BR122019025118B1 - SYSTEM AND METHOD FOR GENERATING A TIME EXTENDED SIGNAL AND/OR A FREQUENCY TRANSPOSED SIGNAL FROM AN INPUT SIGNAL AND COMPUTER READABLE STORAGE MEDIUM - Google Patents

Abstract

The present invention relates to an efficient implementation of vector product-enhanced high-frequency reconstruction (HFR), wherein a new component in frequency is generated on the basis of existing components in and . The invention provides a block-based harmonic transposition, in which a time block of complex subrange samples is processed with a common phase modification. The superposition of several modified samples has the net effect of limiting undesirable intermodulation products, thereby enabling coarser frequency resolution and/or a lower degree of oversampling to be used. In one embodiment, the invention further includes a window function suitable for use with block vector product enhanced HFR. A hardware embodiment of the invention may include an analysis filter bank (101), a control data-configurable subrange processing unit 102 (104), and a synthesis filter bank (103).

Description

[0001] Split from BR112013005676-2 deposited on September 5, 2011.

TECHNICAL FIELD

[0002] The present invention relates to audio source coding systems that make use of a harmonic transposition method for high frequency reconstruction (HFR), for digital effects processors, such as exciters; which generate harmonic distortion to add brightness to a processed signal, and for time stretchers which extend a signal duration with maintained spectral content.

BACKGROUND OF THE INVENTION

[0003] In WO98/57436 the concept of transposition was established as a method for recreating a high frequency range from a low frequency range of an audio signal. Substantial bitrate savings can be achieved by using this concept in audio coding. In an HFR-based audio coding system, a low bandwidth signal is presented to a center waveform encoder and the higher frequencies High frequencies are regenerated using transposition and additional side information from the very low bit rate describing the target spectral shape on the decoder side. For low bit rates where the bandwidth of the central encoded signal is narrow, it becomes increasingly important to recreate a high band with perceptually pleasing characteristics. The harmonic transposition defined in WO98/57436 works very well for complex musical material in a low transition frequency situation.

[0004] The principle of a harmonic transposition is that a sinusoid with frequency is mapped to a sinusoid with frequency where is an integer defining the order of transposition. In contrast to this, an HFR-based single-sideband (SSB) modulation maps a sinusoid with frequency for a sinusoid with frequency where is a fixed frequency deviation. Given a center signal with low bandwidth, a dissonant ringing artifact will result from SSB transposition.

[0005] In order to achieve the best possible audio quality, prior art methods of high quality HFR harmonics employ complex modulated filter banks with very fine frequency resolution and a high degree of oversampling to achieve audio quality. required. Fine resolution is necessary to avoid undesirable intermodulation distortion arising from the nonlinear treatment of sinusoid sums. With sufficiently narrow subbands, high-quality methods aim to have at most one sinusoid in each subband. A high degree of time oversampling is necessary to avoid synonym-like distortion, and a certain degree of frequency oversampling is necessary to avoid pre-echoes for transient signals. The obvious disadvantage is that the computational complexity becomes very high.

[0006] Another common disadvantage associated with harmonic transposers becomes evident for signals with a prominent periodic structure. Such signals are superpositions of sinusoids harmonically related to frequencies where is the fundamental frequency.

[0007] After the harmonic transposition of order the output sinusoids have frequencies which, in the case of it is just a strict subset of the desired full harmonic series. In terms of audio quality resulting "phantom" tone for the fundamental transposed frequency will normally be noticed. Harmonic transposition often results in a "metallic" sound character of the encoded and decoded audio signal.

[0008] In WO2010/081892, which is incorporated herein by reference, the vector product method was developed to handle the above problem of phantom tilt degree in the case of high-quality transposition. Given or transmitted partial or complete information about the value of the fundamental frequency of the dominant harmonic part of the signal to be transposed with greater fidelity, the nonlinear subrange modifications are supplemented with nonlinear combinations of at least two different analysis subranges, where the Distances between the analysis subband indices are related to the fundamental frequency. The result is to regenerate the missing partials in the transposed output, which happens, however, at considerable computational cost.

[0009] In view of the above-mentioned deficiencies of the available HFR methods, it is an object of the present invention to provide a more efficient implementation of the improved HFR vector product. In particular, it is an objective to provide such a method allowing high fidelity audio reproduction with reduced computational effort compared to available techniques.

[0010] The present invention achieves at least one of these objectives by providing devices and methods, as set forth in the embodiments.

[0011] In a first aspect, the invention provides a system configured to generate a time-stretched signal and/or frequency transposed signal from an input signal. The system comprises: a bank of analysis filters configured to generate Y number of subrange signal analyses.

[0012] From the input signal, wherein each analysis subrange signal comprises a plurality of complex-valued analysis samples, each having a phase and a magnitude; a subrange processing unit configured to determine a synthesis subrange signal from the analysis subrange Y signals using a subrange transposition factor and a subrange stretching factor at least one and S being greater than one, wherein the subrange processing unit comprises: a block extractor configured to: i) form Y frames from L input samples, each frame being extracted from said plurality of complex valued analysis samples in an analysis subrange signal and the frame length being L >I; and ii) applying a block jump size of h samples to said plurality of analysis samples, prior to forming a subsequent frame of L input samples, thereby generating a sequence of frames of input samples; a non-linear frame processing unit configured to generate, based on Y corresponding frames of input samples formed by the extractor block, a frame of processed samples by determining a phase and magnitude for each processed sample of the frame, in that, for at least one processed sample: i) the phase of the processed sample is based on the respective phases of the corresponding input sample in each of the Y input sample frames, and ii) the magnitude of the processed sample is based on the magnitude of the corresponding input sample in each of the Y input sample frames, and

[0013] An overlay and addition unit configured to determine the synthesis subrange signal by overlaying and adding the samples from a sequence of processed sample frames, and

[0014] a bank of synthesis filters configured to generate the time elongated and/or frequency transposed signal of the synthesis subband signal.

[0015] The system can be operable for any positive integer value of Y. However, it is operable for at least Y = 2.

[0016] In a second aspect, the invention provides a method for generating a time elongated and/or frequency transposed signal from an input signal. The method comprises: extracting a series of Y>2 analysis subrange signals from the input signal, wherein each analysis subrange signal comprises a plurality of complex valued analysis samples, each having a phase and a magnitude; forming Y frames from L input samples, each frame being extracted from said plurality of complex-valued analysis samples into an analysis subrange signal and the length of the frame being L>1; applying a block jump size of h samples to said plurality of analysis samples, prior to extracting a subsequent frame of L input samples, thereby generating a sequence of frames of input samples; generating, based on Y corresponding frames of input samples, a frame of processed samples by determining a phase and magnitude for each processed sample in the frame, where, for at least one processed sample:

[0017] The phase of the processed sample is based on the respective phases of the corresponding input sample in at least one of the Y input sample frames, and the magnitude of the processed sample is based on the magnitude of the corresponding input sample in each of the Y frames of input samples;

[0018] Determination of the synthesis subrange signal by superposition and addition of samples from a sequence of processed sample frames, and generation of the time-stretched signal and/or the transposed frequency signal of the synthesis subrange signal.

[0019] Here, Y is an arbitrary integer greater than one. The system according to the first aspect is operable to carry out the method at least for Y = 2.

[0020] A third aspect of the invention provides a computer program product including a computer-readable medium (or data carrier) storing software instructions for causing a programmable computer to execute the method according to the second aspect.

