Audio encoder and bandwidth extension decoder

[0001] The embodiments of the invention relate to the processing of an audio signal, in particular an audio encoder, a method for providing an output signal and a decoder for bandwidth enlargement and a method for providing an audio signal with an increased bandwidth.

[0002] Coding of audio signals adapted to hearing to reduce the amount of data for efficient recording and transmission of these signals has been accepted in many fields. Coding algorithms are known, e.g. MPEG 1/2 LAYER 3 "MP3" or MPEG 4 AAC. The coding algorithm used for this, in particular when obtaining the lowest bit rates, reduces the audio quality, which is often caused mainly by the bandwidth limits on the encoder side of the audio signal bandwidth to be transmitted. The low-pass filtered signal is encoded using a so-called core coder, and the higher frequency region is parameterized, so that they can be roughly reproduced from the low-pass filtered signal.

It is known from document WO 98 57436 to subject the audio signal to bandwidth on the encoder side in this situation and to encode only the lower band of the audio signal by means of a high quality audio encoder. The upper band is described only very roughly, i.e. by a set of parameters that allow the playback of the original spectral envelope of the upper band. On the decoder side, the upper band is then synthesized. To this end, a harmonic transposition is proposed in which the bottom band of the decoded audio signal is delivered to the filter bank. Bottom band filter banks are connected to the upper bandpass channel channels or are "transposed" and the transposed bandpass signal is subjected to a boundary matching. The synthesis filterbank belonging to the special analysis filter bank receives in this case the bandpass signals of the audio signal in the lower band and the bandpass signals with the matching lower band envelope, which have been harmonically transposed into the upper band. The synthesis filter bank output signal is an audio signal with an increased bandwidth relative to its audio bandwidth, which was transmitted from the encoder side to the decoder side with very low bit rate. In particular, the calculation of the filter bank and transposition in the field of the filter bank can become a large computational load. The synthesis filter bank output signal is an audio signal with an increased bandwidth relative to its audio bandwidth, which was transmitted from the encoder side to the decoder side with very low bit rate. In particular, the calculation of the filter bank and transposition in the field of the filter bank can become a large computational load. The synthesis filter bank output signal is an audio signal with an increased bandwidth relative to its audio bandwidth, which was transmitted from the encoder side to the decoder side with very low bit rate. In particular, the calculation of the filter bank and transposition in the field of the filter bank can become a large computational load.

[0004] Ways of reduced complexity, bandwidth extension of limited bandwidth audio signals, instead use the function of copying the low-frequency portion (LF) of the signal to the high frequency range (HF) to approximate information missing due to bandwidth limitation. Such methods are described in publications by M. Dietz, L. Liljeryd, K. Kjorling and 0. Kunz, "Spectral Band Replication, and a novel audio coding approach", 112th AES Convention, Munich, May 2002; S. Meltzer, R. Bohm and F. Henn, "SBR enhanced audio codecs for digital broadcasting such as" Digital Radio Mondiale "(DRM)", 112th AES Convention, Munich, May 2002; T. Ziegler, A. Ehret, P. Ekstrand and M. Lutzky, "Enhancing mp3 with SBR: Features and Capabilities of the new mp3 PRO Algorithm", 112th AES Convention, Munich, May 2002; International Standard ISO / IEC 14496-3: 2001 / FPDAM 1, "Bandwidth Extension," ISO / IEC, 2002, or "Speech bandwidth extension method and apparatus", Vasu Iyengar et al., U.S. Patent No. 5,455,888.

[0005] In these methods, harmonic transposition is not performed, but adjacent channels of the subband bandpass filter bank are artificially inserted into neighboring upper band filter channels. This results in a coarse approximation of the upper audio signal band. This coarse signal approximation is in the next stage refined by additional control parameters obtained from the original signal. For example, the MPEG-4 standard uses scaling factors to adjust the spectral envelope, a reverse filtering combination with the addition of noise background to match tonality and input sinusoidal signal portions to complement the tonal components.

[0006] In addition, there are further methods, such as so-called "blind bandwidth enlargement", described in E. Larsen, RM Aarts and M. Danessis, "Efficient high-frequency bandwidth extension of music and speech", AES 112th Convention, Munich , Germany, May 2002, where information on the original high frequency range is not used. In addition, there is also a method for so-called "artificial bandwidth extension" which is described in K. Kayhko, A Robust Wideband Enhancement for Narrowband Speech Signal; Research Report, Helsinki University of Technology, Laboratory of Acoustics and Audio signal Processing, 2001.

[0007] In the publication of J. Makinen et al .: AMR-WB +: a Broadcasts, IEEE, ICASSP '05 describes a method for increasing bandwidth, in which the operation of copying low-frequency components to the high-bandwidth. it is implemented by a reflection operation, for example, by upsampling a low-pass filtered signal.

[0008] Alternatively, a one -band modulation may be used that is substantially equivalent to a copying operation in the field of a filter bank. The methods that allow harmonics to increase the bandwidth usually use the step of determining the pitch (pitch tracking), the non-linear distortion stage (see, for example, the publication U. Kornagel, Spectral Widening of the excitation signal for telephone-band speech enhancement, Proceedings of the IWAENC , Darmstadt, Germany, September 2001, pp. 215-218 ") or use phase vocoders, as for example shown in US provisional patent application" F. Nagel, S. Disch "Apparatus and method of harmonic bandwidth extension in audio signals" with application number US 61/025129.

