AR080477A1

Apparatus and method for downsampling an audio signal

Abstract

An apparatus for processing an input audio signal (2300) is based on a cascade of filter banks, a cascade comprising a bank of synthesis filters (2304) for synthesizing an intermediate audio signal (2306) from the signal of input audio (2300), where the input audio signal is represented by a plurality of first subband signals (2303) generated by a bank of analysis filters (2302), where the number of filter bank channels of the synthesis filter bank (2304) is less than a number of channels of the analysis filter bank (2302). The apparatus further comprises an additional analysis filter bank (2307) for generating a plurality of second subband signals (2308) of the intermediate audio signal (2306), where the additional filter bank has a number of channels that differs from the number of channels of the synthesis filter bank (2304), whereby the sampling frequency of a subband signal of the plurality of second subband signals (2308) is different from the sampling frequency of a first subband signal of the plurality of first subband signals (2303).

AR080477A1, drawing sheet 1 — Sheet 1 of 98

Term

No projected expiry on record.

1 claim: 1 independent, 0 dependent

1
Un aparato para procesar una senal de audio de entrada (2300) se basa en una cascada de bancos de filtros, cascada que comprende un banco de filtros de sintesis (2304) para sintetizar una senal de audio intermedia (2306) a partir de la senal de audio de entrada (2300), donde la senal de audio de entrada està representada por una pluralidad de senales de primera subbanda (2303) generadas por un banco de filtros de anâlisis (2302), donde el nùmero de canales de bancos de filtros del banco de filtros de sintesis (2304) es menor que un nùmero de canales del banco de filtros de anâlisis (2302). El aparato comprende ademâs un banco de filtros de anâlisis adicional (2307) para generar una pluralidad de senales de segunda subbanda (2308) de la senal de audio intermedia (2306), donde el banco de filtros adicional tiene un nùmero de canales que difiere del nùmero de canales del banco de filtros sintesis (2304), por lo que la frecuencia de muestreo de una senal de subbanda de la pluralidad de senales de segunda subbanda (2308) es diferente de la frecuencia de muestreo de una primera senal de subbanda de la pluralidad de senales de primera subbanda (2303).

346 paragraphs in 7 sections, as filed

“APPLIANCE AND METHOD FOR PROCESSING AN INPUT AUDIO SIGNAL USING FILTER BANKS IN CASE EACH”

Requested by:

223194

APP / RF

FRA UNHOFER-GESELLSCHAFT ZUR

FÖRDERUNG DER ANGEWANDTEN

FORSCHUNG EV * DOLBY INTERNAL TIONAL AB, residents in Hansastrasse 27c, 80686 Munich, GERMANY *

Atlas Complex, Africa Building Hoosoorddreef

9 „1101 BA Amsterdam, NETHERLANDS.

APPLIANCE AND METHOD FOR PROCESSING AN INPUT AUDIO SIGNAL USING FILTER BANKS IN CASCADA

TECHNICAL FIELD

The present invention relates to audio source coding systems that make use of a harmonic transposition method for high frequency reconstruction (HFR), and with digital effects processors, e.g. the so-called exciters, where the generation of harmonic distortion adds sharpness to the processed serial and with the time stretchers, where the duration of a serial is prolonged while maintaining the spectral content of the originai.

BACKGROUND OF THE INVENTION

In PCT WO 98/57436 the concept of transposition was established as a method to recreate a high frequency band from a lower frequency band of an audio serial. Substantial savings in bit rate can be obtained using this concept in audio coding. In an HFR-based audio coding system, a low bandwidth serial is processed by a core waveform encoder and the higher frequencies are regenerated using additional transposition and additional complementary information of very low bit rate which describes the objective spectral shape of the decoder side. In the case of low bit rates, where the bandwidth of the coded core serial is narrow, it becomes increasingly important to recreate a high band with pleasant perceptual characteristics. The harmonic transposition defined in PCT WO 98/57436 gives very good results in the case of complex musical material in a situation of

223,194. B.

Low crossover frequency The principle of a harmonic transposition is that a sinusoid with frequency ω is mapped against a sinusoid with frequency 7 \ and where T> \ is an integer that defines the order of transposition. In contrast, an HFR method based on single sideband modulation (SSB) maps a sinusoid with frequency ω against a sinusoid with frequency ω + Δω where Δω is a fixed frequency offset. Given a serial core with low bandwidth, a dissonant chime sound anomaly can occur as a result of SSB transposition.

To obtain the best possible audio quality, high-quality harmonic HFR methods of the current state of the art employ complex modulated filter banks, e.g. a Short Term Fourier Transform (STFT), with high frequency resolution and a high degree of oversampling to achieve the required audio quality. Fine resolution is necessary to avoid the harmful intermodulation distortion that arises from the nonlinear processing of sinus sums. With a sufficiently high frequency resolution, that is narrow subbands, high quality methods aim to have a maximum of one sinusoid in each subband. A high degree of oversampling over time is necessary to avoid distortion of the alias type and a certain degree of oversampling in the frequency is necessary to avoid previous echoes corresponding to transient serial. The obvious disadvantage is that computing complexity can be increased.

Harmonic transposition based on subband blocks is another HFR method used to suppress intermodulation products, in which case a filter bank with a thicker frequency resolution and a lower degree of oversampling is used, e.g. a multi-channel QMF bank. In this method, a time block of complex subband samples is processed by a common phase modifier, while the

Overlapping several modified samples forms an output subband sample. This has the effect of suppressing intermodulation products that would otherwise appear when the input subband serial consists of several sinusoids. Transposition based on block-based subband processing has a much lower computing complexity than high quality transponders and obtains almost the same quality with respect to many signals. However, the complexity is even much higher than in the case of trivial SSB-based HFR methods, since a plurality of analysis filter banks are needed, each of which processes signals of different transposition orders T , in a typical HFR application to synthesize the proposed side width. In addition, a common strategy is to adapt the sampling rate of the input signals to conform to the analysis filter banks of a constant size, even though the filter banks process signals of different transposition orders. It is also customary to apply bandpass filters to the input signals in order to obtain output signals, processed from different transposition orders, with spectral densities of non-overlapping power.

The storage or transmission of audio signals is often subject to strict bit rate restrictions. In the past, it was imperative that encoders drastically reduce the bandwidth of the transmitted audio when only a very low bit rate was available. Modern audio codes today can encode broadband signals using bandwidth extension (BWE) methods [1-12j. These algorithms are based on a perimetric representation of the aft frequency (HF) content that is generated from the low frequency (LF) part of the decoded signal by means of the

transposition to the spectral region of HF ("patching") and the application of a postprocessing driven by the parameters. The LF part is encoded with any audio or voice encoder. For example, the bandwidth extension methods described in [1-4] are based on the single sideband modulation method (SSB), often referred to as the “copying” method for generating multiple HF patches.

Recently, a new algorithm has been introduced that employs a bank of phase vocoders [15-17] for the generation of different patches [13] (see Fig. 20). This method has been developed to avoid the auditory harshness that is frequently observed in signals subjected to an SSB bandwidth extension. However, since the BWE algorithm is executed on the decoder side of a chain of codes, computer complexity is a serious problem. The methods of the current state of the technique, especially HBE based on phase vocoders, are performed at the expense of a greatly increased computer complexity compared to SSB-based methods.

As outlined above, existing bandwidth extension schemes apply a patching method on a given block of signals at a time, either SSB-based patching [1-4] or HBE-based vocoder patching [15]. -17]. In addition, modern audio encoders [19-20] offer the possibility of switching the patching method globally based on blocks of time between alternative patching schemes.

The SSB copy patching introduces harmful roughness in the audio signal, although it is simple from the computer point of view and retains the temporary envelope of the transients. Moreover, computer complexity is

significantly increases with respect to the very simple SSB copy method.

SUMMARY OF THE INVENTION

With regard to a reduction in complexity, sampling rates are of particular importance. This is due to the fact that a high sampling frequency translates into a high complexity and a low sampling frequency generally represents a low complexity due to the reduced number of operations required. On the other hand, however, the situation in bandwidth extension applications is particularly such that the sampling rate of the core serial output serial is generally so low that this sampling frequency is too low for a serial total bandwidth. In other words, when the sampling frequency of the decoder output serial is, for example, 2 or 2.5 times the maximum encoder frequency of the core encoder output serial, then a bandwidth extension from, for example a factor of 2 means that an operation to increase the number of samples is necessary so that the sampling frequency of the extended bandwidth serial is so high that the sampling can "cover" the additionally generated high frequency components.

In addition, filter banks such as analysis filter banks and synthesis filter banks are in charge of a considerable amount of processing operations. Therefore, the filter banks' chart, to say if a filter bank is a 32-channel filter bank, a 64-channel filter bank or even a filter bank with a high number of channels, includes significantly in the complexity of the audio processing algorithm. In general, a large number of filter bank channels can be said to require more processing operations and, therefore, greater

complexity than a small number of filter bank channels. With this in mind, bandwidth extension applications, as well as in other audio processing applications where different sampling rates need to be addressed, such as in vocoder type applications or any other audio effects application , there is a specific interdependence between complexity and sampling frequency or audio bandwidth, which means that the operations of increasing the number of samples or filtering in subbands can dramatically increase the complexity without specifically influencing the audio quality in a good way when the wrong tools or algorithms are chosen for the specific operations.

An objective of the present invention is to propose an improved audio processing concept, which on the one hand allows processing with low complexity and, on the other hand, good audio quality.

This objective is achieved by an apparatus for processing an input audio signal according to claim 1 or 18, a method for processing an input audio signal according to claim 20 or 21, or a computer program according to with claim 22.

The embodiments of the present invention are based on a specific cascade placement of analysis and / or synthesis filter banks to obtain low complexity resampling without sacrificing audio quality. In one embodiment, an apparatus for processing an input audio serial comprises a bank of synthetic filters to synthesize an intermediate audio serial from the input audio signal, where the input audio signal is represented by a plurality of first subband signals generated by an analysis filter bank placed in the processing direction before the synthesis filter bank, where the number of filter bank channels

of the synthesis filter bank is less than the number of channels of the analysis filter bank. The intermediate signal is further processed by an additional analysis filter bank to generate a plurality of second subband signals from the intermediate audio signal, where the additional filter bank has a number of channels that differs from the number of channels of the synthesis filter bank whereby the sampling frequency of a subband signal of the plurality of subband signals is different from the sampling frequency of a first subband signal of the plurality of first subband signals generated by the analysis filter bank.

The cascade of a bank of synthesis filters and an additional filter bank connected subsequently produces a conversion of the sampling frequency and, in addition, a modulation of the bandwidth portion of the original audio input signal that has been input in the synthesis filter bank in a baseband. This time-intermediate serial, which has now been extracted from the original input audio serial which may be, for example, the output serial of a core decoder of a bandwidth extension scheme, is now preferably represented in the form of a modulated critical sampling signal according to the baseband and it has been discovered that this representation, that is to say the resampled output signal, when processed by an additional analysis filter bank to obtain a representation of subbands of the other processing operations that may or may not take place and, for example, may be processing operations related to the extension of the bandwidth taies as operations of non-line subband followed by high frequency reconstruction processing and a combination of the subbands in the final synthesis filter bank.