[0021] The invention is based on the finding that the general concept of an improved HFR vector product will provide better results when data is processed arranged in blocks of complex subrange samples. Among other things, this makes it possible to apply frame-shaped phase compensation to samples, which has been found to reduce intermodulation products in some situations. Furthermore, it is possible to apply a magnitude adjustment, which can lead to similar advantageous effects. The implementation of the improved HFR vector product invention includes subrange block-based harmonic transposition, which can significantly reduce inter-modulation products. Thus, a filter bank with a coarser frequency resolution and/or a smaller degree of oversampling (such as a QMF filter bank) can be used while preserving high output quality. In subrange block-based processing, a time block of complex subrange samples is processed with a common phase modification, and the superposition of several modified samples to form an output subrange sample has the net effect of suppressing phase products. intermodulation which would otherwise occur when the subrange input signal consists of several sinusoids. Transposition based on subrange block-based processing has much lower computational complexity than high-resolution transposers and achieves almost the same quality for many signals.

[0022] For the purpose of this description, it is noted that in embodiments in which Y> 2, the non-linear processing unit uses as input Y "corresponding" frames of input samples in the sense that the frames are synchronous or close to synchronous. For example, samples in respective frames may be related to time steps that have substantial time overlap between frames. The term "corresponding" is also used with respect to samples to indicate that they are synchronous or approximately so. Furthermore, the term "frame" will be used as a synonym for "block". Consequently, the "block jump size" may be equal to the frame length (possibly adjusted with respect to downsampling, if such is applied) or may be smaller than the frame length (possibly adjusted with respect to downsampling). resolution, if applicable), in which case consecutive frames overlap, in the sense that an input sample may belong to more than one frame. The system does not necessarily generate all processed samples in a frame, by identifying their phase and magnitude based on the phase and magnitude of all Y corresponding frames of the input samples; Without departing from the invention, the system can generate the phase and/or magnitude of the same processed samples based only on a smaller number of input samples.

[0023] In one embodiment, the analysis filter bank is a mirror quadrature filter bank (QMF) or pseudo QMF bank with any number of taps and points. It could, for example, be a 64-point QMF bank.

[0024] The bank of analysis filters can also be chosen from the class of discrete Fourier transforms with windows or a small wave transform. Advantageously, the synthesis filter bank coincides with the analysis filter bank by being, respectively, an inverse QMF bank, an inverse QMF pseudo-bank, etc. It is known that such filter banks can have relatively coarse frequency resolution and/or a relatively low degree of oversampling. Unlike the prior art, the invention can be realized using such relatively simple components, without necessarily suffering from a decrease in output quality; Such embodiments consequently represent an economic advantage over the prior art; in one embodiment, one or more of the following is true for the analysis filter bank: an analysis time period is an analysis frequency spacing is • the analysis filter bank includes analysis subranges N> 1 indexed by an analysis subrange index n = 0,..., N - 1; • an analysis subband is associated with a frequency range of the input signal.

[0025] In one embodiment, one or more of the following is true for the synthesis filter bank: a synthesis time period is a synthesis frequency spacing is • the synthesis filter bank includes synthesis subranges M> 1 indexed by a synthesis subrange index m=0,... M - 1; • a synthesis subband is associated with a frequency band of the time-stretched signal and/or transposed frequency.

[0026] In one embodiment, the non-linear frame processing unit is adapted to input two frames (Y-2) in order to generate a frame from the processed samples, and the subrange processing unit includes a subrange control unit. cross processing to generate cross processing control data. Therefore, by specifying the quantitative and/or qualitative characteristics of subrange processing, the invention achieves flexibility and adaptability. The control data may specify subbands (e.g., identified by indices) that differ in frequency by a fundamental frequency of the input signal. In other words, the indices that identify the subbands can be different by an integer approximating the ratio of such fundamental frequency divided by the analysis frequency spacing. This will lead to a psycho-acoustically pleasing output, as the new spectral components generated by the harmonic transposition will be compatible with the natural harmonic series.

[0027] In a further development of the previous embodiment, the analysis (input) and (output) synthesis subrange indices are chosen so as to satisfy equation (16) below. One parameter that appears in this equation makes it applicable to both banks of oddly and evenly stacked filters. When the subrange indices obtained as an approximate solution (e.g., least squares) to equation (16), the new spectral component obtained by harmonic transposition will likely be to be compatible with the natural harmonic series. Thus, HFR will be likely to provide a faithful reconstruction of an original signal which has had its high frequency content removed.

[0028] Another development of the previous embodiment provides a way to select the parameter r appearing in equation (16) and which represents the order of the vector product transposition. Given an output subrange index m, each value of transposition order r will determine two analysis subrange indexes n1, n2. This further development evaluates the magnitudes of the two subranges for a series of r options and selects the value that maximizes the minimum of the two magnitudes of the analysis subranges. This way of index selection can avoid the need to restore sufficient magnitude by amplifying weak components of the input signal, which can lead to poor output quality. In this regard, subrange magnitudes can be calculated in a manner known per se, such as by square rooting the squared input samples forming a frame (block) or part of a frame. The subrange magnitude can also be calculated as a magnitude from a central sample, or near the center in a frame. This calculation can indicate a simple, but adequate measure of magnitude.

[0029] In a further development of the previous embodiment, a synthesis subrange can receive contributions from instances of harmonic transposition according to both direct processing and vector product-based processing. In this regard, decision criteria can be applied to determine whether a particular possibility of regenerating a partial fault by vector product-based processing should be used or not. For example, this further development may be adapted to refrain from using a subrange cross-processing unit if one of the following conditions is met: the ratio of the magnitude Ms of the direct source term analysis subrange producing the subrange of synthesis and the smallest magnitude Mc in an optimal pair of cross-source terms producing the synthesis subrange is greater than a predetermined constant. the synthesis subrange already receives a significant contribution from a direct processing unit; a fundamental frequency is smaller than the analysis filter bank spacing

[0030] In one embodiment, the invention includes decreasing the resolution (decimation) of the input signal.

[0031] In effect, one or more of the input sample frames can be determined by decreasing the resolution of the complex-valued analysis samples by a subrange, as can be performed by the block extractor.

[0032] In a further development of the previous embodiment, the resolution reduction factors that must be applied satisfy equation (15) below. Not both downsampling factors are allowed to be zero, as this corresponds to a trivial case. Equation (15) defines a relationship between the resolution decrease factors D 1, D2, with the subrange stretching factor S and the subrange transposition factor :'=•?, and further with the phase coefficients T1 , T2 that appear in an expression (13) to determine the phase of a processed sample.

[0033] This ensures a correspondence between the phase of the processed samples, with the other components of the input signal, to which the processed samples must be added.

[0034] In one embodiment, the processed sample frames are provided with windows before being overlaid and added. A windowing unit can be adapted to apply a finite length window function to the processed sample frames. Appropriate window functions are enumerated in the embodiments.

[0035] The inventor realized that vector product methods of the type disclosed in WO2G1G/081892 are not fully compatible with processing techniques based on subrange blocks from the outset. Although such a method can be satisfactorily applied to one of the subrange samples in a block, it may lead to distortion if it were simply extended to the other samples in the block. To this end, one embodiment applies window functions comprising window samples that sum - when weighted by complex weights and offset by a jump size - to a substantially constant sequence. The jump size can be the product of the block jump size h and the subband stretching factor S. Using such window functions reduces the impact of distortion artifacts. Alternatively or additionally, such window functions may also allow other measures to reduce artifacts, such as phase rotations of the processed samples.

[0036] Preferably, complex consecutive weights, which are applied to evaluate the condition in the window samples, differ only by a fixed phase rotation. Still preferably, said fixed phase rotation is proportional to the fundamental frequency of the input signal. The phase rotation may also be proportional to the transposition order of the vector product to be applied and/or the physical transposition parameter and/or the difference in the resolution reduction factors and/or the analysis time period. The phase rotation can be given by equation (21), at least in an approximate sense.