[0009] WO 02/41302 A1, for example, shows a method for improving the efficiency of a coding system that uses high frequency reconstruction methods. It indicates how to improve the overall performance of such systems by adjusting the crossover frequency between the lower band encoded by the core coder and the upper band encoded by the high frequency reconstruction circuit. In this method, the core coder must be capable of operating with different crossover frequencies on the encoder side as well as on the decoder side. Accordingly, the complexity of the core coder is increased.

[0010] Further techniques for increasing bandwidth are described, for example, in the publications "RM Aarts, E. Larsen and O. Ouweltjes, A unified approach to low- and high-threshold bandwidth extension, AES 115th Convention, New York, USA, October 2003 ", E. Larsen and RM Aarts: Audio Bandwidth Extension - Application for psychoacoustics, Signal Processing and Loudspeaker Design. John Wiley & Sons, Ltd, 2004 ", E. Larsen, RM Aarts and M. Danessis: Efficient high-frequency bandwidth extension of music and speech, AES 112th Convention, Munich, Germany, May 2002", "J. Makhoul: Spectral Analysis of Speech by Linear Prediction. IEEE Transactions on Audio and Electroacoustics, AU-21 (3), June 1973 "," United States Patent Application 08/951, 029, Ohmori et al., Audio band extending system and method "and" United States Patent 6895375,

[0011] Ways of harmonically increasing bandwidth often have high complexity, while methods of increasing bandwidth with reduced complexity are characterized by a loss of quality. In a particular case, when the low bit rate is combined with a small width of the lower band, artifacts such as roughness and timbres feel unpleasant. The reason for this is that the approximate part of HF is based on a copy operation that does not maintain harmonic relations between the tonal parts of the signal. This takes place both in the harmonic relationship between LF and HF, as well as in the harmonic relationship between consecutive areas transposed in the very part of HF. For example, in SBR, a combination of coded and duplicated components appearing at the border between low and high bands, it can generate a rough sound experience. The reason is illustrated in Fig. 18, in which the tonal parts copied from the LF range to the Hf range spectrally adhere to the tonal parts of the LF range.

[0012] Fig. 18a shows the original spectrogram 1800a of a signal consisting of three tones. Accordingly, Fig. 18b is a graph of the signal with an enlarged bandwidth corresponding to the original signal of Fig. 18a. The abscissa indicates the time and the axis of ordinates indicates the frequency. In particular, in the last tone you can see potential problems 1810 (blurred lines 1810).

[0013] If the known methods take into account harmonic relationships, this is always done based on the F0 estimation, as inter alia shown in document US 2004/028244 A1. In this case, the success of these methods depends mainly on the reliability of this estimation.

[0014] In general, known methods for increasing bandwidth provide audio signals at low bit rates, but with poor audio quality or good audio quality at high bit rates.

[0015] The object of the present invention is to provide an improved coding method for audio signals.

[0016] This object is achieved by the audio encoder defined in claim 1, the bandwidth extension decoder defined in claim 3 and claim 7 and the method defined in claims 11, 12 and 13.

[0017] An embodiment of the invention, as shown in independent claim 1, provides an audio encoder for providing an output signal using an audio input signal. The audio encoder contains a generator of transposed areas, a comparator and an output interface.

[0018] The transposed region generator is configured to generate at least one high frequency bandwidth extension signal. The high bandwidth extension signal includes a high frequency band, wherein the high frequency band of the high bandwidth extension signal is based on the low frequency audio input band. Different high frequency bandwidth signals include different frequencies in their high frequency bands if different high frequency bandwidth signals are generated.

[0019] The comparator is configured to compute a plurality of comparison parameters. The comparison parameter is calculated based on the comparison of the input audio signal and the high frequency bandwidth signal generated. Each comparison parameter from many comparison parameters is calculated based on a different frequency shifted between the input audio signal and the high frequency bandwidth signal generated. In addition, the comparator is configured to determine a comparison parameter from a plurality of comparison parameters, the determined comparison parameter meeting a predetermined criterion.

[0020] In other words, for example, the comparator may be configured to determine a comparison parameter from among a plurality of comparison parameters that meets the pre-determined criterion preferably.

[0021] The output interface is configured to provide an output signal for transmission or recording. The output signal contains an indication of the parameter based on the offset frequency corresponding to the determined comparison parameter.

[0022] In other words, the output signal may include a selected comparison parameter indicating the optimal offset frequency.

[0023] Another embodiment of the invention, as set forth in independent claim 3, provides a bandwidth extension decoder for providing a high bandwidth extension frequency signal based on an audio input signal and a parameter signal. The parameter signal contains the indication of the shifted frequency and the indication of the power density parameter. The bandwidth extension decoder includes a transposed area generator, a link module and an output interface.

[0024] The transposed region generator is configured to generate a high frequency bandwidth extension signal including a high frequency band. The high frequency band of the high frequency bandwidth extension signal is generated based on the frequency shift of the frequency band of the audio input signal. The frequency shift is based on the offset frequency.

Furthermore, the translational region generator is configured to amplify or suppress the high frequency band of the high bandwidth extension signal by a factor, respectively equal to the power density parameter or equal to the inverse value of the power density parameter.

[0026] The link module is configured to combine a high bandwidth amplification signal and an input audio signal to obtain an audio signal with an increased bandwidth.

[0027] The linking module is further configured to ignore a portion of the high frequency band of the high bandwidth extension signal, wherein the ignored high frequency bandwidth of the high bandwidth extension includes frequencies lower than the upper frequency of the audio input signal.

[0028] The output interface is configured to provide an audio signal with an increased bandwidth.

[0029] A further embodiment of the invention, as set forth in independent claim 7, provides a bandwidth extension decoder for providing an audio signal with increased bandwidth based on the input audio signal. The bandwidth extension decoder includes a transposed area generator, comparator, link module and output interface.