The present application discloses different aspects of the devices, methods or computer programs for processing audio signals in the context of bandwidth extension and in the context of other audio applications that are not related to the extension of bandwidth The characteristics of the individual aspects described and claimed below can be combined partially or totally, although they can also be used separately, since the individual aspects already give rise to advantages with respect to perceptual quality, computer complexity and resources. processor / memory when implemented in a computer or microprocessor system.

The embodiments present a method to reduce the computer complexity of an HFR method based on subband blocks by means of efficient filtering and conversion of the sampling frequency of the input signals to the analysis stages of the HFR filter banks. . Moreover, it can be shown that the bandpass filters applied to the input signals are obsolete in a transponder based on subband blocks.

The present embodiments contribute to reducing the computational complexity of the harmonic transposition based on subband blocks by efficiently implementing several transposition orders based on subband blocks within the framework of a single pair of banks of analysis and synthesis filters. Depending on the compromise between perceptual quality and computer complexity, only one subset of orders can be executed ο all transposition orders together within a couple of filter banks. Moreover, there is a combined transposition scheme where only certain transposition orders are calculated directly, while the rest of the bandwidth is filled with replication of available transposition orders, that is to say previously calculated (eg. 2nd order) and / or the width of

band coded by the core. In this case, patching can be carried out using every imaginable combination of ranges of origin for replication.

In addition, the embodiments offer a method to improve both the high-quality harmonic HFR methods and the harmonic HFR methods based on subband blocks by means of the spectral alignment of the HFR tools. In particular, improved efficiency is obtained by aligning the spectral edges of the signals generated by HFR with the spectral edges of the envelope setting of the frequency table. Likewise, the spectral edges of the limiting tool are aligned by the same principle with the spectral edges of the signals generated by HFR.

Other embodiments are configured to improve the perceptual quality of the transients and at the same time reduce the computer complexity, for example, by applying a patching scheme that applies mixed patching consisting of harmonic patching and copy patching.

In specific embodiments, the individual filter banks of the cascading filter bank structure are quadrature mirror filter banks (QMF), all of which are based on a prototype low pass filter or window modulation using a series of modulation frequencies that define the central frequencies of the filter bank channels. Preferably, all the prototype window or filter functions have mutual dependence in such a way that the filters of the filter banks with different sizes (channels of the filter banks) also depend on each other. Preferably, the largest filter bank of a cascade filter bank structure comprising, in some embodiments, a first analysis filter bank, an immediately connected filter bank, another analysis filter bank and, in some cases post processing status, a bank of final synthesis filters, has an answer to the

Window function or prototype filters with a certain number of window function coefficients or prototype filters. The smaller banks of filters are, in all cases, a subsampled version of this window function, which means that the window functions of the other filter banks are subsampled versions of the “large” window function. For example, if a filter bank has half the size of a large filter bank, then the window function has half the number of coefficients and the coefficients of smaller size filter banks are derived by subsampling. In this situation, subsampling means that, for example. a filter coefficient is taken in the middle for the smallest filter bank that is half the size. However, when there are other relationships between the sizes of the filter banks that do not have integer values, then a certain type of interpolation of the window coefficients is executed, so that, ultimately, the filter bank window of smaller, once again, a subsampled version of the larger filter bank window.

The embodiments of the present invention are particularly advantageous in situations where only a portion of the input audio serial is needed for further processing, and this situation occurs especially in the context of the harmonic bandwidth extension. In this context, processing operations of the vocoder type are especially preferred.

An advantage of these embodiments is that they offer a lower complexity of a QMF transponder through efficient time and frequency domain operations and improved audio quality for harmonic replication of spectral bands based on QMF and DFT using the spectral alignment.

The embodiments relate to audio source coding systems that employ, for example, a harmonic transposition method based on subband blocks for high frequency reconstruction (HFR), and with digital effects processors, e.g. the so-called exciters, in which the generation of harmonic distortion brings sharpness to the processed signal, and also with time stretchers, in which the duration of the signal is extended while maintaining the spectral content of the original. The embodiments offer a method to reduce the computer complexity of a harmonic HFR method based on subband blocks by means of efficient filtering and conversion of the sampling frequency of the input serials prior to the filter bank analysis stages. of HFR. Moreover, the embodiments demonstrate that conventional bandpass filters applied to the input signals are obsolete in an HFR system based on subband blocks. In addition, the embodiments offer a method to improve both the high quality harmonic HFR methods and the harmonic HFR methods based on subband blocks by means of the spectral alignment of the HFR tools. In particular, the embodiments demonstrate how improved efficiency is obtained by aligning the spectral edges of the signals generated by HFR with the spectral edges of the envelope adjustment frequency table. Moreover, the spectral edges of the limiting tool are aligned, by the same principle, with the spectral edges of the signals generated by HFR.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention is now described by way of illustrative examples, which do not limit the scope or spirit of the invention, with reference to the accompanying drawings, in which:

Fig. 1 illustrates the operation of a block-based transponder that uses the transposition orders of 2, 3 and 4 in an HFR-powered decoder frame;

Fig. 2 illustrates the operation of the non-linear subband stretching units of Fig. 1;

Fig. 3 illustrates an efficient implementation of the block-based transponder of Fig. 1, where the band samplers and filters that precede the HFR analysis filter banks are implemented using time samplers in the multi-speed time domain and band pass filters based on QMF;

Fig. 4 illustrates an example of blocks for the construction of an efficient implementation of a multi-speed resampler in the time domain of Fig. 3;

Figs. 5a-5f illustrate the effect of an example of signal processed by different blocks of Fig. 4 for a transposition order of 2;

Fig. 6 illustrates an efficient implementation of the block-based transponder of Fig. 1, where the band samplers and filters that precede the HFR analysis filter banks are replaced by small banks of sub-sampled synthesis filters operating in subbands selected from a 32-band analysis filter bank;

Fig. 7 illustrates the effect of an example of signal processed by a subsampled filter bank of Fig. 6 corresponding to a transposition order of 2;

Figs. 8a-8f illustrate the implementation blocks of an efficient multi-speed reducer of the number of samples in the time domain of a factor of 2;

Figs. 9a-9-e: illustrate the implementation blocks of an efficient multi-speed reducer of the number of samples in the time domain of a 3/2 factor;

Fig. 10 illustrates the alignment of the spectral edges of the HFR transponder signals with the edges of the envelope adjustment frequency bands in an HFR-enhanced encoder;

Fig. 11 illustrates a situation in which anomalies arise due to the non-aligned spectral edges of the HFR transponder signals;

Fig. 12 illustrates a situation in which the anomalies of Fig. 11 are avoided as a result of the aligned spectral edges of the HFR transponder signals;

Fig. 13 illustrates the adaptation of the spectral edges in the limiting tool to the spectral edges of the serials of the HFR transponder;

Fig. 14 illustrates the principle of harmonic transposition based on subband blocks;

Fig. 15 illustrates an example of a situation for the application of transposition based on subband blocks using several transposition orders in an audio code enhanced by HFR;

Fig. 16 illustrates a type situation of the prior art with respect to the operation of a transposition based on multi-order subband blocks by applying a separate filter filter bank for each transposition order;

Fig. 17 illustrates a typical situation according to the invention with respect to the efficient operation of a transposition based on multi-order subband blocks applied by a single bank of 64-band QMF analysis filters;

io Fig. 18 illustrates another example of forming a subband signal processing;

Fig. 19 illustrates a single sideband modulation (SSB) patching;

Fig. 20 illustrates a harmonic bandwidth extension (HBE) patching;

Fig. 21 illustrates a mixed patching, where the first patching is generated due to the combination of frequencies and the second patch is generated by an SSB copy of a low frequency portion;

Fig. 22 illustrates an alternative mixed patching that uses the first HBE patch for an SSB copy operation to generate a second patch;

Fig. 23 illustrates a preferred cascade structure of analysis and synthesis filter banks;

Fig. 24a illustrates a preferred implementation of the small filter bank of Fig. 23;

Fig. 24b illustrates a preferred implementation of the additional filter bank of Fig. 23;

Fig. 25a illustrates general reviews of certain analysis and synthesis filter banks of ISO / IEC 14496-3: 2005 (E), and especially an implementation of an analysis filter bank that can be used for the filter bank of analysis of Fig. 23 and an implementation of a synthesis filter bank that can be used for the final synthesis filter bank of Fig. 23;

Fig. 25b illustrates an implementation in the form of a flow chart of the analysis filter bank of Fig. 25a;

Fig. 25c illustrates a preferred implementation of the synthesis filter bank of Fig. 25a;

Fig. 26 illustrates a general review of the framework in the context of bandwidth extension processing and

Fig. 27 illustrates a preferred implementation of signal processing of subbands produced as output from the additional filter bank of Fig. 23.

DESCRIPTION OF PREFERRED EMBODIMENTS

The embodiments described below are merely illustrative and may offer a reduction in the complexity of a QMF transponder through efficient time and frequency domain operations and improved audio quality of both the harmonic SBR based on QMF as in DFT through spectral alignment. It is understood that other people with technical training will consider the modifications and

variations of the provisions and details described here. Therefore, they are intended to be limited only by the scope of the patent claims set forth below and not by the specific details presented by way of description and explanation of the present embodiments.

Fig. 23 illustrates a preferred implementation of the apparatus for processing an input audio signal, where the input audio signal can be an input signal in the online time domain 2300 produced as output, for example, from a decoder of audio of the 2301 core. The input audio serial is input to a first bank of analysis filters 2302 which consists, for example, of a bank of analysis filters with M channels. In particular, the analysis filter bank 2302 produces as output M subband signals 2303, which may have a sampling frequency fs-fs / M. This means that the analysis filter bank is an analysis filter bank with critical sampling. This means that the bank of analysis filters

2302 it provides, for each block of M input samples on line 2300, a single sample for each subband channel. Preferably, the analysis filter bank 2302 is a complex modulated filter bank, which means that each subband sample has a magnitude and a phase or, equivalently, a real part and an imaginary part. Therefore, the input audio serial in line 2300 is represented by a plurality of first subband signals 2303 that are generated by the analysis filter bank 2302.

A subset of all the signals of the first subband is entered in a synthesis filter bank 2304. The synthesis filter bank 2304 has

Ms channels, where Ms is less than M. Consequently, not all subband signals generated by filter bank 2302 are entered as input to synthesis filter bank 2304, but only a subset, that is a

a certain smaller number of channels, as indicated in 2305. In the embodiment of Fig. 23, the sub-series 2305 covers a certain intermediate bandwidth although, on the other hand, the sub-series can also cover a bandwidth that begins with the channel 1 of the filter bank of the filter bank 2302 to a cove that has a number of channel lower than Μ, or on the other hand, substring 2305 can also cover a group of subband signals aligned with the highest channel M and extend to a lower channel having a channel number greater than channel number 1. On the other hand, you can start indexing the channels with zero depending on the notation actually used. Preferably, however, in the case of bandwidth extension operations a certain intermediate bandwidth is represented represented by the group of subband signals indicated at 2305 in the synthesis filter bank 2304.

