to claim 42, characterized in that the window function has a length, which corresponds to the length L of the frame, and the window function is one of the following:
the Gaussian window,
cosine window,
the raised cosine window,
Hamming window,
box Hannah,
rectangular window,
the Bartlett window, and
window Blackman.

44. A method according to claim 42, wherein the windowing function includes a number of windowed discrete values, and superimposed and folded windowed discrete values of a number of window functions when weighing by complex weighting coefficients and the shift to the value of the jump form Sh is largely consistent sequence.

45. A method according to claim 44, characterized in that the successive complex weighting coefficients differ only by a fixed phase sequence.

46. A method according to claim 45, characterized in that the phase rotation is proportional to the fundamental frequency of the input signal.

47. A method according to claim 35 or 36, characterized in that said determination of the signal synthesized sub-band includes the imposition of successive frames of the processed discrete values by applying the magnitude of the jump equal to the value h of the jump unit, multiplied by a stretching factor S subbands.

48. A method according to claim 35 or 36, characterized in that
frame obrabatyvaemykh values based on Y=2 corresponding frames of input discrete values; and
the amplitude of the processed discrete values is defined as the average value of the amplitude of the corresponding input discrete values in the first frame of input digital values and the amplitude of the corresponding input digital values in the second frame of input digital values.

49. A method according to claim 48, characterized in that the average value of the amplitudes is a geometric weighted average.

50. A method according to claim 49, characterized in that the geometric parameters weighting the amplitudes are ρ and 1-ρ, where ρ is a real number that is inversely proportional to the ratio Q transform sub-bands.

51. A method according to claim 35 or 36, characterized in that
the frame of the processed discrete values based on Y=2 corresponding frames of input discrete values; and
the phase of the processed discrete values is defined as a linear combination with nonnegative integer coefficients (T1, T2) of the respective phases of the corresponding input digital values in the first and second frames of input discrete values.

52. A method according to claim 51, characterized in that the sum of these nonnegative integer coefficients is the product Q×S stretching ratio and the conversion factor.

53. A method according to claim 51 wherein what phase of the processed discrete values corresponds to a specified linear combination of the plus option in the correction phase.

54. A method according to claim 35 or 36, characterized in that at least one of the input discrete value is obtained by interpolating two or more of the analyzed discrete values.

55. A method according to claim 35 or 36, characterized in that it further includes receiving control data that should be considered when the specified generating a frame of the processed discrete values.

56. A method according to claim 55, characterized in that
the frame of the processed discrete values based on Y=2 corresponding frames of input discrete values;
these control parameters include a fundamental frequency Ω0input signal; and
two of the analyzed sub-bands, of which extracts the input discrete value in each frame represent frequency than the fundamental frequency.

57. A method according to claim 35 or 36, characterized in that said generating a frame of the processed discrete values includes the stages on which: change the amplitude scale of at least one discrete input values; and
calculated the processed discrete value as a weighted complex productof coefficients equal to the corresponding input dis�specific value in at least two of the Y frames of the input of discrete values, at least one of the coefficients (vm=um/|um|β, m∈M≠⌀) is a discrete input value with the changed amplitude scale.

58. A method according to claim 35 or 36, characterized in that it includes a stage at which generate a number of intermediate signals of synthesized sub-bands, where each of them generates, based on the number of relevant frames of input discrete values using different values of the coefficient Q transform sub-bands and/or the stretching factor S subbands,
moreover, the definition of the signal synthesized sub-band includes the merger of the respective intermediate signals of synthesized sub-bands.

59. A method according to claim 58, characterized in that it further includes the stages on which:
decode the bit stream for the purpose of receiving input, from which to obtain the signals of the analyzed sub-bands; and
apply the information of the spectral bands obtained from the bitstream, the signal synthesized sub-band, for example, by performing the formation of the spectrum of the signal synthesized sub-band.

60. A method according to claim 58, characterized in that at least one of the intermediate signals of synthesized sub-bands generated by direct processing of the subbands on the basis� one of the signals of the analyzed sub-band using the conversion coefficient Q sub-bands and the stretch factor S subbands, and at least one of the intermediate signals of synthesized sub-bands generated by processing the cross-pieces on the basis of two signals of synthesized sub-bands using the conversion coefficient Q sub-bands and the stretch factor S subbands, which are independent from the first two coefficients.