[0030] The transposed region generator is configured to generate at least one high frequency bandwidth extension signal comprising a high frequency band based on the input audio signal, wherein the lower frequency of the frequency band of the high frequency bandwidth generated signal is below the upper frequency of the bandwidth. input audio signal. The different high frequency bandwidth signals generated contain different frequencies in their high frequency bands if different high frequency bandwidth signals are generated.

[0031] The comparator is configured to compute multiple comparison parameters. The comparison parameter is calculated based on the comparison of the input audio signal and the high frequency bandwidth signal generated. Each comparison parameter from many comparison parameters is calculated based on a different frequency shifted between the input audio signal and the high frequency bandwidth signal generated. In addition, the comparator is configured to determine a comparison parameter from a plurality of comparison parameters, the determined comparison parameter meeting a predetermined criterion.

[0032] In other words, for example, the comparator is configured to determine a comparison parameter from among a plurality of comparison parameters that meets the pre-determined criterion preferably.

[0033] The link module is configured to combine an input audio signal and a high bandwidth extension signal to obtain an audio signal with an increased bandwidth, wherein the high bandwidth magnification signal used to obtain an audio signal with increased bandwidth is based on a shifted frequency corresponding to the designated comparison parameter.

[0034] The output interface is configured to provide an audio signal with an increased bandwidth.

[0035] Embodiments of the present invention are based on the central idea that a high frequency bandwidth signal, which is also called a transposed area, can be generated and compared to the original audio input signal. By using a different frequency of the shifted high frequency bandwidth signal or several high frequency bandwidth amplification signals with different offset frequencies, many comparison parameters can be calculated corresponding to different shifted frequencies. The comparison parameters may refer to the amount associated with the audio quality. Therefore,

[0036] The bit rate for transmission or recording of the encoded audio signal may be reduced by using a parameter indication based on the offset frequency corresponding to the determined comparison parameter, for reconstructing the high frequency band of the original audio signal and the parameter indication must be saved or transmitted.

[0037] The terms comparison parameter, xover frequency and parameter indication will be defined later.

[0038] Some preferred embodiments of the invention relate to a cross-corre- sponding comparator for comparing the input audio signal and the generated high-frequency bandwidth extension signal for calculating a comparison parameter.

[0039] Some further preferred embodiments of the invention relate to a transposed area generator that generates a high frequency bandwidth extension signal in the time domain based on single-side modulation.

[0040] An advantage of the preferred embodiments of the invention is that they provide an improved method for encoding audio signals that enables an increase in the audio quality and / or a reduced bit rate for transmission or recording.

[0041] Embodiments according to the invention will then be explained in detail with reference to the accompanying drawings, in which:

Fig. 1 is a block diagram of an audio encoder;

Fig. 2 is a schematic illustration of high frequency broadband signal generation, comparison of input audio signal and high frequency bandwidth signal generated and high frequency bandwidth signal matching of the bandwidth;

Fig. 3 is a schematic illustration of high frequency broadband signal generation, comparison of input audio signal and high bandwidth extension signal and matching of high frequency bandwidth signal power;

Fig. 4 is a block diagram of a bandwidth extension encoder;

Fig. 5 is a block diagram of a bandwidth extension decoder;

Fig. 6 shows a block diagram of a bandwidth extension decoder;

Fig. 7 is a flowchart of a method for providing an output signal based on an input audio signal;

Fig. 8 is a flowchart of a method for providing an audio signal with an increased bandwidth;

Fig. 9 is a flowchart of a method for providing an output signal based on an input audio signal;

Fig. 10 shows a method for calculating a comparison parameter;

Fig. 11 is a schematic illustration of frequency shifted interpolation;

Fig. 12 is a block diagram of a bandwidth extension decoder;

Fig. 13 is a flowchart of a method for providing an audio signal with an increased bandwidth;

Fig. 14 is a block diagram of a method for providing an audio signal with an increased bandwidth;

Fig. 15 is a block diagram of a bandwidth extension encoder;

Fig. 16a shows a three-tone spectrogram using variable shifted frequency;

Fig. 16b shows a spectrogram of an original audio signal with three tones;

Fig. 17 is a graph of the power spectrum of an original audio signal, an audio signal with an enlarged bandwidth using a constant crossover frequency and an audio signal with an increased bandwidth using a variable crossover frequency;

Fig. 18a shows a three-tone spectrogram using a known method for increasing bandwidth; and

Fig. 18b shows the spectrogram of the original audio signal with three tones.

[0042] In the following, the same reference numerals are partially used for objects and functional units having the same or similar functional features and their description in the context of the figure also applies to other figures to reduce the redundancy in the description of the embodiments.

[0043] Fig. 1 is a block diagram of an audio encoder 100 for providing an output signal 132 according to an embodiment of the invention using an audio input signal 102. The output signal is suitable for a decoder for bandwidth enlargement. Accordingly, the audio encoder is also referred to as the bandwidth extension encoder. The bandwidth enlargement encoder 100 includes a generator 110 of transposed areas, a comparator 120 and an output interface 130. The 110 transposed area generator is connected to the comparator 120, and the comparator 120 is connected to the output interface 130.

[0044] Generator 110 of the transposed areas generates at least one high frequency bandwidth 112 signal. The high bandwidth extension frequency signal 112 includes a high frequency band, wherein the high frequency bandwidth 112 of high bandwidth extension frequency band is based on the low frequency audio input bandwidth 102. If different high frequency bandwidth 112 signals are generated, different signals 112 high bandwidth enhancement frequencies contain different frequencies in their high frequency bands.