The other channels that do not belong to group 2305 are not entered as input to the synthesis filter bank 2304. The synthesis filter bank 2304 generates an intermediate audio signal 2306, which has a sampling frequency equal to fs Ms/ M. Since Ms is less than M, the sampling frequency of intermediate serial 2306 must be less than the sampling frequency of the input audio signal on line 2300. Therefore, intermediate signal 2306 represents a signal with reduced and demodulated number of samples represented by subbands 2305, where the serial is demodulated to the baseband, since the lowest channel of group 2305 is entered into channel 1 of the bank of synthesis filters Ms and the highest channel of block 2305 is entered at the highest entrance of block 2304, apart from certain zero-fill operations corresponding to the lowest or highest channel in order to avoid problems of overlapping ("aliasing) at the edges of sub-series 2305. The apparatus for processing an input audio signal

It also includes an additional analysis filter bank 2307 to analyze intermediate serial 2306 and the additional filter bank has MTO channels, where Mto is different from Ms and preferably it is greater than Ms. When MTO is greater than Ms, then the sampling frequency of the subband serials produced as output from the additional filter bank 2307 and indicated at 2308 is less than the sampling frequency of a subband serial 2303. However, when MTO it is less than Ms, then the sampling frequency of a subband serial 2308 must be greater than a sampling frequency of a subband serial of the plurality of first subband serials 2303.

Therefore, the cascade of filter banks 2304 and 2307 (and preferably 2302) produces operations of increasing or reducing the number of very efficient and high quality samples or, in general terms, a very efficient resampling processing tool. The plurality of second subband serials 2308 is further processed, preferably, in a processor 2309 that performs the processing with the data resampled by the cascade of filter banks 2304, 2307 (and preferably 2302). In addition, it is preferable that block 2309 also executes an operation to increase the number of copies corresponding to bandwidth extension processing operations, so that, ultimately, the subbands provided as output by block 2309 have the same sampling frequency as the subbands provided as output of block 2302. Next, in a bandwidth extension processing application, these subbands are entered together with the additional subbands indicated in 2310, which are preferably the low band subbands generated, for example, by the analysis filter bank 2302, in a bank of synthesis filters 2311, which finally produces a serial processed in the time domain, for example

an extended bandwidth serial that has a sampling frequency of 2fs. This sampling frequency provided as output of block 2311 is, in this embodiment, twice the sampling frequency of the serial of line 2300, and this sampling frequency provided as output of block 2311 is high enough for the width of Additional band generated by the processing in block 2309 may be represented in the serial processed in the time domain with high audio quality.

Depending on the determined application of the present invention of the cascade filter banks, the filter bank 2302 may be in a separate device and an apparatus for processing an input audio serial may comprise only the synthesis filter bank 2304 and the additional filter bank 2307. In other words, the analysis filter bank 2302 can be distributed separately from a "post'-processor comprising blocks 2304, 2307 and, depending on the implementation, also blocks 2309 and 2311.

In other embodiments, the application of the present invention that implements cascading filter banks may be different due to the fact that a given device comprises the analysis filter bank 2302 and the smaller synthesis filter bank 2304, and the serial intermediate is sent to a different processor distributed by a different distributor or through a different distribution channel. Then, the combination of the analysis filter bank 2302 and the smaller synthesis filter bank 2304 represents a very efficient way to reduce the number of samples and, at the same time, demodulate the serial bandwidth represented by substring 2305 to the baseband. This reduction in the number of samples and demodulation to the baseband has been carried out without any loss of audio quality, and

especially without any loss of audio information and, therefore, is a high quality processing.

The table in Fig. 23 illustrates certain exemplary numbers corresponding to the different devices. Preferably, the analysis filter bank 2302 has 32 channels, the synthesis filter bank has 12 channels, the additional filter bank has twice the channels of the synthesis filter bank, such as 24 channels, and the bank of final synthesis filters 2311 has 64 channels. In general terms, the number of channels in the analysis filter bank 2302 is high, the number of channels io present in the synthesis filter bank 2304 is low, the number of channels in the additional filter bank 2307 is medium and the number of channels of the synthesis filter bank 2311 is very high. The sampling frequencies of the subband serials from the output of the analysis filter bank 2302 are fs/ M. The intermediate serial has a sampling frequency fs · Ms / M. The subband channels of the additional filter bank indicated in

2308 have a sampling frequency of fs · Ms / (MMTO), and the synthesis filter bank 2311 produces an output serial with a sampling frequency of 2fs, when the processing executed in block 2309 doubles the sampling frequency. However, when processing in the block

2309 it does not double the sampling frequency, then the sampling frequency produced as output from the synthesis filter bank must be correspondingly lower. Other preferred embodiments related to the present invention are described below.

Fig. 14 illustrates the principle of transposition based on subband blocks. The serial in the entry time domain is fed to a bank of analysis filters 1401 that produces a multitude of subbands of complex value. These are fed to the unit of

Subband processing 1402. The multitude of subband serials of complex value is fed to the synthesis filter bank 1403 which, in turn, outputs the serial in the modified time domain. Subband processing unit 1402 executes nonlinear processing operations of block-based subbands, such that the serial in the modified time domain is a transposed version of the input serial compositions in a transposition order 7> 1 . The notion of block-based subband processing is characterized by comprising nonlinear operations in blocks of more than one sample of the subband at a time, where subsequent blocks are placed in windows and aggregated in overlapping form to generate serial subband bands. exit.

Filter banks 1401 and 1403 can be of any complex exponential modulated type such as QMF horn or a window DFT. They can be stacked evenly or oddly in the modulation and can be defined by a wide range of filters or prototypical windows. It is important to know the quotient NfsIEfTO of the following two parameters of the filter banks, measured in physical units.

• Δ /λ : the frequency spacing of subbands of the analysis filter bank 1401;

• Nfs : the frequency spacing of subbands of the synthesis filter bank 1403.

Regarding the configuration of the sub-band processing 1402, it is necessary to find the correspondence between the sub-band indices of origin and destination. It is observed that an input sinusoid of the physical frequency Ω gives rise to a main contribution that takes place in the input subbands with the index «» Ω / Δ / ;. A sinusoid of physical frequency output

transposed pretendici ΤΩ must be the result of feeding to the synthesis subband of the index m ~ TΩ / & fs. Therefore, the appropriate index of origin values of the processing by subbands corresponding to an objective subband index m must comply with the following <I ΔΛ T (D

Fig. 15 illustrates an example of a situation corresponding to the transposition application based on subband blocks using several transposition orders in an HFR-enhanced audio codec. A stream of io bits transmitted in the decoder of core 1501 is received, which provides a decoded core signal of low bandwidth at the sampling frequency fs. The low frequency is resampled to the 2fs output sampling frequency by means of a 32-band modulated complex QMF analysis bank 1502 followed by a 64-band QMF synthesis bank (reverse QMF) 1505. the two filter banks 1502 and 1505 have the same physical resolution parameters Nfs = EfTO and the processing unit of HFR 1504 simply passes the lower unmodified subbands corresponding to the low bandwidth core signal. The high frequency content of the output signal is obtained by feeding the higher subbands of the synthesis bank

QMF of the synthesis band QMF of 64 bands 1505 with the output bands of the multiple transponder unit 1503, subjected to spectral modification and training executed by the processing unit HFR 1504. The multiple transponder 1503 takes the input of the decoded core as input. and produces a multitude of subband signals representing the analysis of 64 QMF bands of an overlap or combination of several transposed signal components. The goal is to go through

high HFR processing; each component corresponds to a physical transposition of an entire number of the core signal, (7 = 2,3,.,.).

Fig. 16 illustrates an example of prior art situation corresponding to the operation of a multi-order transposition based on subband blocks 1603 by applying a bank of separate analysis filters for each transposition order. In this case, three transposition orders 7 = 2,3,4 have to be produced and transferred in the domain of a 64-band QMF operating at a 2fs output sampling rate. The combination unit 1604 simply selects and combines the relevant subbands of each transposition factor branch to obtain a single multitude of QMF subbands to be fed to the HFR processing unit.

Consider first case 7 = 2. The objective is specifically that the processing chain of a 64-band QMF analysis 1602-2, a sub-band processing unit 1603-2 and a 64-band QMF synthesis 1505 gives place to a physical transposition of 7 = 2. Identifying these three blocks with 1401, 1402 and 1403 of Fig. 14, it is found that tysliffTO = 2, so that (1) gives rise, in the specification corresponding to 1603-2, that the correspondence between the subbands of origin and destination m is given by n = m.

As for case 7 = 3, the illustrative system includes a sample rate converter 1601-3 that converts the input sample rate by reducing it by a factor 3/2 of fs to 2fs / 3. The objective is specifically that the processing chain of the 64-band QMF analysis 1602-3, the sub-band processing unit 1603-3 and a synthesis of 64-band QMF 1505 results in a physical transposition of 7 = 3. Identifying these three blocks with 1401, 1402 and 1403 of Fig. 14,

it finds, due to resampling, that 44/44 = 3, so that (1) produces the specification corresponding to 1603-3 that the correspondence between the subbands of origin "and of destination m is given, once again, by n = m .

In the case of case T = 4, the illustrative system includes a sample rate converter 1601-4 that converts the input sample rate by reducing it by a factor of two from fs to fs / 2. The objective is specifically that the processing chain of the 64-band QMF analysis 1602-4, the sub-band processing unit 1603-4 and a synthesis of 64-band QMF 1505 results in a physical transposition of

T = 4. Identifying these three blocks with 1401, 1402 and 1403 of Fig. 14, it is found, due to resampling, that bfsllffTO= 4, so (1) produces the specification corresponding to 1603-4 that the correspondence between the origin and destination subbands m is also given by «= w.

Fig. 17 illustrates an example of a situation according to the invention of an efficient multi-order transposition operation based on subband blocks by applying a single bank of 64-band QMF analysis filters. Incidentally, the use of three separate QMF analysis banks and two sample rate converters in Fig. 16 they give rise to a rather high computing complexity, as well as certain disadvantages of implementing table-based processing due to the conversion of sampling frequency 1601-3. The present embodiments are based on the replacement of the two branches 1601-3 -> 1602-3 -> 1603-3 and 1601-4 -> 1602-4 -> 1603-4 by processing subbands 1703-3 and 1703- 4, respectively, while the branch 1602-2 - * 1603-2 remains unchanged compared to Fig 16. In this case, the three transposition orders must be executed in the domain of a filter bank with reference to Fig. 14, where 44/44 = 2. In the case r = 3, the specification

for 1703-3 given by (1) is that the correspondence between the subbands of origin "and of destination m is given by" "2w / 3. As for the case T = 4, the specifications corresponding to 1703-4 given by (1) are that the correspondence between the subbands of origin "and of destination m is given by n" ïm. To further reduce complexity, some transposition orders can be generated by copying the transposition orders already calculated or the output of the core decoder.