[0037] In one embodiment, the present invention enables harmonic transposition of improved harmonic transposition vector product by modifying the opening of synthesis windows in response to a fundamental frequency parameter.

[0038] In one embodiment, successive frames of processed samples are added with a certain overlap. To achieve adequate overlap, the processed frame frames are appropriately shifted by a jump size that is the block jump size h raised by the subband stretch factor S. Thus, if the overlap of consecutive frames of input samples is L - h, then the overlap of consecutive frames of processed samples can be S(L - h).

[0039] In one embodiment, the system according to the invention is operable to generate not only a processed sample based on Y = 2 input samples, but also based on Y = 1 single sample.

[0040] Thus, the system can regenerate missing partials not only through a vector product-based approach (such as by equation (13)), but also by a direct subrange approach (such as by equation (5) or ( 11)). Preferably, a control unit is configured to control the operation of the system, including which approach is to be used to regenerate a particular missing partial.

[0041] In a further development of the preceding embodiment, the system is further adapted to generate a processed sample, based on more than three samples, i.e., for Y > 3. For example, a processed sample that can be obtained by multiple instances of vector product-based harmonic transposition may contribute to a processed sample, by multiple instances of direct subrange processing, or by a combination of vector product transposition and direct transposition. This transposition method adaptation option provides a powerful and versatile HFR. Therefore, in this embodiment it is operable to carry out the method according to the second aspect of the invention, for Y = 3, 4, 5 etc..

[0042] One embodiment is configured to determine a processed sample as a complex number that has a magnitude that is the average value of the respective magnitudes of the corresponding input samples. The average value can be an arithmetic mean (weighted), geometric mean (weighted), or harmonic mean (weighted) of two or more input samples. In the case Y = 2, the average is based on two complex input samples. Preferably, the magnitude of the processed sample is a weighted geometric value. More preferably, the geometric value is weighted using the parameters- and 1 - , as in equation (13). Here, the weighting parameter of the geometric magnitude is a real number inversely proportional to the transposition factor Q of the subrange. The parameter it may also be inversely proportional to the stretching factor S.

[0043] In one embodiment, the system is adapted to terminate a processed sample as a complex number having a phase that is a linear combination of the respective phases of corresponding input samples in the input sample frames. In particular, the linear combination may comprise phases relative to two input samples (Y=2). The linear combination of two phases can apply to non-zero integer coefficients, the sum of which is equal to the stretching factor S multiplied by the subrange transposition factor Q. Optionally, the phase obtained by such linear combination is further adjusted by a fixed phase correction parameter. The phase of the processed sample can be given by equation (13).

[0044] In one embodiment, the block extractor (or an analogous step in a method according to the invention) is adapted to interpolate two or more analysis samples from an analysis subrange signal in order to obtain a Input sample that will be included in a frame (block). Such interpolation can allow synthesis of the input signal by a non-integer factor. The analysis samples to be interpolated may or may not be consecutive.

[0045] In one embodiment, the subrange processing configuration may be controlled by control data provided from outside the unit performing the processing. The control data may be related to momentary acoustic properties of the input signal. For example, the system itself may include a section adapted to determine momentary acoustic properties of the signal, such as the fundamental (dominant) frequency of the signal. Knowledge of the fundamental frequency provides guidance in selecting the analysis sub-ranges from which processed samples should be obtained. Suitably, the spacing of the analysis subbands is proportional to such fundamental frequency of the input signal. As an alternative, control data may also be provided from outside the system, preferably by being included in an encoding format suitable for transmission as a bit stream over a digital communications network. In addition to control data, such a coding format may include information relating to lower frequency components of a signal (e.g., components in pos. 701 in Figure 7).

[0046] However, in the interest of saving bandwidth the preferred format does not include complete information relating to the higher frequency components, (pos. 702), which can be regenerated by the invention. The invention may in particular provide a decoding system with a control data receiving unit configured to receive such control data if included in a received bit stream that also encodes the input signal or received as a separate signal or stream. of bits.

[0047] One embodiment provides a technique for efficiently carrying out calculations caused by the method of the invention. To this end, a hardware implementation may include a pre-normalizer to scale the magnitudes of the corresponding input samples in some of the Y frames on which a frame of processed samples is to be based. After such resizing, a processed sample can be calculated as a complex (weighted) product of the resized and possibly unrescaled input samples. An input sample that appears as a rescaled factor in the product must normally not reappear as an unscaled factor. With the possible exception of the phase correction parameter θ, it is possible to evaluate equation (13) as a product of complex (possibly rescaled) input samples. This represents a computational advantage compared to separate treatments of the magnitude and phase of a processed sample.

[0048] In one embodiment, a system configured for the case of Y = 2 comprises two block extractors adapted to form a frame of input samples from each, in parallel operation.

[0049] In a further development of embodiments representing Y > 3, a system may comprise a plurality of subrange processing units, each of which is configured to determine a synthesis subrange intermediate signal using a subrange transposition factor and/or a different subrange stretching factor and/or transposition method differing in that it is based on vector or direct product. The subrange processing units may be arranged in parallel, for parallel operation in this embodiment, the system may further comprise a merging unit disposed downstream of the subrange processing units and upstream of the synthesis filter bank. The blending unit may be adapted to combine (e.g., by mixing together) corresponding intermediate synthesis subrange signals to obtain the synthesis subrange signal. As already mentioned, the intermediate synthesis subrange that is merged may have been obtained both based on direct product and cross product harmony transposition. A system according to the embodiment may further comprise a decoder core for decoding a stream of bits in an input signal. It may also comprise an HFR processing unit adapted to apply spectral range information, in particular by performing spectral modeling. The operation of the HFR processing unit can be controlled by information encoded in the bit stream.

[0050] One embodiment provides HFR of multidimensional signals, for example, in an audio reproduction system in a stereo format, comprising Z channels such as left, right, center, surround, etc.. In one possible implementation for the processing an input signal with a plurality of channels, the processed samples of each channel are based on the same number of input samples, although the stretching factor S and the transposition factor Q for each track may vary between channels. To this end, the implementation may comprise an analysis filter bank for producing analysis subrange Y signals from each channel, a subrange processing unit for generating the Z subrange signals, and a synthesis filter bank for generating elongated time and/or transposed frequency signals which form the output signal.

[0051] In variations of the previous embodiment, the output signal may comprise output channels that are based on different numbers of analysis subrange signals. For example, it may be advisable to dedicate a greater amount of computational resources to HFR of acoustically important channels; for example, channels to be reproduced by audio sources located in front of a listener may be favored over surround or rear channels.

[0052] It is emphasized that the invention refers to all combinations of the above characteristics, even if these are reported in different embodiments.

BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The present invention will now be described by means of illustrative examples, without limiting the scope or spirit of the invention, with reference to the attached drawings.

[0054] Figure 1 illustrates the principle of harmonic transposition based on a subrange block.

[0055] Figure 2 illustrates the non-linear subrange block processing operation with a subrange input.

[0056] Figure 3 illustrates the non-linear subrange block processing operation with two subrange inputs.

[0057] Figure 4 illustrates vector product-enhanced subrange block-based harmonic transposition.

[0058] Figure 5 illustrates an exemplary scenario for applying subrange block-based transposition using various transposition orders in an enhanced HFR audio codec.

[0059] Figure 6 illustrates an exemplary scenario for the operation of a multi-order subrange block-based transposition applying a 64-band QMF analysis filter bank.

[0060] Figures 7 and 8 illustrate the experimental results of the described subrange block-based transposition method.