[0045] Comparator 120 calculates many comparison parameters. The comparison parameter is calculated based on the comparison of the input audio signal 102 and the high frequency bandwidth 112 generated signal. Each comparison parameter from a plurality of comparison parameters is calculated based on a different frequency shifted between the input audio signal 102 and the high frequency bandwidth 112 generated signal. In addition, the comparator 120 determines a comparison parameter from a plurality of comparison parameters, wherein the determined comparison parameter satisfies a predetermined criterion.

[0046] The output interface 130 provides an output signal 132 for transmission or recording. The output signal 132 includes an indication of the parameter based on the offset frequency corresponding to the determined comparison parameter.

Also, it may be possible to combine the generation of a plurality of high frequency bandwidth 112 signals with different shifted frequencies and to shift their high frequency bands to other different shifted frequencies. For example, five different high frequency bandwidth 112 signals are generated and each of them is shifted five times by a fixed frequency offset.

[0048] Fig. 2 shows a schematic illustration of 200 generating a high bandwidth extension signal, comparing a high bandwidth magnification signal and an audio input signal, and optional high frequency bandwidth signaling enhancements for a case where only one high frequency bandwidth signal the bandwidth is generated and shifted by different shifted frequencies.

[0049] The first schematic diagram 210 "power versus frequency" schematically illustrates the input audio signal 102. Based on this input audio signal 102, the 110 transposed region generator can generate a high bandwidth extension frequency signal 112, e.g., by shifting the low frequency band 222. the input audio signal 102 to higher frequencies (as indicated by a reference number). For example, the low frequency band is shifted by a frequency equal to the crossover frequency of the core coder, not shown in Fig. 1, which may be part of the bandwidth enlargement encoder 100, or another preset frequency.

The high frequency bandwidth 112 generated signal can then be shifted by different offset frequencies 232, and for each offset frequency 232 (as shown by reference numeral 230) a comparison parameter can be calculated by the comparator 120. The offset frequency 232 can be e.g. defined relative to the crossover frequency of the core coder with respect to another specific frequency, or can be defined as an absolute frequency value.

[0051] Next, the comparator 120 determines a comparison parameter that meets the predetermined criterion. In this way, a high frequency bandwidth signal can be determined with an offset frequency corresponding to a designated comparison parameter (as indicated by reference numeral 240).

[0052] In addition, power parameter 252 may also be determined (as indicated by reference numeral 250). The power density parameter 252 may indicate the ratio of the high frequency bandwidth of the high bandwidth extension signal to the offset frequency corresponding to the determined comparison parameter and the corresponding frequency band of the input audio signal. For example, the ratio may refer to a power density ratio or other ratio of magnitudes related to the power density of the frequency band.

[0053] Alternatively, Fig. 3 shows a schematic illustration 300 of generating a high bandwidth extension frequency signal, comparing a high bandwidth widening signal generated and an audio input signal and an optional high frequency bandwidth signaling power amplification for a case where a plurality of signals are generated high frequencies of bandwidth extension with different shifted frequencies.

[0054] In contrast to the sequence shown in Fig. 2, the generator of 110 transposed areas generates a plurality of high frequency bandwidth 112 signals with different offset frequencies 232 (as indicated by reference numeral 320). This can be done by shifting the frequency band 222 of the low input audio signal 102 to a higher frequency. The low frequency band of the input audio signal 102 may be shifted by a fixed frequency plus an individual frequency offset 232 of high frequency bandwidth broadening signals 112. The frequency constant may be equal to the crossover frequency of the core coder or other specific frequency.

[0055] The comparison parameter for each high frequency bandwidth 112 signal can then be calculated and a comparison parameter can be determined that satisfies the predetermined criterion by the comparator 120.

[0056] The power density parameter may be determined 250 as previously described.

[0057] The concepts shown in Figs. 2 and 3 may also be combined.

[0058] A comparison of the input audio signal 102 and the generated high frequency bandwidth 112 signal can be performed by cross-correlation of both signals. In this case, the comparison parameter may be, for example, a cross-correlation result for a specific frequency shifted between the input audio signal 102 and the high frequency bandwidth 112 generated signal.

The parameter of the output signal 132 may be the offset frequency itself, a quantized frequency or another quantity based on a shifted frequency.

[0060] By transmitting or storing only the parameter indication, instead of the frequency band of the high input audio signal 102, the transmission or recording rate may be reduced. By selecting a parameter based on a shifted frequency corresponding to a comparison parameter meeting a pre-determined criterion, a better audio quality can be obtained than when only audio with limited bandwidth is decoded.

[0061] The predetermined criterion may be to determine a comparison parameter from a plurality of comparison parameters indicative of, for example, high frequency bandwidth 112 signal with corresponding shifted frequency corresponding to the audio input signal 102 better than 70% high frequency bandwidth 112 signals with other shifted frequencies 112 , indicating the high frequency bandwidth 112 signal with the respective shifted frequency being one of the three best corresponding to the input audio signal 102 or indicating the best matched high frequency bandwidth 112 signal with the corresponding offset frequency. This applies to the case,

[0062] The comparison parameter may be a cross-correlation result or other size indicating how well the high frequency bandwidth 112 signal with a specific shifted frequency corresponding to the audio input signal 102.

[0063] The bandwidth enlargement encoder 100 may comprise a core coder for encoding the low frequency audio input band 102. This core coder includes a crossover frequency that may correspond to the upper frequency of the frequency coded low audio frequency band 102. The crossover frequency of the core encoder may be constant or variable over time. The introduction of a crossover frequency variable may increase the complexity of the core coder, but may also increase coding flexibility.