Fig. 1 illustrates the operation of a transponder based on subband blocks using transposition orders of 2, 3 and 4 in an HFR-powered decoder framework, such as SBR [ISO / IEC 14496-3: 2009 , Information technology - Coding of audio-visual obje cts - Part 3: Audio]. The bit stream is decoded to the time domain by the core 101 decoder and transferred to the HFR 103 module, which generates a high frequency serial from the core band core serial. Once generated, the serial generated by HFR is dynamically adjusted to match the original serial to the maximum possible through the complementary information transmitted. This adjustment is executed by the HFR 105 processor in the subband serial obtained from one or several QMF analysis banks. A typical situation is a case where the core decoder operates on a serial in the time domain sampled at half the frequency of the input and output serials, that is to say that the HFR decoder module effectively resamples the serial from the core to double of the sampling frequency. This conversion of the sampling frequency is usually obtained in the first filtering step of the serial of the core encoder by means of a 32-band QMF analysis bank 102. The subbands below the so-called crossover frequency, ie the lowest series of the 32 subbands containing all the energy of the

core encoder signal is combined with the series of subbands that carry the serial generated by HFR. Usually, the number of subbands thus combined is 64, which, after filtering by the synthesis QMF bank 106, gives rise to a signal of the core encoder with converted sampling frequency combined with the output of the HFR module.

In the transponder based on subband blocks of the HFR 103 module, three transposition orders T = 2, 3 and 4 must be produced, and transferred in the domain of a 64-sided QMF that operates at a 2fs output sampling frequency . The signal in the entry time domain is filtered by band pass in blocks 103-12, 103-13 and 103-14. This is done to make the output signals, processed by the different transposition orders, have non-overlapping spectral contents. The number of samples of the signals is also reduced (103-23,103-24) to adapt the sampling frequency of the input signals to fit the banks of analysis filters of a constant size (in this case 64). It can be noted that the explanation of the increase in the sampling frequency, from fs to 2fs, may be that the sample rate converters use factors to reduce the number of samples of 772 instead of T, where the latter would give rise to signals of subbands transposed with the same sampling frequency as the input serial. The signals with reduced number of samples are fed to separate HFR analysis filter banks (10332, 103-33 and 103-34), one for each transposition order, which produces a multitude of subband signals of complex value. These are fed to the stretch units of non-linear subbands (103-42, 103-43 and 10344). The multitude of output subbands of complex value is fed to the Fusion / Combination module 104 together with the output of the subsampled analysis bench 102. The Fusion / Combination unit simply melts the

protruding subbands of the core filter analysis bank 102 and each branch of stretch factor in a single multitude of QMF subbands to be fed to the processing unit HFR 105.

When the serial spectra of different transposition orders are adjusted so that they do not overlap, that is, the spectrum of the serial of the 7th transposition order should begin where the serial spectrum of the order 7-1 ends, the serial transposed must be of a band pass character. From there, the filters pass traditional band 103-12-10314 of Fig. 1. However, by means of a simple exclusive selection io among the subbands available through the Fusion / Combination Unit 104, the separate band pass filters are redundant and can be avoided. On the contrary, the inherent band pass characteristic provided by the QMF bank is used by feeding the different contributions of the transponder branches independently to different subband channels in 104.

It is also enough with the application of time stretching only to the bands that are combined in 104.

Fig. 2 illustrates the operation of a non-linear subband stretching unit. The block extractor 201 samples a finite frame of samples of the complex value input serial. The table is defined by an entry pointer position. This frame undergoes non-linear processing at 202 and is then poisoned by a finite length window 203. The samples thus obtained are added to the previous output samples of the overlapping unit and sum 204 where the position of the output frame is defined by an exit pointer position. The input pointer is increased by a fixed amount and the output pointer is increased by a subband stretch factor multiplied by the same amount. An iteration of this chain of operations produces an output serial whose

Duration is the subband stretch factor multiplied by the duration of the input subband serial, up to the length of the synthesis window.

While the SSB transponder used by SBR [ISO / IEC 144963: 2009, Information technology - Coding of audio-visual objects - Part 3:

Audio] takes advantage, as a whole, of the entire baseband, excluding the first subband, to generate the high band serial, a harmonic transponder generally uses a smaller part of the spectrum of the core encoder. The amount used, the so-called origin range, depends on the order of transposition, the bandwidth extension factor io and the rules applied to the combined result, e.g. if the serials generated by different transposition orders are allowed to overlap spectrally or not. As a consequence of it, only a limited part of the output spectrum of the harmonic transponder corresponding to a particular transposition order is actually used by the HFR 105 processing module.

Fig. 18 illustrates another embodiment of an example of implementation of processing for the processing of a single subband serial. The single subband serial has been subjected to some sort of decimation before or after its filtration by a bank of analysis filters that is not illustrated in Fig. 18. Therefore, the duration in time of the subband serial Unique is less than the duration in time before forming the decimation. The single subband serial is entered into a block extractor 1800, which can be identical to the block extractor 2s01, although it can also be implemented in another way. Block extractor 1800 of Fig. 18 operates using a sample / block feed value illustratively called e. The sample / block advance value can be variable or it can be fixed and is illustrated in Fig. 18 in the form of a float towards the box 1800 block extractor.

output of block extractor 1800, there is a plurality of extracted blocks. These blocks are highly overlapping, since the sample / block advance value e is significantly lower than the block length of the block extractor. An example is that the block extractor extracts the blocks of 12 samples. The first block comprises samples 0 to 11, the second block comprises samples 1 to 12, the third block comprises samples 2 to 13, and so on. In this embodiment, the sample / block feed value e is equal to 1, and there is an overlap of 11 times.

The individual blocks are entered in a 1802 poisoner to place the blocks in a window using a window function for each block. In addition, a 1804 phase calculator is included that calculates one phase for each block. The phase calculator 1804 can use the individual block before the poisoning or after the poisoning. Next, a pxky phase adjustment value is calculated and entered into a phase adjuster

1806 The phase adjuster applies the adjustment value to each sample in the block.

Moreover, the k factor is equal to the bandwidth extension factor. When, for example, the bandwidth extension in factor 2 must be obtained, then the calculated phase p is multiplied with respect to a block extracted by block extractor 1800 by factor 2 and the adjustment value applied to each sample of the block in the phase adjuster 1806 is p multiplied by 2. This is an illustrative value / rule. On the other hand, the corrected phase for the synthesis is k * p, p + (k-1) * p. Therefore, in this example the correction factor is 2 if multiplied or 1 * p if added. Other values / rules can be applied to calculate the phase correction value.

In one embodiment, the unique subband signal is a complex subband serial and the phase of a block can be calculated in a plurality of different ways. One way is to turn the sample in the center or

around the center of the block and calculate the phase of this complex sample. It is also possible to calculate the phase for each sample.

Although illustrated in Fig. 18 of the way a phase adjuster operates after the poisoning device, these two blocks can also be exchanged, so that the phase adjustment is performed on the blocks removed by the block extractor and a subsequent poisoning operation is executed. Since both operations, that is, the poisoning and the phase adjustment are multiplications of real value or complex value, these two operations can be summarized in a single operation that uses a complex multiplication factor which, in itself, is the product of a multiplication factor for phase adjustment and a poisoning factor.

The blocks with adjusted phases are entered in an overlap / sum block and amplitude correction 1808, where the poisoned blocks and with adjusted phase are added with overlap. The important thing, however, is that the sample advance / block value of block 1808 is different from the value used in block extractor 1800. In particular, the sample / block advance value of block 1808 is greater than the value e used in block 1800, whereby a time stretch of the signal provided as output of block 1808 is obtained. Consequently, the serial of processed subband provided as output of block 1808 has a length greater than the serial subband input to block 1800. When a bandwidth extension of two must be obtained, then the sample / block advance value is used, which is twice the corresponding value of block 1800. This results in a time stretch by a factor of two. When, otherwise, other time stretch factors must be used, other sample / block advance values can be used so that the output of block 1808 has the necessary time duration.

To address the overlap problem, an amplitude correction is preferably executed in order to solve the problem of the different overlaps in blocks 1800 and 1808. This amplitude correction could also be introduced in the poisoner / adjuster multiplication factor phase, although amplitude correction can also be performed after overlapping / processing.

In the previous example with a block length of 12 and a sample / block advance value in the block extractor of one, the sample / block advance value corresponding to the overlap / sum block 1808 would be equal to two, when a bandwidth extension of a factor of two is executed. This, however, resulted in an overlap of five blocks. When a bandwidth extension of a factor of three must be executed, then the sample / block advance value used by block 1808 would be equal to three and the overlap would decrease to a three-fold overlap. When a bandwidth extension of four times is to be executed, then the overlap / sum block 1808 would have to use a sample advance / block value of four, which would still give rise to an overlap of more than two blocks

Large computing savings can be obtained by restricting the input signals to the transponder branches so that they contain only the range of origin, and this at a sampling rate adapted to each transposition order. The basic block scheme of that type of system corresponding to an HFR generator based on subband blocks is illustrated in Fig. 3. The input serial of the core encoder is processed by sample number reducers that precede the HFR analysis filter banks.

The essential effect of each reducer in the number of samples is the filtering of the signal from the source range and its transfer to the analysis filter bank at the lowest possible sampling frequency. In this case, the lowest possible refers to the lowest sampling frequency suitable for further processing, not necessarily the lowest sampling frequency that prevented overlap after decimation. The conversion of the sampling frequency can be obtained in various ways. Without limiting the scope of the invention, two examples are presented; the first demonstrates the resampling executed by a multi-speed processing in the time domain and the second illustrates the resampling obtained by processing by QMF in subbands.

Fig. 4 illustrates an example of the blocks of a multi-speed reducer of the number of samples in the time domain by an order of transposition of 2. The input signal, with a bandwidth B Hz and a sampling frequency fs, is modulated by a complex exponential (401) in order to shift in the frequency the beginning of the source range of the DC frequency as follows («) = x (n) · exp ^ -i2nfs

Examples of an input signal and the spectrum after modulation are illustrated in Figs. 5 (a) and (b). The modulated signal is interpolated (402) and filtered by a complex low pass filter with bandwidth limits of 0 and B / 2 Hz (403). The spectra after the respective steps are set forth in Figs. 5 (c) and (d). Next, the filtered signal (404) is decimated and the real part of the signal (405) is computed. The results after these steps are set forth in Figs. 5 (e) and (f). In this specific example, when T = 2, B = 0.Q (on a standardized scale, that is fs = 2), a P is chosen2 from

24, to safely cover the range of origin. The reduction factor of the number of samples is then

327 64 8

P2 ~ 24 _ 3, where the fraction has been reduced by a common factor of 8. Therefore, the interpolation factor is 3 (as shown in Fig. 5 (c)) and the decimation factor is 8. Using Noble Identities ["Multirate Systems And Filter Bank," PP Vaidyanathan, 1993, Prentice Hall, Englewood Cliffs], you can run the completely left and the interpolator completely to the right of Fig. 4. In this way, modulation and filtering are carried out with the lowest possible sampling frequency and thus the computing complexity is further reduced.