61. A method according to claim 60, characterized in that said generating the intermediate signal synthesized sub-band by processing the cross-pieces is suspended in response to satisfaction of one of the following:
(a) the ratio of the amplitude of the Msanalyzed sub-band member from direct source, resulting in the synthesized sub-band, and at least the amplitude of the Mwithin the optimal pair of members from a cross of the source, resulting in the synthesized sub-band, is greater than a pre-defined constant q;
(b) synthesized sub-band contains a significant contribution from the direct processing module;
c) the fundamental frequency Ω0less than the frequency spacing ΔfAunit of the analyzing filters.

62. A method according to claim 35 or 36, characterized in that the get the x Y Z of the analyzed sub-bands;
form Y x Z input of discrete values;
to generate Z frames processed discrete values, use x Y Z �sootvetstvujushij frames input of discrete values;
define Z signals of synthesized sub-bands; and
generate Z stretched in time and/or converted frequency signals.

63. Media data in the memory which are stored machine-readable commands intended to perform the method according to claim 35 or 36.



 

Same patents:

FIELD: physics, computer engineering.

SUBSTANCE: present invention relates to signal processing means. An encoder sets an interval including 16 frames as interval section to be processed, outputs high-frequency band encoded data to obtain the high-frequency band component of an input signal and low-frequency band encoded data obtained by encoding the low-frequency band signal of the input signal for each section to be processed. In this case, for each frame, a coefficient used in estimation of the high-frequency band component is selected and the section to be processed is divided into continuous frame segments including continuous frames from which the coefficient with the same section to be processed is selected. In addition, high-frequency band encoded data are produced which include data including information indicating the length of each continuous frame segment, information indicating the number of continuous frame segments included in the section to be processed and a coefficient index indicating the coefficient selected in each continuous frame segment.

EFFECT: improved sound quality with frequency band expansion.

23 cl, 51 dwg

FIELD: radio engineering, communication.

SUBSTANCE: invention relates to data transmission and is intended for encoding and decoding data. A bandpass filter splits an input signal into multiple-subband signals; a parameter calculating means calculates a parameter value using a selected multiple-subband signal and/or the input signal; a high-frequency subband power estimating means calculates a high-frequency subband power estimate based on the calculated parameter value; a high-frequency signal component generating means generates a high-frequency signal component based on multiple-subband signals selected by the bandpass filter, and the high-frequency subband power estimate calculated by high-frequency subband power estimating means. A frequency band extension apparatus extends the frequency band of the input signal using the high-frequency signal component.

EFFECT: high accuracy of reproducing data and high quality of sound owing to frequency band extension.

36 cl, 30 dwg

FIELD: physics, acoustics.

SUBSTANCE: group of inventions relates to means of analysing time variations of audio signals. Disclosed is an apparatus for obtaining a parameter describing variation of a signal characteristic of a signal based on actual transform-domain parameters describing an audio signal in transform-domain which includes a parameter determiner. The parameter determiner is configured to determine one or more model parameters of a transform-domain variation model describing evolution of the transform-domain parameters depending on one or more model parameters representing a signal characteristic, such that a model error, representing deviation between a modelled temporal evolution of the transform-domain parameters and evolution of the actual transform-domain parameters, is brought below a predetermined threshold value or minimised.

EFFECT: designing highly reliable means for obtaining a parameter describing time variation of a signal characteristic.

27 cl, 9 dwg

FIELD: information technology.

SUBSTANCE: audio signal decoder designed to provide a decoded representation of an audio signal based on an encoded representation of the audio signal, which includes information on evolution of a temporary deformation loop, includes a temporary deformation loop computer, a device for changing the scale of the temporary deformation loop data and a deformation decoder. The temporary deformation loop computer is designed to generate temporary deformation loop data through multiple restarting from a predefined starting value of the temporary deformation loop based on information on evolution of the temporary deformation loop, which describes time evolution of the temporary deformation loop. The device for changing the scale of temporary deformation loop data is designed to change the scale of at least part of temporary deformation loop data to avoid, reduce or eliminate non-uniformity during restart in a scaled version of the temporary deformation loop. The deformation decoder is designed to provide a decoded representation of an audio signal based on an encoded representation of the audio signal and by using the scaled version of the temporary deformation loop.

EFFECT: supporting low bit rate with reliable reconstruction of the required temporary deformation information at the decoder side.