[0064] The method shown in Fig. 2 and / or Fig. 3 can be repeated for frequency bands higher or areas of transposition. For example, the low frequency band of the audio input signal 102 includes an upper limit frequency of 4 kHz. Accordingly, if the low frequency band of the input audio signal 102 is shifted by the upper low frequency band frequency to generate high frequency bandwidth 112, the high frequency bandwidth 112 signal comprises a high frequency band at a low frequency of 4 kHz and the upper limit frequency of 8 kHz. The method may be repeated by shifting the frequency band of low input audio signal 102 by twice the upper frequency of the low frequency band. Thus, the new high frequency broadening signal 112 generated by the bandwidth extension comprises a high frequency band with a low frequency of 8 kHz and an upper frequency of 12 kHz. This can be repeated until the desired top frequency is reached.

[0065] Alternatively, it can also be implemented by generating one high frequency bandwidth signal with a plurality of different high frequency bands.

[0066] As shown in this example, the bandwidth, low frequency audio input bandwidth and bandwidth, high frequency bandwidths of the high bandwidth extension bandwidth can be the same. Alternatively, the low frequency audio input band may be separated and shifted to generate a high bandwidth extension signal.

[0067] Determination of the high bandwidth magnification signal 112 at an offset frequency 232 corresponding to the comparison comparison parameter may leave a gap between the low frequency audio input bandwidth 102 and the high frequency bandwidth 112 high frequency bandwidth frequency response depending on the offset frequency 242. This break it may be filled by generating portions of frequencies matching this gap containing eg noise with limited bandwidth. Alternatively, the gap may be left blank, because the audio quality may not be too much.

[0068] Fig. 4 shows a block diagram of a bandwidth extension encoder 400 for providing an output signal 132 using an audio input signal 102 according to a preferred embodiment of the invention. The bandwidth extension encoder 400 includes a generator 110 of transposed areas, a comparator 120, an output interface 130, a core coder 410, a bandpass filter 420, and a parameter acquisition unit 430. The core coder 410 is connected to the output interface 130 and the transponder 110 generator 110, the transphrane area generator 110 is connected to the comparator 120, the comparator 120 is connected to the parameter acquisition unit 430, the parameter acquisition unit 430 is connected to the output interface 130, and the bandpass filter 420 it is connected to comparator 120.

[0069] The 110 transposed region generator can be implemented as a modulator for generating a high bandwidth extension frequency signal 112 based on the audio input signal 102. The comparator 120 can perform a comparison of the audio input signal 102 filtered by the bandpass filter 420 and the high frequency 112 signal generated. increasing the bandwidth, by cross-correlation. Determining the comparison parameter that meets the pre-determined criterion can also be called the delay estimation.

[0070] The output interface 130 may also have the functionality of a bit stream reformer module and may include a combiner for combining a low frequency signal provided by the core coder 410 and a parameter signal 432 including the parameter indication based on the offset frequency provided by the parameter acquisition unit 430. In addition, the output interface 130 may include an entropy encoder or differential encoder to reduce the bit rate of the output signal 132. The combiner and entropy or difference encoder may be part of the output interface 130 as shown in this example, or they may be independent entities.

[0071] The audio signal 102 may be divided into a part of low frequencies and a part of high frequencies. This can be done by means of a core coder low-pass filter 410 and a bandpass filter 420. The low-pass filter can be part of the core coder 410 or an independent low-pass filter connected to the core coder 410.

[0072] A part of the low frequencies is processed by a core coder 410, which may be an audio coder, for example, conforming to the MPEG1 / 2 Layer 3 "MP3" or MPEG 4 AAC standard, or speech encoder.

[0073] A portion of the low frequencies may be shifted by a constant value, for example, by means of one-side modulation or fast Fourier transformation (FFT, Fast Fourier transformation), such that it is above the original low frequency region in the target area of ​​the respective transposed area. This can be done by means of an independent low-pass filter connected to the 110 transposed areas generator.

[0074] At regular time intervals, a cross-correlation may be calculated between the amplitudes spectra of the windowed signal sections, between the original part of the high frequency (input audio signal) and the obtained part of the high frequencies (high frequency bandwidth signal). In this way, a delay (frequency shifted) can be determined for the maximum correlation. This delay may have the significance of a correlation coefficient in the range of the original one-stream modulation, i.e. one-side modulation may be further adjusted by a delay to maximize cross-correlation. In other words, a shifted frequency may be determined, which is also called a delay corresponding to a comparison parameter meeting the pre-determined criterion,

[0075] In addition, the ratios of the absolute values ​​of the amplitude spectra may be determined. In this way, it can be determined with which factor the obtained high frequency signal should be suppressed or amplified. In other words, a power density parameter may be determined indicating a power ratio, power density, absolute values ​​of the amplitude spectra, or other value associated with a power density ratio between the high frequency bandwidth 112 of the high bandwidth extension frequency and the corresponding frequency band of the original audio input signal 102. this can be done by means of a power density comparator, which can be part of the parameter acquisition unit 430, as in the example shown, or it can be an independent unit. To determine the power density parameter, for example, a high frequency bandwidth 112 signal may be used, which has been generated by shifting the low frequency audio input band 102 to a fixed frequency or high frequency bandwidth 112 signal corresponding to the determined comparison parameter or other high frequency bandwidth 112 generated signal. A corresponding frequency band in this case means, for example, a frequency band with the same frequency range. For example, if the high frequency band of the high bandwidth extension signal includes frequencies from 4 kHz to 8 kHz, the corresponding frequency band of the input audio signal also includes a range of 4 kHz to 8 kHz.

[0076] The resulting correction factors (offset frequency, power density parameter) corresponding to the delay and the corresponding absolute amplitude values ​​can be interpolated over time. In other words, a designated parameter for the windowed signal section (for the time frame) can be interpolated for each signal section time step.