Another strategy is to use the subband outputs of the sub-sampled 32-band QMF analysis bank 102 that is already present in the HFR SBR method. The subbands covering the ranges of origin corresponding to the different branches of the transponder are synthesized to the time domain by small sub-sampled QMF banks that precede the banks of HFR analysis filters. This type of HFR system is illustrated in Fig. 6. The small QMF sides are obtained by sub-sampling the original 64-band QMF bank, where the prototypical filter coefficients are found by interpolation linei of the original prototype filter. Following the notation of Fig. 6, the synthetic QMF bank that precedes the 2nd order transponder branch has Q2= 12 bands (subbands with zero base indices from 8 to 19 in the 32-band QMF). To prevent overlap in the synthesis process, the first (index 8) and the last (index 19) band are set to zero. The spectral output thus obtained is set out in Fig. 7. Note that the block-based transponder analysis filter bank has 2Q2= 24 bands, is

say the same number of bands as in the example based on the multi-speed reducer of the number of samples in the time domain (Fig. 3).

When comparing Fig. 6 with Fig. 23, it becomes clear that element 601 of Fig. 6 corresponds to the bank of analysis filters 2302 of the

Fig. 23. Moreover, the synthesis filter bank 2304 of Fig. 23 corresponds to element 602-2, and the additional filter bank 2307 of Fig. 23 corresponds to element 603-2. The block 604-2 corresponds to the block 2309 and the combiner 605 may correspond to the synthesis filter bank 2311 although, in other embodiments, the combiner may be configured to produce as output subband signals and then another filter bank may be used. of synthesis connected to the combiner. However, depending on the implementation, a certain high frequency reconstruction can be performed as explained in the context of Fig. 26 moreover, before filtering performed by the synthesis filter bank 2311 or the combiner 205, or it can be executed after the synthesis filtering executed in the synthesis filter bank 2311 of Fig. 23 or after the combiner in the block 605 of Fig. 6.

The other branches that extend from 602-3 to 604-3 or that extend from 602-T to 604-T are not exposed in Fig. 23, although they can be implemented in a similar way, but with different bank sizes of filters, where T corresponds, in Fig. 6, to a transposition factor. However, as described in the context of Fig. 27, transposition by a factor of 3 and transposition by a factor of 4 can be introduced into the processing branch consisting of element 602-2 to 604-2 so that block 604-2 not only produces a transposition in a factor of 2 but also a transposition in a factor of 3 and a factor of 4 is used, along with certain

synthesis filter bank according to what is described in the framework of Figs. 26 and 27.

In the embodiment of Fig., Q2 corresponds to Ms and Ms is equal, for example, to 12. Also, the size of the additional filter bank 603-2 corresponding to element 2307 is equal to 2Ms such as horn 24 in this embodiment.

Moreover, as outlined above, the lower subband channel and the higher subband channel of the synthesis filter bank 2304 can be fed with zeros in order to avoid overlapping problems.

io The system outlined in Fig. 1 can be considered a special simplified case of resampling outlined in Figs. 3 and 4. To simplify the arrangement, modulators are omitted. Moreover, all HFR analysis filtering is obtained using banks of 64-band analysis filters. Therefore, P2 = P3 = P4 = 64 of Fig. 3, and the reduction factors of the number of samples are 1, 1.5 and 2 for the transponder branches of 2nd, 3rd and 4th order, respectively.

An advantage of the present invention is that, in the context of the critical sampling processing of the invention, the subband serials of the 32-band analysis QMF bank corresponding to block 2302 of Fig. 23 0 601 of Fig. .6 in accordance with MPEG4 (ISO / IEC 14496-3). The definition of this analysis filter bank in the MPEG-4 Standard is illustrated in the upper portion of Fig. 25a and is illustrated in the form of a flow chart in Fig. 25b, which has also been taken from the MPEG-4 Standard. The SBR (spectral bandwidth replication) portion of this standard is incorporated herein by reference. In particular, the analysis filter bank 2302 of Fig.

or the 32-band QMF of Fig. 6 in accordance with that illustrated in Fig. 25a, in its upper portion and the flow chart of Fig. 25b.

Moreover, the synthesis filter bank illustrated in block 2311 of Fig. 23 can also be implemented as indicated in the lower portion of Fig. 25a and as illustrated in the flow chart of Fig. 25c. However, any other definition of filter bank can be applied, although at least in the case of analysis filter bank 2302, the implementation illustrated in Figs. 25a and 25b due to the robustness, stability and high quality provided by this MPEG-4 analysis filter bank consisting of 32 channels, io at least in the context of taies bandwidth extension applications such as width replication of spectral band or, in general terms, in high frequency reconstruction processing applications.

The synthesis filter bank 2304 is configured to synthesize a subset of subbands that cover the range of origin corresponding to a transponder. This synthesis is performed to synthesize the intermediate serial 2306 in the time domain. Preferably, the synthesis filter bank 2304 is a small QMF bank of real value sub-sampled.

The 2306 time domain output of this filter bank is then fed to a complex value analysis QMF bank twice the size of the filter bank. This QMF bank is represented by block 2307 of Fig. 23. This procedure allows a substantial saving of the computer complexity since only the relevant origin range is transformed to the QMF subband domain with the duplicate frequency resolution. The small banks of QMF are obtained by sub-sampling the original 64-band QMF bank, where the prototypical filter coefficients are obtained by interpolating the original prototype filter line. Preferably, the prototype filter associated with the MPEG-4 synthesis filter bank with 640 is used

samples, where the MPEG-4 analysis filter bank has a window of 320 window samples.

The processing of the subsampled filter banks is illustrated in Figs. 24a and 24b, which present flow charts. First, the following variables are determined:

«G '« «" l, =SMrtS1lbb-4 »d2H.ftTO,Ml »» where Ms is the size of the subsampled synthesis filter bank and kL represents the subband index of the first channel of the 32-band QMF bank to enter the sub-sampled synthesis filter bank. The startSubband2kL (start sub-band2kL) matrix is listed in Table 1. The floor {x} function rounds the argument x to the nearest integer towards the negative infinity.

Table 1 - y = startSubband2kL (x)

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

4444466 6 8 8 8 8 8 10 10 10 12 12 12 12 12 12

Therefore, the value Ms defines the size of the synthesis filter bank 2304 of Fig. 23 and KL it is the first channel of sub-series 2305 indicated in Fig. 23. Specifically, the value included in the phthabia base equation is defined in ISO / IEC 14496-3, section 4.6.18.3.2, which is also incorporated into this reference horn . It should be borne in mind that the value Ms undergoes increments of 4, which means that the size of the synthesis filter bank 2304 can be 4, 8, 12, 16, 20, 24, 28 or 32.

Preferably, the synthesis filter bank 2304 is a synthesis filter bank of real value. For this purpose, a series of real value subband samples Ms is calculated from the new subband samples of

complex value Ms according to the first step of Fig. 24a. For this purpose the following equation is used

F (ttx) = Re {'

Xu, / *) «P

(fc + 0.5) 191 64

In the equation, exp () denotes the complex exponential function, i is the imaginary unit and ki_ is already defined above.

• Displacement of matrix samples goes 2MS positions. The oldest 2M samples are discarded.

• The samples of reai Ms subbands are multiplied by the matrix N, that is to say that the matrix-vector product NV is computed, where

Sr · (k + 0.5) · (2 · n - Ms ) 5 <k <Ms 0 <n <2MS

The output of this operation is stored in positions 0 to 2 / V7s — 1 of matrix v.

• Samples of v are extracted according to the flow chart of Fig. 24a to create the 10M matrixs elements g.

• The samples of matrix g are multiplied by the window Cj to produce matrix w. The window coefficients Cj are obtained by the linear interpolation of the coefficients c, that is by means of the equation c, (n) = p (n) c (p (n) + 1) + (1 - /? (« )) c (μ (η)), 0 <n <\ QMs where μ (η) and p (n) are defined as whole number and component fractions of 64-n / Msrespectively. The window coefficients of c can be found in Table 4.A.87 of ISO / IEC 14496-3: 2009.

Therefore, the synthesis filter bank has a prototype window function calculator to calculate a prototype window function by subsampling or interpolation using a stored window function for a different size filter bank.

• Ms new output samples are calculated by adding the samples of the matrix w according to the last step of the flow chart of Fig. 24a.

The preferred implementation of the additional filter bank 2307 of Fig. 23 together with the flow chart of Fig. 24b is illustrated below.

• Matrix samples are displaced by 2M positions according to the first step in Fig. 24b. The 2Ms older samples are discarded and the 2Ms new samples are stored at positions 0 to 2Ms ~ 4.

• Matrix samples x are multiplied by window coefficients c2i.

The window coefficients C2i are obtained by the linear interpolation of the coefficients c, that is to say by means of the equation c2, (w) = /? («) <? (// («) + 1) + (1 - /? («)) c (// («)), 0 <n <20Afy where μ (η) and p (n) are defined as the whole number and the component fractions of 22-nl Msrespectively. The window coefficients of c can be found in Table 4.A.87 of ISO / IEC 14496-3: 2009.

Thus, the additional filter bank 2307 has a prototype window function calculator to calculate a prototype window function by subsampling or interpolation using a stored window function for a filter bank of different size.

• The samples are added according to the flow chart formula of the

Fig. 24b to generate the 4M matrixS u elements

• 2Ms new samples of complex value subbands are calculated by multiplying matrix-vector M u, where Z i-TT- (k + 0.5) (2-n-4Ms/

0 <k <2MS 0 <n <4M s

In the equation, exp () denotes the complex exponential function and i is the imaginary unit.

A diagram of a reducer of the number of samples of factor 2 is illustrated in Fig. 8 (a). The low-pass filter of actual value can now be represented as H (z) = B (z) / A (z) , where B (z) is the non-recursive part (FIR) and ^ (z) is the recursive part (IIR). However, for efficient implementation, using Noble identities to reduce computing complexity, it is convenient to distinguish a filter in which all poles have a multiplicity of 2 (double poles) in terms of A (z2). Therefore, the filter can be factored in the manner illustrated in Fig. 8 (b). Using the Identity of Noble 1, the recursive part beyond the decimator illustrated in Fig. 8 (c) can be run. The non-recursive filter B (z) can be implemented using a standard two-component polyphonic decomposition according to

N. I / V. two

B (z) = £ ό (η) ζ ~η = £ z-'E, (z2 ), where E, (z) = £ b (2n + l) zn n = Q 1 = 0 / i-Ü

Accordingly, the sample number reducer may be structured according to Fig. 8 (d). After using the Identity of Noble 1, the FIR part is computed at the lowest possible sampling frequency, as illustrated in Fig. 8 (e). From Fig. 8 (e) it is clear that the operation of FIR (delay, decimators and polyphonic components) can be considered a window-summing operation that uses an input advance of two samples. For every two input samples, a new output sample is produced, which effectively results in a reduction in the number of samples by a factor of 2.