14 cl, 40 dwg

FIELD: information technology.

SUBSTANCE: disclosed are a context-based arithmetic encoding apparatus and method and a context-based arithmetic decoding apparatus and method. The context-based arithmetic decoding apparatus may determine a context of a current tuple of N elements to be decoded, determine a most significant bit (MSB) context corresponding to an MSB symbol of the current tuple of N elements, and determine a probability model using the context of the tuple of N elements and the MSB context. The context-based arithmetic decoding apparatus may then perform decoding for a MSB based on the determined probability model, and perform decoding for a least significant bit (LSB) based on a bit depth of the LSB derived from the decoding process for an escape code.

EFFECT: high efficiency of encoding and reducing the amount of memory required for encoding.

79 cl, 29 dwg

FIELD: information technology.

SUBSTANCE: signal is subjected to high- and low-frequency filtration. After high-frequency filtration, the signal is decimated and after low-frequency filtration, the signal is correlated with unit readings of a given maximum frequency, followed by reduction of frequency and repeated search until minimum frequency of the readings is achieved, with comparison and storage of values of the correlation function.

EFFECT: simple algorithms for implementing the method, high speed of operation, lower requirements for computational power of hardware resources.

5 dwg

FIELD: information technology.

SUBSTANCE: in one version, a method of processing signals includes encoding a low-frequency portion of a speech signal into at least an encoded narrowband excitation signal and a plurality of narrowband filter parameters; and generating a highband excitation signal based on a narrowband excitation signal. The narrowband excitation signal is based on an encoded narrowband excitation signal. The method also includes encoding a high-frequency portion of a speech signal into at least a plurality of highband filter parameters according to at least a highband excitation signal. The encoded narrowband excitation signal includes a time warping and the method includes applying a time shift to the high-frequency portion based on information related to the time warping.

EFFECT: improved method.

FIELD: information technology.

SUBSTANCE: apparatus for processing an audio signal which contains a transient signal includes a transient signal removing unit (100), a processor (110) and a signal inserting module (120) which inserts part of the audio signal into the processed audio signal at that point where the transient signal was removed before the step of processing by said unit; thus the processed audio signal contains a transient signal which was not altered during processing; vertical conformity of the transient signal does not go through the step of processing by the processor (110) which might disrupt it.

EFFECT: improved quality when processing an audio signal.

15 cl, 17 dwg

FIELD: information technology.

SUBSTANCE: band-pass filters 13 obtain a plurality of subband signals from an input signal. A frequency envelope extracting circuit 14 extracts a frequency envelope from the plurality of subband signals. A highband signal generating circuit 15 generates highband signal components on the basis of the frequency envelope obtained by the frequency envelope extracting circuit 14, and the plurality of subband signals obtained by the band-pass filters 13. A frequency band extension apparatus 10 extends the frequency band of the input signal on the basis of the highband signal components generated by the highband signal generating circuit 15. The highband signal generating circuit includes a gain calculating circuit that finds a gain for each subband from the frequency envelope and applies the gain to the plurality of subband signals. The gain is calculated by a mapping function obtained by performing learning in advance with a wide-band signal as teacher data.

EFFECT: higher sound quality of reproducing a music signal owing to frequency band extension.

26 cl, 17 dwg

FIELD: information technology.

SUBSTANCE: speech transmission method involves time warping a residual low-band speech signal into a stretched or compressed version of the residual low-band speech signal, time warping a high-band speech signal into a stretched or compressed version of the high-band speech signal and combined the time warped low-band and high-band speech signals to obtain a complete time warped speech signal. Time warping of the high-band speech signal includes determination of a set of periods of the fundamental tone from the low-band speech signal, using periods of the fundamental tone from the low-band speech signal and overlapping/summation of one or more periods of the main tone if the high-band speech signal is compressed, and overlapping/summation or repetition of one or more periods of the fundamental tone if the high-band speech signal is stretched. The method may also involves steps on which speech segments are classified and linear predictive coding with code excitation, linear predictive coding with noise excitation or 1/8 frame (pause) coding is carried out.

EFFECT: high quality of time warping frames and reducing the computational load.

51 cl, 10 dwg

FIELD: physics, acoustics.