[0077] This modulation signal (control) (parameter signal), or its parameterized representation, can be saved or sent to the decoder. In other words, the parameter signal 432 may be coupled to a frequency band of the low input audio signal 102 processed by the core coder 410 to obtain an output signal 132 that can be stored or sent to the decoder.

[0078] In addition, further parameters may be determined for adjusting, for example, the level of noise and / or tonality. This can be done by means of parameter acquisition unit 430. The next parameters can be added to the parameter signal 432.

[0079] The example shown in Fig. 4 illustrates the coder side calculation of a time modulation variable. The temporally variable modulation in this case refers to high frequency bandwidth 112 signals with other shifted frequencies. The offset frequency corresponding to the designated comparison parameter meeting the pre-determined criterion may change over time.

[0080] FIG. 5 is a block diagram of a bandwidth extension decoder 500 for providing an audio signal 532 with an increased bandwidth based on an audio input signal 502 and a parameter signal 504 according to an embodiment of the invention. The parameter signal 504 includes an indication of the offset frequency and an indication of the power density parameter. The bandwidth extension decoder 500 includes a translational field generator 510, a combiner 520, and an output interface 530. The transposition area generator 510 is connected to the combiner 520, and the combiner 520 is connected to the output interface 530.

The translational region generator 510 generates a high bandwidth extension frequency signal 512 comprising a high frequency band based on the audio input signal 502. The high frequency bandwidth signal band 512 of the high bandwidth extension frequencies is generated based on the frequency shift of the audio input band 502, wherein the frequency shift is based on a shifted frequency.

[0082] Furthermore, the translational region generator 510 amplifies or suppresses the high frequency bandwidth signal band 512 of the bandwidth extension, by a factor equal to the power density parameter value or equal to the inverse value to the power density parameter.

[0083] The combiner 520 combines a high bandwidth extension frequency signal 512 and an audio input signal 502 to obtain an audio signal 532 with an increased bandwidth, and the output interface 530 provides an audio signal 532 with an increased bandwidth.

[0084] Generation of the high bandwidth extension frequency signal 112 based on the offset frequency may allow an improved continuation of the frequency response of the audio input in the high frequency region, for example if the offset frequency is determined as previously described. This can improve the audio quality of the 532 audio signal with increased bandwidth.

[0085] In addition, the power density of the continuation of the input audio signal 502 can be performed in a very efficient manner by amplifying or suppressing the high frequency bandwidth of high frequency signal 512 by the power density parameter. As a result, normalization may not be necessary.

[0086] The translational region generator 510 can generate a high bandwidth extension frequency signal 512 by shifting the frequency response of the audio input signal 512 to a fixed frequency plus the offset frequency. The link module is configured to ignore a portion of the high frequency bandwidth signal 512 of the high bandwidth enhancement frequencies comprising the frequencies lower than the upper border frequency of the audio input signal 502.

[0087] The transposed region generator 510 may generate a high bandwidth extension frequency signal 512 in the time domain or in the frequency domain, the transposed region generator 510 may generate a high bandwidth extension frequency signal 512 based on the single-side modulation.

[0088] Additionally, the output interface may amplify the output signal before it is delivered.

[0089] Fig. 6 shows a block diagram of a bandwidth extension decoder 600 for providing an audio signal 532 with an increased bandwidth based on an audio input signal 502 and a parameter signal 504 according to a preferred embodiment of the invention. The bandwidth extension decoder 600 includes a generator of transposed areas 510, a combiner 520, an output interface 530, a core decoder 610, and a parameter acquisition unit 620. The core decoder 610 is connected to the transponder 510 generator and the combiner 520, the parameter acquisition unit 620 is connected to the transposed area generator 510 and the output interface 530, the transposed region generator is connected to the combiner 520 and the combiner 520 is connected to the output interface. 530.

[0091] The core decoder 610 may decode the received bit stream 602 and provide the audio input 502 to the transposed region generator 510 and the combiner 520. The audio input signal 502 may include an upper crossover frequency of the crossover frequency in the core decoder 610. This crossover frequency may be constant or variable over time. Variable in time means, for example, a variable for different time periods or time frames, but constant for one time interval or time frame.

[0092] The parameter acquisition unit 620 can separate the parameter signal 504 from the received bit stream 602 and provide it to the transposed generator 510. In addition, the parameter signal 504 or the acquired noise and / or tonality parameter may be provided to the output interface 530.

[0093] The transposed region generator 510 may modulate the audio input signal 502 based on the frequency shifted to obtain the high bandwidth increase signal 512 and may amplify or suppress the high bandwidth extension signal 512 based on the power density parameter included in the parameter signal 504. . A high bandwidth extension signal 512 is provided to the combiner 530. In other words, the transposed region generator 510 may modulate the audio input signal 502 based on the shifted frequency and the power density parameter to obtain the high frequency signal. This can be done, for example, in the time domain by means of single-side modulation 634 with interpolation and / or filtering 632 for each time step.

[0094] The combiner 520 combines the input audio signal 502 and the high frequency bandwidth generated signal 512 to obtain an audio signal for increasing the bandwidth 532.

[0095] The output interface 530 provides an audio signal 532 with an increased bandwidth and may further include a correction unit. The correction unit may perform tonality correction and / or noise correction based on the parameters provided by the parameter acquisition unit 620. The correction unit may be part of the output interface 530 as shown in Fig. 6, or it may be an independent unit. The correction unit may also be provided between the transposed region generator 510 and the combiner 520. In this way, the correction unit can only correct the tonality and / or noise of the high frequency bandwidth generated signal 512. Correction of the tonality and noise of the audio input signal 512 is not necessary because the input audio signal 502 corresponds to the original audio signal.