A block diagram of the sample number reducer of factor 1.5 = 3/2 is illustrated in Fig. 9 (a). The low pass filter with reai value can also be written as follows H (z) = ß (z) / rt (z), where £ (z) is the non-recursive part (FIR) and A (z) is the recursive part (IIR). As before, for an efficient implementation, using Noble Identities to reduce computing complexity, it is advantageous to design a filter where all the poles have a multiplicity of 2 (double poles) or a multiple of 3 (triple poles)

horn A (z2) ο Α (ζ3) respectively. In this case, the double poles are chosen since the design algorithm for the low pass filter is more efficient, although the recursive part actually becomes 1.5 times more complex to implement compared to the triple approach. pole. Therefore, the filter can be factorized in the manner illustrated in Fig. 9 (b). Using the Identity of Noble 2, the recursive part can be run along the interpolator as indicated in Fig. 9 (c). The non-recursive filter B (z) can be implemented using the standard polyphonic decomposition of 2-3 = 6 components as follows

N. 5 A '-.' Or

B (z) = £ b (n) z-n = £ z ~ 'E, (z6 ), where E, (z) = £ è (6 · «+ /) z 'n n = 0 / = 0 n = 0

Therefore, the sample number reducer can be structured as in Fig. 9 (d). After using both the Noble Identity 1 and 2, the part of FIR is computed at the lowest possible sampling frequency as set forth in Fig. 9 (e). From Fig. 9 (e), it is easy to see that the even index output samples are computed using the lowest group of three polyphase filters (E0(z), E2(z), E4(z)) while odd index samples are computed from the highest group (£, (z), £3(z), ^ (z)). The operation of each group (delay chain, decimators and polyphonic components) can be considered as the operation of adding windows using an input advance of three samples. The window coefficients used in the upper group are the odd index coefficients, while the lower group uses the even index coefficients of the original filter B (z). Therefore, in the case of a group of three input samples, two new output samples are produced, which effectively results in a reduction in the number of samples of a factor 1.5.

The signal in the time domain from the core decoder (101 in Fig. 1) can also be subsampled by using

transformed from smaller subsampled syntheses in the core decoder. The use of a synthetic transformation of smaller size even offers a further reduction in computer complexity. Depending on the crossover frequency, that is, the bandwidth of the core encoder serial, the ratio of the synthesis transform table and the nominal range Q (Q <1), it gives rise to an output serial of the encoder of the core with a sampling frequency Qfs. To process the subsampled serial of the core encoder in the examples outlined in this application, all the analysis filter banks in Fig. 1 (102, 103-32, 103-33 and 103-34) must be scaled in the factor Q, as well as the sample number reducers (301-2, 301-3 and 301-T) of Fig. 3, decimator 404 of Fig. 4 and the analysis filter bank 601 of Fig. 6. Obviously, you must choose Q so that all the filter bank's tables are whole numbers.

Fig. 10 illustrates the alignment of the spectral edges of the serials of the HFR transponder with the spectral edges of the envelope setting frequency table in an encoder powered by HFR, such as SBR [ISO / IEC 14496-3: 2009 , Information technology - Coding of audiovisual objects - Part 3: Audio]. Fig. 10 (a) illustrates a stylistic graph of the frequency bands comprising the envelope adjustment table, the so-called scale factor bands, covering the frequency range from the crossover frequency kx at the cessation frequency ks. The bands of scale factors constitute the frequency grid used in an HFR-powered encoder when adjusting the energy level of the regenerated high-band frequency, that is the frequency envelope. To adjust the envelope, the energy of the serials is averaged over an entire time / frequency block limited by the edges of the scale factor bands and

the selected time edges. If the signals generated by different transposition orders are not aligned with the scale factor bands, according to that illustrated in Fig. 10 (b), anomalies may arise if the spectral energy changes drastically in the vicinity of the edge of a transposition band, since the envelope adjustment process maintains the spectral structure within a band of scale factor one. Therefore, the proposed solution consists in adapting the frequency edges of the transposed signals to the edges of the scale factor bands as set forth in Fig. 10 (c). In the Figure, the upper edge of the io signals generated by the transposition orders of 2 and 3 (7 = 2, 3) are poorly reduced, compared to Fig. 10 (b), to align the frequency edges of the transposition bands with the edges of the existing scale factor bands.

Fig. 11 illustrates a realistic situation that demonstrates the potential anomalies produced when non-aligned edges are used. 11. Fig. 11 (a) illustrates, once again, the scale factor band edges. Fig. 11 (b) illustrates the unadjusted HFR generated signals of transposition orders 7 = 2, 3 and 4 together with the serial core decoded core band. Fig. 11 (c) illustrates the serial with tight envelope when an objective flat envelope is presumed. Blocks with squared areas represent bands of scale factors with high intra-band variations, which can cause anomalies in the output serial.

Fig. 12 illustrates the situation in Fig. 11, although this time using aligned edges. Fig. 12 (a) illustrates the edges of the scale factor band, Fig. 12 (b) illustrates the serials generated by unadjusted HFR of the transposition orders T = 2, 3 and 4 together with the serial of the decoded baseband of the core and, in line with Fig. 11 (c), Fig. 12 (c) illustrates the serial

with tight envelope when an objective flat envelope is presumed. As seen in this figure, there is no non-band of scale factors with high intra-band energy variations due to the lack of alignment of the transposed serial bands and the scale factor bands, and consequently the potential anomalies

Fig. 13 illustrates the adaptation of the HFR limiting band edges, as described, for example, in SBR [ISO / IEC 14496-3: 2009, Information technology - Coding of audio-visual objects - Part 3 : Audio] to harmonic patches in an encoder powered by HFR. The limiter operates in frequency bands with a much thicker resolution than the scale factor bands, although the principle of operation is practically the same. In the limiter, an average gain value is calculated for each of the bands of the limiter. Individual gain values, ie envelope gain values for each of the scale factor bands, are not allowed to exceed the average gain value of the limiter in more than a given multiplication factor. The objective of the limiter is to suppress large variations in the gains of the bands of scale factors within each of the bands of the limiter. Although the adaptation of the bands generated by the transponder to the scale factor bands guarantees small variations of the intra-band energy within a range of scale factors, the adaptation of the edges of the limiter bands to the edges of The transponder bands, according to the present invention, handle the larger-scale energy differences between the bands processed by the transponder. Fig. 13 (a) illustrates the frequency limits of the serials generated by HFR of the transposition orders T = 2, 3 and 4. The energy levels of the different transposed serials can be substantially different. Fig. 13 (b) illustrates the frequency bands of the

limiter, which are usually of constant width on a logarithmic frequency scale. The edges of the frequency bands of the transponder are added as constant edges of the limiter and the rest of the edges of the limiter are recalculated to maintain the logarithmic relations as narrow as possible, as illustrated, for example, in Fig. 13 (c ). Although some aspects have been described in the context of an apparatus, it is obvious that these aspects also represent a description of the corresponding method, where a block or device corresponds to a step of the method or a characteristic of a step of the method. Similarly, the aspects described in the context of a step of the method also represent a description of a corresponding block or element or characteristic of a corresponding apparatus.

Other embodiments employ a mixed patching scheme as illustrated in Fig. 21, where the mixed patching method is executed within a time block. For complete coverage of the different regions of the HF spectrum, a BWE comprises several patches. in HVE, higher patches require high transposition factors within phase vocoders, which particularly deteriorate the perceptual quality of the transients.

Accordingly, the embodiments generate the highest order patches that occupy the upper spectral regions, preferably by means of a computer-efficient SSB copy patching and the lower order patches that cover the intermediate spectral regions, for which they are It aims to preserve the harmonic structure, preferably by patching HBE. The individual mixing of the patching methods can be static over time or, preferably, can be signaled in the bit stream.

In the case of the copy operation, the low frequency information can be used, as illustrated in Fig. 21. On the other hand, the data of the patches generated using HBE methods can be used as illustrated in Fig. 21. The latter leads to a less dense tonal structure for the upper patches. Apart from these two examples, any combination of copying and HBE is conceivable.

The advantages of the proposed concepts are • Improved perceptual quality of the transients • Reduced computer complexity

Fig. 26 illustrates a preferred processing chain intended for bandwidth extension, where different processing operations can be performed within the processing in non-linear subbands indicated in blocks 1020a, 1020b. The cascade of filter banks 2302, 2304, 2307 is represented, in Fig. 26, by block 1010. Moreover, block 2309 may correspond to elements 1020a, 1020b and envelope adjuster 1030 may be located between block 2309 and block 2311 of Fig. 23 or may be located after processing executed in block 2311. In this implementation, the selective processing of bands of the serial processed in the time domain, such as the extended bandwidth serial, is executed in the time domain instead of in the subband domain, which exists before of synthesis filter bank 2311.

Fig. 26 illustrates an apparatus for generating an extended bandwidth audio serial from a low band input serial 1000 according to another embodiment. The apparatus comprises a bank of analysis filters 1010, a non-linear subband processor by subbands 1020a, 1020b, an envelope adjuster subsequently connected 1030 or, in general terms, a high frequency reconstruction processor that

operates according to the high frequency reconstruction parameters such as the input of the parameter line 1040. The envelope adjuster, or in general terms, the high frequency reconstruction processor processes the signals of individual subbands for each subband channel and enter the subband signals processed by each subband channel into a bank of 1050 synthesis filters. Synthesis filter bank 1050 receives, in its input signals of the lower channels, a subband representation of the low band core decoder serial. Depending on the implementation, the low band may also derive from the outputs of the analysis filter bank 1010 of Fig. 26. The transposed subband serials are fed to the upper filter bank channels of the synthesis filter bank to perform high frequency reconstruction.

The filter bank 1050 finally outputs a serial transponder output comprising bandwidth extension in transposition factors 2, 3 and 4, and the serial provided as output of block 1050 is no longer limited in its bandwidth at the crossover frequency, that is to say the highest frequency of the serial coder of the core encoder that corresponds to the lowest frequency of the serial components generated by SBR or HFR.

In the embodiment of Fig. 26, the analysis filter bank performs a sampling twice and has a certain spacing between analysis subbands 1060. The synthesis filter bank 1050 has a spacing of synthesis subbands 1070 which, in this In this embodiment, it has twice the tamarum of the spacing of the analysis subbands, which results in a transposition contribution, as described below in the context of Fig. 27.

Fig. 27 illustrates a detailed implementation of a preferred embodiment of a non-linear subband processor 1020a in Fig. 26. The circuit illustrated in Fig. 27 receives a single subband serial as input.

108, which is processed in three “branches. The upper branch 110a is for transposition in a transposition factor of 2. The center branch of Fig. 27 indicated in 110b is for transposition in a transposition factor of 3 and the lower branch of Fig. 27 is for the transposition in a transposition factor of 4 and is indicated by reference number 110c. However, the actual transposition obtained by each processing element of Fig. 27 it is only 1 (that is, there is no transposition) in the case of branch 110a. The actual transposition obtained by the processing element illustrated in Fig. 27 corresponding to the intermediate branch 110b is equal to 1.5 and the actual transposition obtained by the lower branch 110c is equal to 2. This is indicated by the numbers between parenthesis to the left of Fig. 27, where transposition factors T. are indicated. Transpositions of 1.5 and 2 represent a first transposition contribution obtained including the decimation operations in branches 110b, 110c and a time stretch by the overlapping sum processor. The second contribution, that is to say the duplication of the transposition, is obtained by virtue of the synthesis filter bank 105, which has a spacing of synthetic subbands 107 which is double the spacing of subbands of the analysis filter bank. Therefore, since the synthesis filter bank has twice the spacing of analysis subbands, no decimation function takes place in branch 110a.

Branch 11 Ob, however, has a decimation functionality that serves to obtain a transposition of 1.5. Because the synthesis filter bank has twice the physical spacing between subbands of the analysis filter bank, a transposition factor of 3 is obtained as indicated in Fig. 27 to the left of the block extractor corresponding to the second branch 11 Ob.