[0061] Figure 9 shows a detail of the non-linear processing unit of Figure 2, including a pre-normalizer and a multiplier.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0062] The embodiments described below are merely illustrative of the principles of the present invention, SUB-RANGE BLOCK-BASED HARMONIC TRANSPOSITION IMPROVED BY VECTOR PRODUCT. It is understood that modifications and variations of the arrangements and details described herein will be apparent to others skilled in the art. It is the intention, therefore, that the invention be limited only by the scope of the embodiments and not by the specific details presented by way of description and explanation of the embodiments in this document.

[0063] Figure 1 illustrates the principle of subrange block-based harmonic transposition, time stretching, or a combination of transposition and time stretching. The time domain input signal is fed into an analysis filter bank 101 that provides a plurality of complex valued subrange signals. These are fed to the subrange processing unit 102, the operation of which may be influenced by control data 104. Each output subrange may be obtained from the processing of one or two input subranges, or even as an overlay of the result of several such processed sub-tracks. The plurality of complex valued output subbands is fed to a synthesis filter bank 103, which in turn outputs the modified time domain signal. Optional control data 104 describes the configuration and processing parameters of the subband which can be adapted to the signal to be transposed. For the case of vector product-enhanced transposition, this data can carry information relative to a dominant fundamental frequency.

[0064] Figure 2 illustrates the non-linear subrange block processing operation with a subrange input. Given the physical time stretching and transposition target values, and the physical parameters of the analysis and synthesis filter banks 101 and 103, one deduces subrange time stretching and transposition parameters, as well as a source subrange index for each subband target index. The purpose of subrange block processing then is to perform corresponding transposition, time stretching, or a combination of transposition and time stretching of the source subrange signal of complex values in order to produce the target subrange signal.

[0065] A block extractor 201 samples a finite frame of samples from the complex-valued input signal. The frame is defined by an input pointer position and the subrange transpose factor. This frame undergoes non-linear processing in the processing section 202 and is then provided with windows of finite and possibly variable length in the windowing section 203. The resulting samples are added to previously produced samples in an overlay and add unit. 204 where the position of the output frame is defined by the position of an output pointer. The input pointer is incremented by a fixed value and the output pointer is incremented by the subrange stretch factor times the same value. One iteration of this chain of operations will produce an output signal with the duration being the subrange stretching factor times the duration of the subrange input signal, up to the length of the synthesis window, and with complex frequencies transposed by the subrange transposition factor. subrange. Control signals 104 can influence each of the three sections 201, 202, 203.

[0066] Figure 3 illustrates the non-linear subrange block processing operation with two subrange inputs. Given the physical time stretching and transposition target values, and the physical parameters of the analysis and synthesis filter banks 101 and 103, one deduces subrange time stretching and transposition parameters, as well as two source subrange indices for each subband target index. In the case where non-linear subrange block processing is to be used for creating missing partials by adding the cross product, the configuration of sections 301-1, 301-2, 302, 303, as well as the values of the two indices of the source subrange, may depend on the output 403 of a cross-processing control unit 404. The purpose of block processing of the subrange is to perform corresponding transposition, time stretching, or a combination of transposition and time stretching of two complex-valued subrange signals in order to produce the subrange target signal. A first block extractor 301-1 samples a finite time frame of samples from the first complex-valued source subrange, and the second block extractor 301-2 samples a finite frame of samples from the second complex-valued source subrange. The frames are defined by a common input pointer position and the subrange transpose factor. The two frames undergo non-linear processing at 302 and are subsequently windowed by windows of finite length in the windowing section 303. The overlay and addition unit 204 may have a similar or identical structure to that shown in Figure 2. A iteration of this chain of operations will produce an output signal with the duration being the subrange stretching factor times the two subrange input signals, (up to the length of the synthesis window).

[0067] In case the two subrange input signals carry the same frequencies, the output signal will have complex frequencies transposed by the subrange transposition factor. In the case where the two subband signals carry different frequencies, the present invention teaches that the windowing 303 can be adapted to generate an output signal that has a target frequency suitable for generating missing partials in the transposed signal.

[0068] Figure 4 illustrates the principle of subrange block-based transposition enhanced by vector product, time stretching, or a combination of transposition and time stretching. The direct subrange processing unit 401 may be of the type already described with reference to Figure 2 (section 202) or Figure 3. A cross subrange processing unit 402 is also fed with the multitude of complex valued subrange signals, and its operation is influenced by cross-processing control data 403. The sub-range cross-processing unit 402 performs non-linear processing of sub-range blocks of the type with two sub-range inputs described in Fig. 3, and the target output sub-ranges are added to those of direct subrange processing in adder 405.

[0069] Cross-processing control data 403 may vary for each position of the input pointer and consists of at least: a selected list of target subrange indices; • a pair of source subrange indices for each selected target subrange index; and a synthesis window of finite length.

[0070] A cross-processing control unit 404 provides this cross-processing control data 403 given a portion of the control data 104 describing a fundamental frequency and the plurality of complex-valued subrange signals output from the analysis filter bank 101 Control data 104 may also carry other dependent signal configuration parameters that influence vector product processing.

[0071] In the following text, a description of vector product-enhanced subrange block-based transposition and time stretching principles will be outlined with references to Figures 1-4, by the addition of appropriate mathematical terminology.

[0072] The two main configuration parameters of the general harmonic transposer and/or time stretcher are: the desired physical time stretching factor; It is the desired physical transposition factor.

[0073] The filter banks 101 and 103 can be of any complex exponential modulated type such as QMF or a windowed DFT or a small wave transform. The analysis filter bank 101 and the synthesis filter bank 103 can be stacked oddly and evenly in modulation and can be defined from a wide range of prototype filters and/or windows. details in subsequent design such as phase correction and subrange mapping management, the main system design parameters for subrange processing can generally be obtained from the two quotients of the following four filter bank parameters, all measured in physical units. In the above quotient is the sample time step of the subrange or time step of the analysis filter bank 101 (e.g., measured in seconds [s]); is the subband frequency spacing of the analysis filter bank 103 (e.g., measured in Hertz [1/s]); is the sample time step of the subrange or time step of the synthesis filter bank 103 (e.g., measured in seconds [s]); It is is the subband frequency spacing of the synthesis filter bank 103 (e.g., measured in Hertz [1/s]);

[0074] For the configuration of the sub-range processing unit 102, the following parameters must be calculated: S: the sub-range elongation factor, that is, the elongation factor that is applied within the sub-range processing unit 102 as a proportion of the input and output samples in order to achieve an overall physical time stretching of the time domain signal by S; Q: the subrange transposition factor, that is, the transposition factor that is applied within the subrange processing unit 102 in order to achieve an overall physical frequency transposition of the time domain signal by the Q factor; and the correspondence between source and target subrange indices, wherein n denotes an index of an analysis subrange entering the subrange processing unit 102, and m denotes an index of a corresponding synthesis subrange at the output of the processing unit 102. subrange 102.

[0075] In order to determine the stretching factor of the subband S, it is noted that an input signal to the analysis filter bank 101 of physical duration D corresponds to a number of the analysis subrange samples at the input of the subrange processing unit 302.

[0076] These samples will be extended to by the subband processing unit 102 which applies the subband stretching factor S.

[0077] At the output of the synthesis filter bank103 these samples result in an output signal having a physical duration of Since this last duration must meet the specified value, i.e., since the duration of the time domain output signal must be time elongated compared to the time domain input signal by the lengthening factor S physical time, the following design rule is obtained:

[0078] In order to determine the subrange transposition factor Q which is applied within the subrange processing unit 102 in order to achieve a physical transposition , it is observed that an input sinusoid for the filter bank physical frequency analysis 101 will result in a complex analysis subrange signal with discrete-time angular frequency and the main contribution occurs within the analysis subrange with the index An output sinusoid at the output of the synthesis filter bank 103 of the desired transposed physical frequency will result from feeding the synthesis subrange with the index with a complex signal of discrete angular frequency subrange . In this context, care must be taken to avoid the synthesis of distorted output frequencies other than Typically, this can be avoided by making appropriate second-order choices as discussed, for example, by selecting appropriate analysis and/or synthesis filter banks. The discrete frequency at the output of the subband processing unit 102 must correspond to the discrete time frequency at the input of the subrange processing unit 102 multiplied by the subrange transposition factor Q. That is, by definition equal and the following relationship between the physical transposition factor Q and the subrange transposition factor Q can be determined.