[0096] Briefly summarizing, the bandwidth extension decoder 600 can synthesise and spectrally form a high frequency signal from an audio decoder output signal or a core decoder (input audio signal) by means of a transmitted modulation function. The transmitted modulation function means, for example, a modulation function based on a shifted frequency and a power density parameter. Then, the high frequency signal and the low frequency signal can be combined and further parameters can be applied for matching the noise and tonality levels.

[0097] FIG. 7 is a flow diagram of a method 700 for providing an output signal based on an input audio signal according to an embodiment of the invention. The method includes generating 710 at least one high bandwidth extension signal, computing 720 a plurality of comparison parameters, determining a comparison parameter 730 from a plurality of comparison parameters, and providing a 740 output signal for transmission or recording.

[0098] The high frequency bandwidth signal generated comprises a high frequency band. The frequency band of the high frequency signal of the bandwidth extension is based on the frequency band of the low input audio signal. Different high frequency bandwidth signals include different frequencies in their high frequency bands if different high frequency bandwidth signals are generated.

wherein [0099] The comparison parameter is calculated based on a comparison of the input audio signal and a high bandwidth extension signal generated. Each comparison parameter from many comparison parameters is calculated based on a different frequency shifted between the input audio signal and the high frequency bandwidth signal generated.

[0100] The determined comparison parameter fulfills the pre-determined criterion.

[0101] The output signal includes an indication of a parameter based on a shifted frequency corresponding to a predetermined comparison parameter.

[0102] FIG. 8 is a flow diagram of a method 800 for providing a high bandwidth extension signal based on an input audio signal and a parameter signal according to an embodiment of the invention. The parameter signal contains the indication of the shifted frequency and the indication of the power density parameter. The method comprises a high frequency bandwidth 810 generated signal, an 820 amplification or high frequency bandwidth reduction of the high bandwidth extension bandwidth, a high frequency bandwidth amplification signal 830 and an audio input signal to obtain an audio signal with an increased bandwidth and an 840 audio signal supply. with increased bandwidth.

[0103] The high bandwidth extension signal includes a high frequency band. The high frequency band of the high frequency bandwidth extension signal is generated 810 based on the frequency shift of the frequency band of the audio input signal. The frequency offset is based on the offset frequency.

[0104] The high frequency band of the high bandwidth extension signal is amplified 820 or suppressed by a factor equal to the value of the power density parameter or equal to the inverse value to the power density parameter.

[0105] FIG. 9 is a flow diagram of a method 900 for providing an output signal based on an input audio signal. It illustrates one possibility for the coder algorithm sequence. It can also be formal mathematically described below. Real-time signals may be marked in lower-case Latin letters, Hilbert transform signals in the corresponding Greek letters, and Fourier transform signals in Latin letters large or alternatively Greek.

[0106] The input signal may be called f (n), the output signal o (n). fHFk = f * filtBFk; 1 <k <kmax indicates Fourier transform, j indicates imaginary and transformation

Hilbert H (.) Is defined as <a name="caption1"></a>&lt;p (m): = 7i (f (n)) = P ~ \ -j-sgn (&lt;0) · F (jto))

F (ja)): = xOver can be the limit frequency of the core coder, ^Nen can indicate the time.

k_{It has}x^{&gt; K €}^ may indicate the kth extension or transposed area. ak describes the edge of the band, perceptual bands associated with xOver, for example according to the Barge or ERB scale. Alternatively, ak may, for example, increase linearly, i.e. ak + i-ak = constant. The Hilbert transformation can also be calculated computationally efficiently by filtering the signal by means of a modulated low-pass filter.

[0i07] The analytic modulator function 902 may first be generated with the modulation frequencies ak and the resulting phase steps Fs with time step ^fs (Fs indicates the sampling frequency). This can be described mathematically using the following formulas:

_ _e^ r_kn * max 247 * max ²^^kn μ (η) ~ Σ^ε £ = e

4 = 1 4 = 1 [0108] The sum can only be replaced by n if Yk is independent of n.

[0 and 09] The input audio signal i02 or the actual audio signal f can be filtered low-pass to the bandwidth ak + 1-ak which can be expressed as:

<img file="PL2359366T3_D0001.tif" />

[0110] In this case, each transposed area will have the same bandwidth.

[0111] Alternatively, the input audio signal f 102 may be band pass filtered to bandwidth ak with different bandwidths that can be described as:

<img file="PL2359366T3_D0002.tif" />

[0112] The areas of the original signal may then be determined, which should be reconstructed by means of this method. These areas with limited bandwidth can be indicated as:

f_HFL = f * filt_BFi; l &lt;k &lt;k_mgand are located in intervals (ak, ak + 1).

[0113] The modulation of low pass filtered input signals 904 may be implemented in the frequency domain or in the time domain.

[0114] In the frequency domain, input signals can be first windowed, which can be described as:

ί_ξ(η) = ί (ξ · + mod (n, NFFT) +1) · win (mod (n, NFFT) +1) where NFFT is the number of containers for a fast Fourier transformation (for example 512 containers), ξ is the window number and win (.) is a window function. Windows or time frames can contain time tabs. For example, the formula given above describes the window half time tab. In this way, NeN blocks from the original signal and with them so many connected amplitude spectra Fξ (ω) with ξ &lt;N how many absolute values ​​after the Fourier transform

<img file="PL2359366T3_D0003.tif" />

describes the band edge index k of the Fourier transform.