SUBSTANCE: invention relates to encoding acoustic signals and can be used for transportation in the frequency domain. The high frequency audio signal generator consists of an analyser for analysing an input signal to determine current transient information, a spectral converter for converting the input signal into an input spectral representation, a spectral processor for processing the input spectral representation to obtain a modified spectral representation comprising frequency values higher than the input spectral representation, a time converter for converting the modified spectral representation to a time domain representation, wherein the spectral converter or the time converter is capable of performing frequency domain oversampling for the first component of the input signal having transient information and not performing the frequency domain oversampling for the second component of the input signal without transient features.

EFFECT: efficient generation of a quality high frequency audio signal through separate processing of transient and fixed components of the audio signal.

15 cl, 7 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to radio engineering and is intended for controlling an audio signal, including a transient event. The device comprises a unit for replacing a transient signal, configured to replace the transient part of a signal, which includes a transient event of an audio signal, with part of a replacement signal adapted to energy characteristics of the signal of one or more transient parts of the audio signal, or to the energy characteristic of the signal of the transient part of the signal to obtain an audio signal with a shorter transient process. The device also includes a signal processor configured to process an audio signal with a shorter transient process to obtain a processed version of the audio signal with a shorter transient process. The device also includes a transient signal inserting unit configured to merge the processed version of the audio signal with a shorter transient process with the transient signal, representing in the original or processed form the transient content of the transient part of the signal.

EFFECT: high accuracy of reproducing the signal.

14 cl, 20 dwg

FIELD: physics, acoustics.

SUBSTANCE: group of inventions relates to means of analysing time variations of audio signals. Disclosed is an apparatus for obtaining a parameter describing variation of a signal characteristic of a signal based on actual transform-domain parameters describing an audio signal in transform-domain which includes a parameter determiner. The parameter determiner is configured to determine one or more model parameters of a transform-domain variation model describing evolution of the transform-domain parameters depending on one or more model parameters representing a signal characteristic, such that a model error, representing deviation between a modelled temporal evolution of the transform-domain parameters and evolution of the actual transform-domain parameters, is brought below a predetermined threshold value or minimised.

EFFECT: designing highly reliable means for obtaining a parameter describing time variation of a signal characteristic.

27 cl, 9 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to computer engineering. A method of maintaining speech audibility in a multi-channel audio signal, comprising comparing a first characteristic and a second characteristic of the multi-channel audio signal to generate an attenuation factor, wherein the first characteristic corresponds to a first channel of the multi-channel audio signal that contains speech audio and non-speech audio, wherein the second characteristic corresponds to a second channel of the multi-channel audio signal that contains predominantly non-speech audio; adjusting a gain applied to a second power spectrum until the predicted speech intelligibility meets a criterion; and using the adjusted gain as the attenuation factor once the predicted speech intelligibility meets the criterion; adjusting the attenuation factor according to a speech likelihood value to generate an adjusted attenuation factor; and attenuating the second channel using the adjusted attenuation factor.

EFFECT: improved speech audibility in a multi-channel audio signal.

14 cl, 5 dwg

FIELD: radio engineering, communication.

SUBSTANCE: method of picking up speech signal in presence of interference, which comprises converting an input mixture of an acoustic signal and interference into an electrical signal, filtering with a band-pass filter to obtain a mixture of a speech signal and interference with a given bandwidth, which is amplified in a low frequency amplifier; an analogue-to-digital converter (ADC) generates readings of the mixture of the signal and interference in digital form and transmits said readings to a computing device, which forms pairs of sums of amplitudes of the readings in a certain manner and calculates signal amplitudes for each moment in time using the obtained summation results by solving corresponding systems of linear equations.

EFFECT: high efficiency of picking up a speech signal in the presence of interference.

2 dwg, 1 tbl

FIELD: physics, acoustics.

SUBSTANCE: invention relates to HFR (High Frequency Reconstruction/Regeneration) of audio signals and is intended for performing HFR of audio signals having large variations in energy level across the low frequency range which is used to reconstruct the high frequencies of the audio signal. The system configured to generate a plurality of high frequency subband signals covering a high frequency interval from a plurality of low frequency subband signals. The system comprises means of receiving a plurality of low frequency subband signals; means of receiving a set of target energies, each target energy covering a different target interval within the high frequency interval and being indicative of the required energy of one or more high frequency subband signals lying within the target interval; means of generating a plurality of high frequency subband signals from the plurality of low frequency subband signals and from a plurality of spectral gain coefficients associated with the plurality of low frequency subband signals, respectively; and means of adjusting the energy of the plurality of high frequency subband signals using the set of target energies.