Analogously, the third branch has a decimation functionality that corresponds to a transposition factor of 2, and the final contribution of different subband spacing between the analysis filter bank and the synthesis filter bank corresponds, ultimately , to a transposition factor of 4 of the third branch 110c.

In particular, each branch has a block extractor 120a, 120b, 120c and each of these block extractors can be similar to block extractor 1800 of Fig. 18. Moreover, each branch consists of a phase calculator 122a, 122b and 122c, and the phase calculator may be similar to the phase calculator 1804 of Fig. 18. By addition, each branch has a phase adjuster 124a, 124b, 124c and the phase adjuster may be similar to the phase adjuster 1806 of Fig. 18. In addition, each branch consists of a poisoner 126a, 126b, 126c, where each of these poisoners may be similar to the poisoner 1802 of Fig. 18. However elio, the poisoners 126a, 126b, 126c may also be configured to apply a Rectangular window along with a certain “zero padding”. The transposition or patch signals from each branch 110a, 110b, 110c, in the embodiment of Fig. 27, are entered into adder 128, which adds the contribution of each branch to the current subband serial to finally obtain the so-called transposition blocks at the output of adder 128. Next, an overlap-sum procedure is executed in the adder of overlap and the overlap adder 130 may be similar to the overlap / sum block 1808 of Fig. 18. The overlap adder applies an overlap-sum value of 2 e, where e is the overlap-feed value or "feed value" of block extractors 120a, 120b, 120c, and overlap adder 130 outputs the transposed serial which, in the embodiment of Fig. 27, is a single subband output corresponding to channel k, ie the channel of ^ Vfoliöv-X subband observed at the time. The processing illustrated in Fig. 27 it is executed by each sub-band of analysis or by a certain group of sub-bands of analysis and, as illustrated in Fig. 26, the serials of transposed subbands are entered into the bank of synthesis filters 1050 once processed by block 1030 to obtain , in the last instance, the serial output of the transponder illustrated in Fig. 26 at the output of block 1050.

In one embodiment, block extractor 120a of the first branch of transponder 110a extracts 10 subband samples and then converts these 10 QMF samples to polar coordinates. This output, generated by the phase adjuster 124a, is then forwarded to the poisoner 126a, which extends the output by zeros corresponding to the first and last value of the block, where this operation is equivalent to a poisoned (synthesis) with a rectangular window of length 10. Block extractor 120a of branch 110a does not perform a decimation. Therefore, the samples extracted by the block extractor are mapped against a block extracted with the same sample spacing as when they were extracted.

However, this is different in the case of branches 110b and 110c. The block extractor 120b preferably extracts a block of 8 subband samples and distributes these 8 subband samples in the extracted block with a different spacing of the subband samples. The income of non-whole subband samples corresponding to the extracted block is obtained by interpolation, and the QMF samples thus obtained, together with the interpolated samples, are converted to polar coordinates and processed by the phase adjuster. Then, once again, the poisoning is executed in the poisoner 126b in order to extend the block exit through the phase adjuster 124b by zeros in the case of the first two

samples and the last two samples, operation equivalent to a poisoned (synthesis) with a rectangular window of length 8.

The block extractor 120c is configured to extract a block with a time range of 6 subband samples and performs a decimation of a decimation factor 2, performs a conversion of the QMF samples to polar coordinates and, once again, executes an operation in the phase adjuster 124b, the output is extended again with zeros, although now with respect to the first three subband samples and the last three subband samples. This operation is equivalent to a poisoned (synthesis) with a rectangular window of length 6.

Subsequently, the outputs of the transposition of each branch are added by the sum of outputs 128 to form the combined output of QMF, and finally the combined outputs of QMF are superimposed using the overlap in block 130, where the feed value or overlap-sum step is twice the advance value of block extractors 120a, 120b, 120c described above.

One embodiment comprises a method for decoding an audio signal by using harmonic transposition based on subband blocks, which comprises filtering a decoded signal from the core by means of a bank of M-band analysis filters to obtain a series of subband signals, the synthesis of a subset of said subband signals by means of banks of subsampled synthesis filters with a reduced number of subbands, to get subsampled signals in the range of origin.

One embodiment relates to a method for aligning the edges of the spectral bands of the signals generated by HFR with the spectral edges used in a parametric process.

One embodiment relates to a method for aligning the spectral edges of the signals generated by HFR with the spectral edges of the envelope adjustment frequency table comprising: the search for the highest edge of the frequency adjustment table of envelope that does not exceed the fundamental limits of serial bandwidth generated by HFR of transposition factor T; and that uses the highest edge found as the frequency limit of the serial generated by HFR of the transposition factor T.

One embodiment relates to a method for aligning the spectral edges of the limiting tool with the spectral edges of the serials generated by HFR comprising: add the frequency edges of the serials generated by HFR to the edge table used when creating the edges of the frequency bands used by the limiting tool and force the limiter to use the frequency edges added as constant edges and adjust accordingly remaining edges.

One embodiment relates to the combined transposition of an audio serial comprising transpose orders of whole numbers in a low resolution filter bank domain where the transposition operation is executed in time blocks of subband serials.

Another embodiment relates to the combined transposition, where transposition orders greater than 2 are embedded in a transposition environment of order 2.

Another embodiment relates to the combined transposition, 25 where transposition orders greater than 3 are embedded in a transposition environment of order 3, while transposition orders less than 4 are executed separately.

Another embodiment relates to the combined transposition, where transposition orders (eg transposition orders greater than 2) are generated by replicating transposition orders calculated above (ie, especially the lower orders) including the coded core bandwidth. Any conceivable combination of existing transposition orders and core bandwidths is possible without restrictions.

One embodiment relates to the reduction of computer complexity due to the reduced number of banks of analysis filters that are necessary for transposition.

One embodiment relates to an apparatus for generating an extended bandwidth serial from an input audio serial, comprising: a patching device to introduce a patch into an input audio serial in order to obtain a first serial with patch and a second serial with patch, where the second serial with patch has a different patch frequency from the first serial with patch, where the first serial with patch is generated using a first patching algorithm and the second serial with patch is generated using a second patching algorithm and a combiner to combine the first serial with patch and the second serial with patch to obtain the wide serial Extended band.

Another embodiment relates to this apparatus according to the present invention, in which the first patching algorithm is a harmonic patching algorithm and the second patching algorithm is a non-harmonic patching algorithm.

Another embodiment relates to the previous apparatus, in which the first patching frequency is lower than the second patching frequency or vice versa.

Another embodiment relates to the preceding apparatus, in which the input serial comprises patching information and in which the patching device is configured to be controlled by the patching information extracted from the input serial to vary the First patching algorithm or the second patching algorithm according to the patching information.

Another embodiment relates to a preceding apparatus, in which the patching device fulfills the function of patching (connection) of different blocks of audio serials and in which the patching device is configured to apply the first patching algorithm. and the second patching algorithm to the same block of audio samples.

Another embodiment relates to a preceding apparatus, in which the patching device comprises, in arbitrary orders, a decimator controlled by a bandwidth extension factor, a filter bank and a stretcher for a serial subband of filter bank

Another embodiment relates to the preceding apparatus, in which the stretcher comprises a block extractor for extracting a number of overlapping blocks according to an extraction advance value, a phase adjuster or poisoner to adjust the subband sampling values in each block based on a window function or a phase correction and an overlap adder to execute an overlapping processing of poisoned blocks and with phase adjustment using an overlap advance value greater than the extraction advance value.

Another embodiment relates to an apparatus for extending the bandwidth of an audio serial comprising: a bank of filters for filtering the audio serial in order to obtain serials of subbands with reduced number of samples, a plurality of different subband processors

to process different subband serials in different ways, where subband processors perform different time stretching operations of subband serials using different stretching factors and a merger to merge the processed subband output by means of the plurality of processors of different subbands to obtain an audio serial with extended bandwidth.

Another embodiment relates to an apparatus for reducing the number of samples of an audio serial comprising: a modulator; an interpolator that uses an interpolation factor; a complex low-pass filter and a decimator that uses a decimation factor, where the decimation factor is higher than the interpolation factor.

One embodiment relates to an apparatus for reducing the number of samples of an audio serial comprising: a first filter bank for generating a plurality of subband serials from an audio serial, where a sampling frequency of the subband serial is less than a sampling frequency of an audio serial; at least one synthesis filter bank followed by an analysis filter bank to perform a conversion of the sampling frequency, where the synthesis filter bank has a different number of channels from a number of channels of the filter bank of analysis; a time stretch processor to process the serial with converted sampling frequency and a combiner to combine the serial with time stretch and a low band serial or a serial with different time stretch.

Another embodiment relates to an apparatus for reducing the number of samples of an audio serial in a non-integer sample number reduction factor comprising: a digital filter, an interpolator consisting of an interpolation factor, an element polyphasic that has

Even and odd leads and a decimator that has a decimation factor that is greater than the interpolation factor, where the decimation factor and the interpolation factor are selected such that the ratio of the interpolation factor and the decimation factor is a number not whole.

An embodiment relates to an apparatus for processing an audio serial, comprising: a core decoder that has a synthetic transform size smaller than a nominal transform size by one factor, so an output signal is generated by the core decoder that has a sample rate less than a sample rate nominal that corresponds to the nominal transform size and a post processor that has one or more filter banks, one or more time stretchers and a merger, where a number of filter bank channels of the one or more filter banks has been reduced compared to a number determined by the nominal transform size.

Another embodiment relates to an apparatus for processing a low band signal comprising: a patch generator for generating multiple patches using the low band audio signal, an envelope adjuster to adjust a signal envelope using given scale factors for adjacent scale factor bands that have edges of scale factor bands , where the patch generator is configured to run multiple patches, whereby an edge between adjacent patches coincides with an edge between the bands of adjacent scale factors of the frequency scale.

One embodiment relates to an apparatus for processing a low-band audio serial comprising: a patch generator for generating multiple patches using the low-band audio signal and an envelope adjustment limiter to limit the adjustment values of envelope

corresponding to a signal by limiting adjacent limiting bands consisting of edges of limiting bands, where the patch generator is configured to execute multiple patches so that an edge between adjacent patches matches an edge between adjacent limiting bands in a frequency scale

The process of the invention is advantageous for enhancing audio codes that are based on a bandwidth extension scheme. Especially, if optimal perceptual quality at a given bit rate is extremely important and, at the same time, if the processing power is a limited resource.

Most prominent applications are audio decoders, which are often implemented in manual devices and, therefore, operate with battery power supply.

The encoded audio signal of the present invention may be stored in a digital storage medium or it may be transmitted in a transmission medium such as a wireless transmission medium or a wired transmission medium such as the Internet.

Depending on certain implementation requirements, embodiments of the invention may be implemented in hardware or software. The implementation can be executed using a digital storage medium, for example a soft disk, a DVD, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, with electronically stored readable control signals in the same, that cooperate (or have the capacity to cooperate) with a programmable computer system in such a way that the respective method is executed.

Some embodiments according to the present invention comprise a data carrier consisting of readable control signals.

electronically, with the ability to cooperate with a programmable computing system, so that one of the methods described here can be executed.