[0079] Likewise, the appropriate source or analysis subrange index n of the subrange processing unit 102 for a given target or synthesis subrange index m must comply with:

[0080] In one embodiment, it contains that is, the frequency spacing of the synthesis filter bank 103 corresponds to the frequency spacing of the analysis filter bank 101 multiplied by the physical transposition factor, and the one-to-one mapping of analysis to synthesis subband index n=m can be applied. In other embodiments, the subrange index mapping may depend on the details of the filter bank parameters. In particular, if the frequency spacing fraction of the synthesis filter bank 103 and the analysis filter bank 101 is different from the physical transposition factor one or two source subranges can be assigned to a given target subrange. In the case of two source subranges, it may be preferable to use two adjacent source subranges with index n, n+1, respectively.

[0081] That is, the first and second origin subranges are given by either (n(m), n(m) + 1) or (n(m) +1, n(m)).

[0082] The subrange processing of Figure 2 with a single source subrange will now be described as a function of the subrange processing parameters S and Q. Let x(k) be the input signal to extractor block 201, and let h be the input step of the block. That is, x(k) is a complex analysis subrange signal of an analysis subrange with index n.

[0083] The block extracted by block extractor 201 can, without loss of generality, be considered to be defined by L=R+R2 samples

[0084] Where the integer l is a block count index, L is the length of the block and are non-negative integers.

[0085] It is observed that for Q=1, the block is extracted from consecutive samples, however, for Q>1, a resolution reduction is performed in such a way that the input addresses are extended by the Q factor.

[0086] If Q is an integer this operation is normally simple to perform, whereas an interpolation method may be necessary for non-integer values of Q. This statement is also relevant for non-integer values of the increment h, i.e. of the input block pitch. In one embodiment, short interpolation filters, e.g., filters with two filter taps, can be applied to the complex-valued subband signal.

[0087] For example, if a sample at fractional time index k + 0.5 is required, a two-lead interpolation of the form where the coefficients a, b can be constant or can depend on a sub-range index (see, for example, WO2004/097794 and WO2007/085275), it can ensure sufficient quality.

[0088] An interesting special case of formula (4) is where the extracted block consists of a single sample, that is, the length of the block is L=1.

[0089] With the polar representation of a complex number on what is the magnitude of the complex number and is the complex number phase, the nonlinear processing unit 202 producing the output frame , from producing the input frame is advantageously defined by the phase modification factor until

[0090] Where is a geometric magnitude weighting parameter.

[0091] The case corresponds to a pure phase modification of the extracted block. A particularly attractive value of magnitude weighting is for which some computational complexity relief is obtained regardless of block length and the resulting transient response is somewhat improved with respect to the case The phase correction parameter depends on the filter bank details and the source and target subrange indices. In one embodiment, the phase correction parameter and can be determined experimentally by sweeping a set of input sinusoids. Additionally, the phase correction parameter can be obtained by studying the phase difference of two adjacent target subband complex sinusoids or by optimizing performance for a Dirac pulse type input signal. Finally, with proper design of the analysis and synthesis filter banks 101 and 103, the phase correction parameter S can be set to zero, or omitted. The phase modification factor T must be an integer such that the coefficients Tl and l are

[0092] Integers in the linear combination of phases in the first line of formula (5). With this hypothesis, that is, with the assumption that the phase modification factor T is an integer, the result of the nonlinear modification is well defined even if the phases are ambiguous by identification module

[0093] That is, formula (5) specifies that the phase of a sample of output frames is determined by offsetting the phase of a corresponding sample of input frame by a constant offset value. This constant displacement value may depend on the modification factor T, which in turn depends on the subrange elongation factor and/or the subrange transposition factor. Furthermore, the constant offset value may depend on the phase of a particular input frame sample of the input frame. This particular input frame sample is held fixed for determining the phase of all output frame samples of a given block. In the facility of formula (5), the central sample phase of the input frame is used as the phase of the particular input frame sample.

[0094] The second line of formula (5) specifies that the magnitude of a sample of the output frame may depend on the amplitude of the corresponding sample of the input frame. Furthermore, the magnitude of a sample of an output frame sample may depend on the magnitude of a particular input frame sample. This particular input frame sample can be used for determining the magnitude of all output frame samples. In the case of formula (5), the central sample of the input frame is used as the particular input frame sample. In one embodiment, the magnitude of a sample of the output frame may correspond to the geometric mean of the magnitude of the corresponding sample of the input frame and the particular input frame sample.

[0095] In the window opening unit 203, a window w of length L is applied to the output frame resulting in the windowed output frame:

[0096] Finally, it is assumed that all frames are spanned by zeros, and the overlap and addition operation 204 is defined by:

[0097]

[0098] In which it should be noted that the overlay and addition unit 204 applies a block step of Sh, that is, a time step, which is S times greater than the step h of the input block. Due to this difference in the time steps of formula (4) and (7) the duration of the output signal z (k) is S times the duration of the input signal x (k), i.e. the synthesis subband signal has been stretched by the subrange stretching factor S compared to the analysis subrange signal. It should be noted that this observation normally applies if the window length L is insignificant compared to the signal duration.

[0099] For the case where a complex sinusoid is used as input, for subrange processing 102, that is, an analysis subrange signal corresponding to a complex sinusoid:

[00100] It can be determined by applying formulas (4) - (7) that the output of subband processing 102, that is, the corresponding synthesis subband signal, is given by:

[00101] Regardless of . Thus, a complex sinusoid of discrete time frequency will be transformed into a complex sinusoid with a discrete time frequency as long as the synthesis window moves with a step of Sh sums up to the same constant value K for all k,

[00102] It is illustrative to consider the special case of pure transposition where S=1 and T=Q. If the step of the input block is h: -1 and R1 = 0, R2 = 1, all of the above, i.e. notably formula (5), reduces to the "point-wise" based phase modification rule or sample;

[00103] The subrange processing unit 102 may use control data 104 to set certain processing parameters, for example, the block length of the block extractors.

[00104] In what follows, the description of subrange processing will be extended to cover the case of Figure 3, with two subrange inputs. Let yourself ' be the subrange input signal for the first block extractor 301-1 and let be the subrange input signal for the second block extractor 301-2. Each extractor can use a different downsampling factor, leading to extracted blocks,

[00105] Non-linear processing 302 produces the output frame and can be defined by

[00106] The processing at 303 is again described by (6) and (7) and 204 is identical to the overlap and addition processing described in the context of the single-input case.

[00107] Defining real non-negative parameters and the integer non-negative parameters and the synthesis window w now depend on the desired operating mode. It is observed that if the same substrip is fed to both inputs nneractions (12) and 113) and , operations (i2) and (13) reduce to those of (4) and (5), in the case of simple input.

[00108] In one embodiment, in which the frequency spacing ratio of the synthesis filter bank 103 and the frequency spacing of analysis filter bank 101 is different from the desired physical transposition factor , it may be beneficial to determine samples from a synthesis subrange with index m from two analysis subranges with index n, n+1, respectively. For a given index m, the corresponding index n can be given by the integer value obtained by truncating the value of the analysis index n given by formula (3).