EFFECT: preventing undesirable noise caused by discontinuities of the spectral envelope of the high frequency audio signal.

20 cl, 14 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to means of generating a broadband signal using a low-bandwidth input signal. A processor performs controlled bandwidth expansion using a low-bandwidth input signal and a first set of parameters for generating first frequency content, which continues up to the a first frequency, and performs blind bandwidth expansion using the first frequency content and a second set of parameters for generating second frequency content, which continues up to a second frequency which is higher than the first frequency. The first set of parameters and the input low-bandwidth signal are extracted from the bit stream. The processor comprises a parameter generator for generating the second set of parameters from the first frequency content, wherein the parameter generator is configured to obtain spectral envelope parameters for the second set of parameters for the second frequency content via extrapolation from lower to higher frequencies of information about the energy of the formed spectral envelope of the first frequency content.

EFFECT: wider bandwidth with low bit rate and maintaining high signal quality.

13 cl, 7 dwg

FIELD: physics, acoustics.

SUBSTANCE: invention relates to means of decoding and/or transcoding audio. A first and a second source set of spectral band replication (SBR) parameters are merged into a target set of SBR parameters. The first and second source set comprise a first and second frequency band partitioning, respectively, which are different from one another. The first source set comprises a first set of energy related values associated with frequency bands of the first frequency band partitioning. The second source set comprises a second set of energy related values associated with frequency bands of the second frequency band partitioning. The target set comprises a target set of energy related values associated with an elementary frequency band. The method comprises steps of breaking up the first and the second frequency band partitioning into a joint grid comprising the elementary frequency band; assigning a first value of the first set of energy related values to the elementary frequency band; assigning a second value of the second set of energy related values to the elementary frequency band; and combining the first and second value to yield the target energy related value for the elementary frequency band.

EFFECT: simplifying the process of reducing the number of channels while preserving the relevant high-frequency channel information.

32 cl, 9 dwg

FIELD: physics, acoustics.

SUBSTANCE: invention relates to means of generating an equalised multichannel audio signal. An audio encoder for obtaining an output signal using an input audio signal comprises a patch generator, a comparator and an output interface. The patch generator generates at least one bandwidth extension signal, having a high-frequency band. The high-frequency band of the bandwidth extension signal is based on a low frequency band of the input audio signal. A comparator calculates a plurality of comparison parameters. A comparison parameter is calculated based on a comparison of the input audio signal and a generated bandwidth extension signal. Each comparison parameter of the plurality of comparison parameters is calculated based on a different offset frequency between the input audio signal and a generated high bandwidth signal. Further, the comparator determines a comparison parameter from the plurality of comparison parameters, wherein the determined comparison parameter satisfies a predefined criterion.

EFFECT: improved signal encoding quality at high bit rate.

17 cl, 22 dwg

FIELD: radio engineering, communication.

SUBSTANCE: invention relates to means of filtering a multichannel audio signal, having a speech channel and at least one non-speech channel. The method includes determining at least one attenuation control value which serves as a feature of the extent of similarity between speech-related content which is defined by the speech channel and speech-related content which is defined by the non-speech channel; attenuating the non-speech channel in response to at least one attenuation control value; scaling the raw attenuation control signal (e.g. a gain control signal with suppression of a weak signal with a stronger signal) for the non-speech channel in response to at least one attenuation control value.

EFFECT: high speech intelligibility defined by a signal.

66 cl, 7 dwg

FIELD: physics, computer engineering.

SUBSTANCE: present invention relates to signal processing means. An encoder sets an interval including 16 frames as interval section to be processed, outputs high-frequency band encoded data to obtain the high-frequency band component of an input signal and low-frequency band encoded data obtained by encoding the low-frequency band signal of the input signal for each section to be processed. In this case, for each frame, a coefficient used in estimation of the high-frequency band component is selected and the section to be processed is divided into continuous frame segments including continuous frames from which the coefficient with the same section to be processed is selected. In addition, high-frequency band encoded data are produced which include data including information indicating the length of each continuous frame segment, information indicating the number of continuous frame segments included in the section to be processed and a coefficient index indicating the coefficient selected in each continuous frame segment.

EFFECT: improved sound quality with frequency band expansion.

23 cl, 51 dwg

Up!