In general, the embodiments of the present invention can be implemented in the form of a computer program product with a program code, where the program code is operative to execute one of the methods when the computer program product is executed on a computer. The program code can be stored, for example, in a carrier readable by a machine.

Other embodiments include the computer program for executing one of the methods described herein, stored in a carrier readable by a machine.

In other words, an embodiment of the method of the invention is, therefore, a computer program consisting of a program code for executing one of the methods described herein, when the computer program is run on a computer.

Another embodiment of the methods of the invention therefore consists of a data carrier (or a digital storage medium, or a computer readable medium) comprising, recorded therein, the computer program for executing one. of the methods described here.

Another embodiment of the method of the invention is, therefore, a data stream or a sequence of signals representing the computer program to execute one of the methods described herein. The data stream or the signal sequence may be configured, for example, to be transferred through a data communication connection, for example over the Internet.

Another embodiment comprises a processing means, for example a computer, or a programmable logical device, configured or adapted to execute one of the methods described herein.

Another embodiment comprises a computer that has the computer program installed therein to execute one of the methods described herein.

In some embodiments, a programmable logic device (for example an array of programmable doors) can be used to execute some or all of the functionalities of the methods described herein. In some embodiments, an array of programmable field doors can cooperate with a microprocessor to execute one of the methods described herein. In general, the methods are preferably executed by any hardware device.

The embodiments described above are merely illustrative of the principles of the present invention. It is understood that people with technical training will consider the modifications and variations of the provisions and details described here evident. Therefore, they are intended to be limited only by the scope of the following patent claims and not by the specific details presented by way of description and explanation of the embodiments set forth herein.

Literature:

[1] M. Dietz, L. Liljeryd, K. Kjórling and O. Kunz, "Spectral Band Replication, a novel approach in audio coding," at the 112th AES Convention, Munich, May 2002.

[2] S. Meitzer, R. Böhm and F. Henn, “SBR enhanced audio codées for digital broadcasting such as“ Digital Radio Mondiale ”(DRM),” at the 112th AES Convention, Munich, May 2002.

[3] T. Ziegler, A. Ehret, P. Ekstrand and M. Lutzky, "Enhancing mp3 with SBR: Features and Capabilities of the new mp3PRO Algorithm," at the 112th AES Convention, Munich, May 2002.

[4] International Standard ISO / IEC 14496-3: 2001 / FPDAM 1, “Bandwidth Extension,” ISO / IEC, 2002. Speech bandwidth extension method and apparatus Vasu lyengar et al [5] E. Larsen, RM Aarts and M. Danessis. Efficient high-frequency bandwidth extension of music and speech. At the 112th AES Convention, Munich, Germany, May 2002.

[6] RM Aarts, E. Larsen and O. Ouweltjes. A unified approach to low- and high frequency bandwidth extension. At the 115th AES Convention, New York, United States, October 2003.

[7] K. Käyhkö. A Robust Wideband Enhancement for Narrowband Speech Signal. Research Report, Helsinki University of Technology, Laboratory of Acoustics and Audio Signal Processing, 2001.

[8] E. Larsen and RM Aarts. Audio bandwidth extension - Application to psychoacoustics, Signal Processing and Loudspeaker Design. John Wiley & Sons, Ltd, 2004.

[9] E. Larsen, RM Aarts and M. Danessis. Efficient high-frequency bandwidth extension of music and speech. At the 112th AES Convention, Munich, Germany, May 2002.

[10] J. Makhoul. Spectral Analysis of Speech by Linear Prédiction. IEEE Transactions on Audio and Electroacoustics, AU-21 (3), June 1973.

[11] US Patent Application 08 / 951,029, Ohmori, et al. Audio band width extending System and method [12] US Patent 6895375, Malah, D & Cox, RV: System for frequency bandwidth extension of Narrow-band speech

[13] Frederik Nagel, Sascha Disch, “A harmonie bandwidth extension method for audio codées,” ICASSP International Conference on Acoustics, Speech and Signal Processing, IEEE CNF, Taipei, Taiwan, April 2009 [14] Frederik Nagel, Sascha Disch, Nikolaus Rettelbach, “A phase Vocoder driven bandwidth extension method with novel transient handling for audio codées,” At the 126th AES Convention, Munich, Germany, May 2009.

[15] M. Puckette. Phase-Iocked Vocoder. IEEE ASSP Conference on Applications of Signal Processing to Audio and Acoustics, Mohonk 1995., Röbel, A .: Transient detection and préservation in the phase Vocoder;

io citeseer.ist.psu.edu/679246.html [16] Laroche L., Dolson M .: “Improved phase Vocoder timescale modification of audio”, IEEE Trans. Speech and Audio Processing, voi. 7, no. 3, pp. 323-332, [17] US Patent 6549884 Laroche, J. & Dolson, M .: Phasevocoder pitch-shifting [18] Herre, J .; Faller, C .; Ertel, C .; Hilpert, J .; Hölzer, A .; Spenger, C, “MP3

Surround: Efficient and Compatible Coding of Multi-Channel Audio, ”116th Conv. Aud. Eng. Soc., May 2004 [19] Neuendorf, Max; Gournay, Philippe; Multrus, Markus; Lecomte, Jérémie; Kiss you, Bruno; Geiger, Ralf; Bayer, Stefan; Fuchs, Guillaume; Hilpert,

Johannes; Rettelbach, Nikolaus; Salami, Redwan; Schüller, Gerald; Lefebvre, Roch; Grill, Bernhard: Unified Speech and Audio Coding Scheme for High Ouality at Lowbitrates, ICASSP 2009, April 19-24, 2009, Taipei, Taiwan [20] Bayer, Stefan; Kiss you, Bruno; Fuchs, Guillaume; Geiger, Ralf; Gournay, Philippe; Grill, Bernhard; Hilpert, Johannes; Lecomte, Jérémie; Lefebvre, Roch;

Multrus, Markus; Nagel, Frederik; Neuendorf, Max; Rettelbach, Nikolaus; Robilliard, Julien; Salami, Redwan; Schüller, Gerald: A Novel Scheme for Low

Contents7

81 members in 19 offices

This recordMember, by publication datePriority claim9 more offices are in the list below

Priority claims2

Priority claims
Document	Office	Kind	Date
31212710	United States of America	P	2010-03-09
2011053315	European Patent Office (EPO)	W	2011-03-04

Members81

Family members, oldest first
Document	Office	Kind
CA2792450A1	Canada	A1	2011-09-15
CA2792452A1	Canada	A1	2011-09-15
WO2011110499A1	World Intellectual Property Organization (WIPO)	A1	2011-09-15
WO2011110500A1	World Intellectual Property Organization (WIPO)	A1	2011-09-15
TW201207841A	Taiwan Province of China	A	2012-02-16
TW201207842A	Taiwan Province of China	A	2012-02-16
AR080476A1	Argentina	A1	2012-04-11
AR080477A1This record	Argentina	A1	2012-04-11
MX2012010415A	Mexico	A	2012-10-03
AU2011226211A1	Australia	A1	2012-10-18
AU2011226212A1	Australia	A1	2012-10-18
SG183967A1	Singapore	A1	2012-10-30
MX2012010416A	Mexico	A	2012-11-23
KR20120131206A	Republic of Korea	A	2012-12-04
KR20120139784A	Republic of Korea	A	2012-12-27
EP2545548A1	European Patent Office (EPO)	A1	2013-01-16
EP2545553A1	European Patent Office (EPO)	A1	2013-01-16
CN102939628A	China	A	2013-02-20
US2013051571A1	United States of America	A1	2013-02-28
CN103038819A	China	A	2013-04-10
US2013090933A1	United States of America	A1	2013-04-11
JP2013521538A	Japan	A	2013-06-10
JP2013525824A	Japan	A	2013-06-20
HK1181180A	Hong Kong, China	A	2013-11-01
HK1181180A1	Hong Kong, China	A1	2013-11-01
AU2011226211B2	Australia	B2	2014-01-09
AU2011226212B2	Australia	B2	2014-03-27
RU2012142732A	Russian Federation	A	2014-05-27
JP5523589B2	Japan	B2	2014-06-18
TWI444991B	Taiwan Province of China	B	2014-07-11
TWI446337B	Taiwan Province of China	B	2014-07-21
EP2545553B1	European Patent Office (EPO)	B1	2014-07-30
KR101414736B1	Republic of Korea	B1	2014-08-06
KR101425154B1	Republic of Korea	B1	2014-08-13
JP5588025B2	Japan	B2	2014-09-10
ES2522171T3	Spain	T3	2014-11-13
PL2545553T3	Poland	T3	2015-01-30
CN103038819B	China	B	2015-02-18
CN102939628B	China	B	2015-05-13
MY154204A	Malaysia	A	2015-05-15
US9305557B2	United States of America	B2	2016-04-05
CA2792450C	Canada	C	2016-05-31
RU2586846C2	Russian Federation	C2	2016-06-10
US2017194011A1	United States of America	A1	2017-07-06
US9792915B2	United States of America	B2	2017-10-17
CA2792452C	Canada	C	2018-01-16
US10032458B2	United States of America	B2	2018-07-24
US2018366130A1	United States of America	A1	2018-12-20
EP3570278A1	European Patent Office (EPO)	A1	2019-11-20
US2020279571A1	United States of America	A1	2020-09-03
US10770079B2	United States of America	B2	2020-09-08
BR112012022740A2	Brazil	A2	2020-10-13
BR112012022574A2	Brazil	A2	2021-09-21
BR112012022740B1	Brazil	B1	2021-12-21
BR122021019078B1	Brazil	B1	2022-05-03
BR112012022574B1	Brazil	B1	2022-05-17
BR122021014305B1	Brazil	B1	2022-07-05
BR122021019082B1	Brazil	B1	2022-07-26
BR122021014312B1	Brazil	B1	2022-08-16
EP3570278B1	European Patent Office (EPO)	B1	2022-10-26
US11495236B2	United States of America	B2	2022-11-08
ES2935637T3	Spain	T3	2023-03-08
US2023074883A1	United States of America	A1	2023-03-09
EP4148729A1	European Patent Office (EPO)	A1	2023-03-15
PL3570278T3	Poland	T3	2023-03-20
US11894002B2	United States of America	B2	2024-02-06
US2024135939A1	United States of America	A1	2024-04-25
EP4475124A2	European Patent Office (EPO)	A2	2024-12-11
EP4475124A3	European Patent Office (EPO)	A3	2025-01-15
EP4148729B1	European Patent Office (EPO)	B1	2025-01-22
EP4148729C0	European Patent Office (EPO)	C0	2025-01-22
ES3010370T3	Spain	T3	2025-04-02
US12308036B2	United States of America	B2	2025-05-20
PL4148729T3	Poland	T3	2025-05-26
HUE070311T2	Hungary	T2	2025-05-28
EP4475124B1	European Patent Office (EPO)	B1	2025-10-29
EP4475124C0	European Patent Office (EPO)	C0	2025-10-29
EP4661004A2	European Patent Office (EPO)	A2	2025-12-10
EP4661004A3	European Patent Office (EPO)	A3	2026-02-25
ES3058766T3	Spain	T3	2026-03-12
PL4475124T3	Poland	T3	2026-03-16