[00109] One of the analysis subrange signals, for example, the analysis subrange signal corresponding to index n, is fed into the first block extractor 303-1 and the other analysis subrange signal, e.g. corresponds to an index n +, 1 is fed into the second extractor block 301-2. Based on these two analysis subrange signals a synthesis subrange signal corresponding to index m is determined according to the processing outlined above. The assignment of the analysis subrange signals adjacent to the two block extractors 301 -1 and 302-1 can be based on the remainder that is obtained when truncating the index value of formula (3), i.e., the difference of the value of the exact index given by formula (3) and the truncated integer value obtained from formula (3). If the remainder is greater than 0.5, then the analysis subrange signal corresponding to index n can be assigned to the second block extractor 302-2, otherwise this analysis subrange signal can be assigned to the first block extractor blocks 301-1. In this mode of operation, the parameters can be designed such that subrange input signals sharing the same complex frequency w,

[00110] Leads to a subrange output signal being a complex sinusoid with a discrete time frequency

[00111] This happens if the following relationships hold:

[00112] For the mode of operation to generate missing partials via vector products, the design criteria are different. Returning to the physical transposition parameter The purpose of a cross product addition is to produce an output at the frequencies given inputs on frequencies where is a fundamental frequency belonging to a dominant component with a determined tone of the input signal.

[00113] As described in WO2010/081892, selective addition of such terms will result in a completion of the harmonic series and a significant reduction of the ghost tone artifact.

[00114] A constructive algorithm for operating cross-processing control 404 will now be outlined.

[00115] Given a target subrange index m, the parameter and the fundamental frequency one can deduce the appropriate source subrange indices n and n2 by solving the following system of equations in an approximate sense,

[00116] Where For oddly stacked filter bank modulation (as typically used for QMF and MDCT filter banks) and for uniformly stacked filter bank modulation (as typically used for FFT filter banks).

[00117] With the settings, the fundamental frequency measured in units of the frequency spacing of the analysis filter bank; the synthesis quotient for analysis of spacing subrange frequencies, and the real-valued target to a smaller integer-valued source index, an example of an advantageous approximate solution to (16) is given by selecting n, as the closest integer to as the closest integer to .

[00118] If the fundamental frequency is smaller than the analysis filter bank spacing, that is, if , it may be advantageous to cancel the addition of a cross product.

[00119] As taught in WO2010/081892, a vector product should not be added to an output subrange which already has a significant main contribution from transposition without vector products. Furthermore, at most one of the cases must contribute to the output of the vector product. Here, these rules can be realized by performing the following three steps for each target output subrange index m: Calculate the maximum Mc over all d options; of the minimum of the candidate magnitudes of subranges of Origin and evaluated at (or from a region close to the origin of) the central time step , in which the source subbands can be given by the indices n, and n2 as in equation (16); Calculate the corresponding magnitude for the direct source term |x| obtained from a source subrange with index 3, Activate the cross term of a winning choice for Mc in point 1 above.

[00120] Only if where q is a predetermined threshold value.

[00121] Variations of this procedure may be desirable depending on the configuration parameters of the particular system. One such variation is to replace the hard threshold in point 3 with softer rules depending on the quotient Another variation is to expand the maximization at point 1 to more than options, for example, defined by a finite list of candidate values for fundamental frequency measured in frequency spacing analysis units P. Yet another variation is to apply different measures of subrange magnitudes, such as a fixed sample magnitude, a maximum magnitude, an average magnitude, a magnitude in normal sense. , etc..

[00122] The list of source target tracks m selected by adding a vector product together with the values of constitutes a major part of the cross-processing control data 403. What remains to be described are the configuration parameters non-negative integer parameters which appear in the phase rotation (13) and the synthesis window to be used in sub-range cross-processing 402. Inserting the sinusoidal model for the vector product situation leads to the following signals from the source sub-range;

[00123] Where Likewise, the desired output subrange is of the form:

[00124] Calculations show that this production target can be achieved if (15) is satisfied together with:

[00125] Conditions (15) and (19) are equivalent to:

[00126] Which defines the integer factors for the phase modification in (13) and provides a certain design freedom in defining the values of the resolution reduction factors The magnitude weighting parameter can be advantageously chosen to . As can be seen, these configuration parameters depend only on the fundamental frequency through the selection of r. However, by (18) to contain, a new condition in the synthesis window w emerges, namely:

[00127] A synthesis window w which satisfies (21) exactly or approximately must be provided as the last piece of cross-processing control data 403.

[00128] It is noteworthy that the above algorithm for computing cross-processing of control data 403 based on input parameters, such as an output target subrange index m and a fundamental frequency is of a purely exemplary nature and, as such, does not limit the scope of the invention. Variations of this disclosure within the skill of the art and routine experimentation - for example, an additional block-based subband processing method that provides a signal (18) as output in response to input signals (17) - fall entirely within the scope of the invention. invention.

[00129] Figure 5 illustrates an exemplary scenario for applying subrange block-based transposition using various transposition orders in an enhanced HFR audio codec. A transmitted bit stream is received at a central decoder 501, which provides a low bandwidth central decoded signal at a sampling frequency The low bandwidth center decoded signal is resampled to the output sampling frequency through a complex 32-band QMF modulated analysis bank 502 followed by a 64-band QMF (inverse QMF) synthesis bank 505. The two filter banks 502 and 505 share the same physical parameters The HFR 504 processing unit simply passes through the smaller unmodified subbands corresponding to the low bandwidth central signal. The high-frequency content of the output signal is obtained by feeding the highest sub-bands of the 64-track QMF synthesis bank 505 to the output bands of a multiple transposition unit 503, subject to spectral shaping and modification performed by a transposition unit. HFR processing 504. The multiple transposer 503 takes as input the central decoded signal and outputs a plurality of subrange signals, which represent the QMF 64 range analysis of an overlay or a combination of various components of the transposed signal. The goal is that if HFR processing is bypassed, each component corresponds to an integer physical transposition without lengthening the timing of the center signal. . In the scenario of the invention, the transposer control signal 104 contains data describing a fundamental frequency. This data may be transmitted via the corresponding audio encoder bitstream, deduced by tone detection in the decoder, or obtained from a combination of transmitted and detected information.

[00130] Figure 6 illustrates an exemplary scenario for the operation of a multi-order subrange block-based transpose applying a single 64-band QMF analysis filter bank. Here three orders of transposition must be produced and delivered within the domain of a 64-lane QMF operating at an output sampling rate of The merge unit 603 simply selects and combines the relevant subbands from each transpose factor branch into a single plurality of QMF subbands to be fed into the HFR processing unit. The objective is specifically that the processing chain of a 64-band QMF analysis 601, a sub-range processing unit 602 , and a 64-track QMF synthesis 505 results in a physical transposition of <4 with (i.e. no stretch). Identifying these three blocks with 101, 102 and 1.03 in Figure 1, we discover that like this A project of specific configuration parameters for will be described separately for each case. For all cases, the analysis step is chosen to be h=1 and it is assumed that the normalized fundamental frequency parameter is known.

[00131] Consider the case first Then 602-2 has to perform a subrange stretching of S=2, a subrange transposition of (i.e. none) and the correspondence between source n and target subranges m is given by n=m for direct subrange processing. In the scenario of the vector product addition invention there is only one type of vector product to consider, namely r=1 (see above, after equation (15)), and equations (20) reduce to An example solution is to choose For direct processing synthesis window, a rectangular window of uniform length L=10 with R1=R2=5 can be used, since it satisfies condition (10). For cross-processing synthesis window, a window obtained L=2 can be used with R1=R2=1, in order to keep the additional complexity of adding vector products to a minimum. After all, the beneficial effect of using a long block for subband processing is most notable in the case of complex audio signals, where unwanted intermodulation terms are suppressed; in the case of a dominant tone, such artifacts are less likely to occur. The derivation window L=2 is the shortest that can satisfy (10) as long as h=1 and S=2. Through the present invention, however, the window advantageously satisfies (21). For the parameters at hand, this amounts to: which is fulfilled by choosing It is

[00132] Just in case The specifications for 602-3 given by (1) - (3) are that it has to perform a stretching of the S=2 subrange, a transposition of the Q=3/2 subrange, and that the correspondence between the source subranges ne target m for direct processing of the term is given by There are two types of cross product terms r=1, 2, and equations (20) reduce to

[00133] An exemplary solution consists of choosing the parameters decreasing resolution as D1= 0 and D2 = 3 / 2 for r = 1; D1 = 3/2 and D2 = 0 for r = 2.

[00134] For direct processing synthesis window, a rectangular window of uniform length L= 8 with R1=R2= 4 can be used. For cross-processing synthesis window, a short obtained window L= 2 can be used with R1=R2=1, and that satisfies:

[00135] Which is fulfilled by choosing w(o) =1 and

[00136] Just in case ' The specifications for 602-4 given by (1) - (3) are that it has to perform an elongation of the subrange of S=2, a transposition of the subrange of Q=2, and that the correspondence between the source and target subranges m for direct processing of the term is given by n=2m. There are two types of cross product terms r=1, 2, and equations (20) reduce to

[00137] An exemplary 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;

[00138] For direct processing synthesis window, a rectangular window of uniform length L= 6 with R1=R2 = 3 can be used. For cross-processing synthesis window, a short obtained window L=2 can be used with R1=R2 = 1 and that satisfies:

[00139] Which is fulfilled by choosing w(o) =1 and

[00140] In each of the above cases where more than one value of r is applicable, a selection will occur, for example, in a similar way to the three-step procedure described before equation (17).

[00141] Figure 7 describes the amplitude spectrum of a harmonic signal with fundamental frequency The low frequency part 701 of the signal is to be used as input to a multiple transposer. The purpose of the transposer is to generate a signal as close as possible to the high frequency part 702 of the input signal, so that transmission of the high frequency part 702 becomes non-imperative and the available bit rate can be utilized economically. .

[00142] Figure 8 depicts the output amplitude spectrum of a transposer that has the low frequency part 701 of the signal of Figure 7 as input. The multiple transposer is constructed using 64-band QMF filter banks, sampling frequency and according to the description in Figure 5. For clarity, however, only the two transposition orders are considered. The three different panels 801-803 represent the final output obtained using different settings of the cross-processing control data.

[00143] Top panel 801 depicts the output spectrum obtained if all vector product processing is canceled and only direct subrange processing 401 is active. This will be the case if the cross-processing control 404 receives no tone or Transposition by . generates output in the range of 4 to 8 kHz and transpose by generates output in the range of 8 to 12 kHz. As can be seen, the partials created are increasingly further apart and the output deviates significantly from the high frequency target signal 702. Audible artifacts of double and triple "ghost" tones will be present in the resulting audio output.

[00144] Middle panel 802 depicts the output spectrum obtained if vector product processing is active, the tone parameter is used (which is an approximation to , therefore, a two-lead synthesis window with w(0) = w(-1) = 1, satisfying condition (10), is used for cross-subrange processing. This amounts to a simple combination of subrange block-based processing and vector product-enhanced harmonic transposition. As can be seen, the additional output signal components compared to 801 do not align well with the desired harmonic series. This shows that it leads to insufficient audio quality when using the project's legacy procedure of direct subrange processing for vector product processing.

[00145] The bottom panel 803 describes the output spectrum obtained from the same scenario for the middle panel 802, but now with the cross-subband processing synthesis windows given by the formulas described in the cases of Figure 5. That is, a two-derivation window of the form w(0) =1 and satisfying (21) and with the feature taught by the present invention that depends on the value of •< As can be seen, the combined output signals align very well with the desired harmonic series 702.

[00146] Figure 9 shows a portion of the frame processing nonlinear processing unit 202 including sections configured to receive two input samples and to generate based on these a processed sample w, whose magnitude is given by a geometric mean of the magnitudes of the input samples, that is,

[00147] It is possible to obtain the processed sample w according to this specification by pre-normalizing each of the input samples in a respective prenormalizer 901, 902 and multiplying the prenormalized input samples into a weighted multiplier 910, which generates

[00148] It is clear that the operation of the prenormalizers 901, 902 and the weighted multiplier 910 is determined by the input parameters and It is easy to verify that equations (22) will be fulfilled if The expert will readily be able to generalize this scheme to an arbitrary number of input samples, where a multiplier is provided with input samples , some or all of which have experienced normalization. It is then observed that a common prenormalization (a=b, implying that prenormalizers 901, 902 produce identical results) is possible if the parameter is set to . This results in a computational advantage when many subranges are considered, since a common prenormalization step can be performed on all candidate subranges before multiplication. In an advantageous hardware implementation, a plurality of identically functioning prenormalizers are replaced by a single unit that switches between samples of different subranges in a time-division mode.

[00149] Additional embodiments of the present invention will become apparent to one skilled in the art after reading the above description. Although the present description and drawings disclose embodiments and examples, the present invention is not limited to these specific examples. Numerous modifications and variations may be made without departing from the spirit or scope of the present invention, which is defined by the embodiments.

[00150] The systems and methods described above can be implemented as software, firmware (printed program), hardware or a combination thereof. Certain components or all components may be implemented as software executed by a digital signal processor or microprocessor, or be implemented as hardware or as an application-specific integrated circuit. This type of software may be distributed on computer-readable media, which may comprise computational storage media (or non-transitory media) and communication media (or transient media). As is well known to a person skilled in the art, computational storage media includes both volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer readable instructions, data structures, modules programs or other data. Computational storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile discs (DVD,) or other optical disk storage, magnetic cassettes, magnetic tape, storage on magnetic disk or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be accessed by a computer. Furthermore, it is well known to the skilled artisan that communications media typically embody computer-readable instructions, data structures, program modules, or other data in a modulated data signal, such as a carrier wave or other transport mechanism, and include all means of delivering information.

Claims (14)

1. System configured to generate an extended time signal and/or a frequency transposed signal from an input signal, the system characterized by the fact that it comprises: a bank of analysis filters (101) configured to obtain a number Y > 1 of analysis subrange signals from the input signal, wherein each analysis subrange signal comprises a plurality of complex-valued analysis samples, each having a phase and a magnitude; a subrange processing unit (102) configured to generate a synthesis subrange signal from the Y analysis subrange signals using a subrange transposition factor Q and a subrange stretching factor S, at least one of Q and S being greater than one, wherein the subrange processing unit (102) comprises: a block extractor (201) configured to: i) form Y frames from L input samples, each frame being extracted from the plurality of complex valued analysis samples in an analysis subrange signal and the frame length being L > 1; and ii) applying a block jump size of h samples to the plurality of analysis samples, before forming a subsequent frame of L input samples, thereby generating a sequence of frames of input samples; a non-linear frame processing unit (202) configured to generate, based on Y corresponding frames of input samples formed by the extractor block (201), a frame of processed samples by determining a phase and a magnitude for each processed sample of the frame, wherein, for at least one processed sample: iii) the phase of the processed sample is based on the respective phases of the corresponding input sample in each of the Y input sample frames; and iv) the magnitude of the processed sample is based on the magnitude of the corresponding input sample in each of the Y input sample frames; a windowing unit configured to apply a window function of a length corresponding to the frame length L to the processed sample frame; and an overlay and add unit (204) configured to determine the synthesis subrange signal by overlaying and adding the samples from a sequence of processed sample frames; and a bank of synthesis filters (103) configured to generate the time-elongated signal and/or the transposed frequency signal from the synthesis subband signal, wherein the system is operable for at least Y=2, and in that the block extractor (201) is configured to obtain at least one frame of input samples by downsampling the complex-valued analysis samples into an analysis subrange signal. 2. System according to claim 1, characterized by the fact that the analysis filter bank (101) applies an analysis time step ΔtA to the input signal, the analysis filter bank has a frequency spacing analysis frequency ΔfA; the analysis filter bank has N number of analysis subranges, with N > 1, where n is an analysis subrange index with n = 0,..., N - 1; an analysis sub-range of the N analysis sub-ranges is associated with a frequency range of the input signal; the synthesis filter bank (103) applies a synthesis time step ΔtS to the synthesis subband signal; the synthesis filter bank has a synthesis frequency spacing ΔfS; the analysis filter bank has a number M of synthesis subranges, with M > 1, where m is a synthesis subrange index with m = 0,..., M - 1; and a synthesis sub-band of the M synthesis sub-bands is associated with a time-stretched frequency band and/or transposed frequency signal, wherein the sub-band processing unit (102) is set to Y=2 and further comprises a cross-processing control module (404) configured to generate cross-processing control data (403) by defining subrange indices n1, n2 associated with the analysis subrange signals in such a way that the subrange indices differ by an integer p approaching is the ratio of a fundamental frequency QO of the input signal and the analysis frequency spacing ΔfA. 3. System, according to claim 2, characterized by the fact that the subrange indices n1, n2 are associated with the signals of the analysis subrange and the synthesis subrange index m, which subrange indices are related because they are solutions integer approximate to: where QO is a fundamental frequency of the input signal, α = 0 or 1/2, QΦ is a transposition factor; er is an integer that satisfies 1 < r < QΦ- 1. 4. System according to claim 3, characterized by the fact that the cross processing control unit (404) is configured to generate processing control data such that the subrange indices n1, n2 are based on a value of r that maximizes the minimum of the subrange magnitudes of the two frames formed by extracting analysis samples from analysis subrange signals. 5. System according to any one of claims 1 to 4, characterized by the fact that the window function is one of: Gauss Window, Cosine Window, Raised Cosine Window, Hamming Window, Hann Window, Rectangular Window, Bartlett, and Blackman's Window. 6. System according to claim 5, characterized by the fact that the window function comprises a plurality of window samples, and wherein overlapping and added window samples of a plurality of window functions, when weighted by Complex and shifted weights with a jump size of Sh form a substantially constant sequence. 7. System according to any one of claims 1 to 6, characterized by the fact that it is operable at least for Y=1 and Y=2. 8. System, according to any one of claims 1 to 7, characterized by the fact that the non-linear processing unit (102) comprises: a pre-normalizer (901, 902) configured to resize the magnitudes of the samples corresponding input samples in at least one of the Y input sample frames and a complex multiplier (910) configured to determine the sample processed by computing a weighted complex product of factors equal to the corresponding input sample in at least two of the Y input sample frames, at least one of the factors (v,, m and M ^0) being obtained from a sample with a magnitude rescaled by the pre- normalizer (901, 902). 9. System, according to any one of claims 1 to 8, characterized by the fact that it is configured for Y = 2, comprising: a bank of analysis filters (101) configured to obtain a first and a second signal from the subband of analysis from the input signal; a subrange processing unit (102) configured to determine a synthesis subrange signal from the first and second analysis subrange signals, wherein the subrange processing unit (102) comprises: a first block extractor (301 -1) configured to: i) form a first frame of L input samples from the plurality of complex-valued analysis samples in the first analysis subrange signal, the length of the frame being L> 1; and ii) applying a block jump size of h samples to the plurality of analysis samples, before forming a subsequent frame of L input samples, thus generating a first sequence of input sample frames; a second block extractor (301-2) configured to: i) form a second frame of L input samples from the plurality of complex valued analysis samples in the second analysis subrange signal; and ii) applying the block jump size of h samples to the plurality of analysis samples, before forming a subsequent frame of L input samples, thus generating a second sequence of frames of input samples; a nonlinear frame processing unit (302) configured to generate, based on the first and second input sample frames, a processed sample frame; and an overlap and addition unit (204) configured to form the synthesis subband signal; and a bank of synthesis filters (103) configured to generate the elongated time and/or transposed frequency signal from the synthesis subband signal. 10. System according to any one of claims 1 to 9, characterized by the fact that it further comprises: a plurality of sub-range processing units (401, 402; 503; 602-2, 602-3, 602- 4), each configured to determine an intermediate synthesis subrange signal using a different value of the subrange transposition factor Q and/or the subrange stretching factor S; and a blending unit (405, 603) disposed downstream of the plurality of subrange processing units (401, 402; 503; 602-2, 602-3, 602-4) and upstream of the synthesis filter bank ( 103) configured to merge corresponding synthesis subrange intermediate signals to determine the synthesis subrange signal. 11. System according to claim 10, characterized by the fact that it is configured for Y = 2, wherein at least one of the subrange processing units is a direct subrange processing unit (401), which is configured to determining a synthesis subrange signal from an analysis subrange signal using a subrange transposition factor Q and a subrange stretching factor S, and at least one is a subrange cross-processing unit (402) , which is configured to determine a synthesis subrange signal from two analysis subrange signals using a subrange transposition factor Q and a subrange stretching factor S, which are independent of the first two factors, the system being configured to disable at least one subrange cross-processing unit (402) if, for a given synthesis subrange, one of the following conditions is satisfied: a) the ratio of the magnitude Ms of the direct source term analysis subrange producing the subrange of synthesis and the smallest magnitude Mc in an optimal pair of cross-source terms producing the synthesis subrange is greater than a predetermined constant q: b) the synthesis subrange has a significant contribution from a direct processing unit; c) a fundamental frequency QO is smaller than the analysis filter bank spacing ΔfA. 12. System according to any one of claims 1 to 11, characterized by the fact that: the analysis filter bank (101) is configured to form Y x Z analysis subrange signals from the input signal; The subrange processing unit (102) is configured to generate Z synthesis subrange signals from the Y x Z analysis subrange signals by applying a pair of S and Q values to each group of Y analysis subrange signals in the which a synthesis subband signal is based on; and the synthesis filter bank (103) is configured to generate Z time-stretched and/or frequency transposed signals from the Z synthesis subband signals. 13. Method for generating an elongated time signal and/or a transposed frequency signal from an input signal, the method comprising the steps of: obtaining a number Y > 2 of analysis subrange signals from the input signal, wherein each analysis subrange signal comprises a plurality of complex-valued analysis samples, each having a phase and a magnitude; forming Y frames from L input samples, each frame being extracted from the plurality of complex-valued analysis samples into an analysis subrange signal and the frame length being L > 1; applying a block jump size of h samples to the plurality of analysis samples, before obtaining a subsequent frame of L input samples, thereby generating a sequence of frames of input samples; generate, based on Y corresponding frames of input samples, a frame of processed samples by determining a phase and a magnitude for each processed sample in the frame, wherein, for at least one processed sample: i) the phase of the processed sample is based on the respective phases of the corresponding input sample in each of the Y input sample frames, and ii) the magnitude of the processed sample is based on the magnitude of the corresponding input sample in each of the Y frames of input samples; determine the synthesis subrange signal by applying a window function of a length corresponding to the frame length L to the processed sample frame and superimposing and adding the samples from a sequence of processed sample frames, and generating the time-stretched signal and/or the frequency signal transposed from the synthesis subrange signal, wherein forming the input sample frames includes downsampling complex valued analysis samples into an analysis subrange signal. 14. Non-transitory computer-readable storage medium having the method as defined in claim 13.