[0115] Next, the signal is modulated in the frequency domain by shifting the FFT containers (Fast Fourier transform containers). The default Hilbert transformation is not necessary here, but allows an even formal description of the following steps:

<img file="PL2359366T3_D0004.tif" />

for ω&gt; 0 and

<img file="PL2359366T3_D0005.tif" />

[0116] In the time domain, Hilbert transform 906 firstly implements the input audio signal f 102 to generate analytical signal 908.

<img file="PL2359366T3_D0006.tif" />

then the analytical signal φυ * is modulated in single-side mode 710 using modulator u (n) 902:

Yax ι // Η = Σ ^ (η) · ^ (η) fc = 1 or ψ (η \ = &lt;p_LF(n) · μ (η) [0117] In this way, a high bandwidth extension signal may be generated, which is also called a modulated signal 910.

[0118] Next, windowing (also potentially with a tab) of the input signal 912 and the extended signal 914 and the Fourier transformation 916 is carried out:

&lt;Ρε (η} = φ_ίΡ(ξ- ~ + n) and

ψ. (η) = ψ (ξ · + mod (n, NFFT) +1) · win (mod (n, / \ / F / T) +1) where the NFFT is again the number of containers for a fast Fourier transformation (e.g. 256, 512, 1024 containers or other number between 2⁴ and 2³²), ξ is the window number, and win (.) is a window function. In this way, the NeN blocks 914 are formed from the original signal and in conjunction with this, the number of amplitude spectra Φξ (ω), Ψξ (ω) with ξ &lt;N, how many absolute values ​​Fourier transform 916.

<img file="PL2359366T3_D0007.tif" />

may describe the k edge index in the Fourier transform.

[0119] The time-domain operation is illustrated in Fig. 9.

[0120] The next step is to calculate the 720 cross-correlation Rξ, k (comparison parameter may be equal to the cross-correlation result) of the partial amplitude spectra of the original and extended signal which can be mathematically expressed as:

<img file="PL2359366T3_D0008.tif" />

wherein

<img file="PL2359366T3_D0009.tif" />

δ can indicate the maximum delay (maximum shifted frequency) for which the cross correlation is calculated. If the cross-correlation should be calculated with an offset, i.e. a small delay, and therefore large overlaps would be preferred, β = 0 should be chosen. In contrast, if you want to compensate for the fact that fewer FFT containers (containers for fast Fourier transforms) occur for large delays than for small ones, choose β = 1. Generally, you can arbitrarily choose Ο-β ^ Κ .. Alternatively or additionally, you can choose 2 &lt;^ eN; mod (J, 2&gt; 0 for selecting a cross-correlation area that is slightly larger than the transposed area. In this case, the area that is δ_ considered by cross correlation can be extended by² at both spectral ends of a given transposed area.

[0121] Based on these cross-correlation results, a cross-correlation maximum of 730 m can be determined_{£ k} : = Max (/?_IJT(v)) and delay dξ, k maximum correlation

<img file="PL2359366T3_D0010.tif" />

[0122] In addition, the 920 energy or power ratios can be determined in the areas transposed by power density spectra:

Λ-i. ,,

Σ | φ (ο&gt; |

.... ^ohZK =

L ϊ "

And a) ssy_k [0123] If a clear maximum 924 can not be determined, the delay is restored to 0 (as shown in reference numeral 922). Otherwise, the estimated delay 918 may be a delay corresponding to a cross-correlation maximum. For this purpose, the appropriate criterion of the threshold value dξ, k&gt; τ with the selected τ may be determined. Alternatively, the curvature or spectral flatness (SFN) of cross-correlation Rξ, k can be observed, for example:

<img file="PL2359366T3_D0011.tif" />

or

<img file="PL2359366T3_D0012.tif" />

[0124] Here

R: = dv ⁵'dv [0125] Delays for ξ power density parameters ζξΛ can be interpolated 926 to obtain values ​​for each time step:

Δfn): = interp (ωj; 4 (n) = interp (d ^) [0126] Then a general modulation function, amplitude modulated and frequency shifted can be generated:

for example,_k(n) = ę_k(N) en

^k&lt;m fi (n) = ^ ę_k(N) e ^m[0127] This general modulation function or parameters of the general modulation function may be provided 740 with an output signal for recording or transmission.

[0128] Additionally, additional parameters for noise correction and / or tonality correction may be determined.

[0129] The decoder modulation may be implemented by means of

<img file="PL2359366T3_D0013.tif" />

and adding k partial frequency modulations (if there is more than one area transposed). For this purpose, the general modulation function μ ^^ or μ ^ or the parameters Zk (n) and λ ^^ or c and k dξk of the general modulation function may be respectively encoded, for example, by quantization. Optionally, the sampling frequency may be reduced and hysteresis may be introduced.

[0130] Delay calculation may be omitted if there is no tonal signal, for example in silence, transient or noise locations. In these cases, the delay may be set to zero.

[0131] Fig. 10 shows in more detail example 1000 of delay determination.

[0132] For a time frame or window ξ = ΐ 1010, the delay v is set to minus λ as the start value. Then the computed 720 is the cross-correlation Rą, k (v). If v is less than Λ 1030, v is increased 1032 and then the computed 720 is the cross-correlation comparison parameter. If v is equal to or greater than Λ 1030, designated 730 may be the delay corresponding to the cross correlation maximum. If the maximum can be clearly identified 924, the determined delay is then used as parameter pa, k 918. Otherwise, the delay is set to 0 and used as parameter pa, k = 0 922.

[0133] Next, the entire method is repeated 1040 for the next time frame ξ = ξ + 1 1050. The determined delays can be interpolated 926 to obtain a parameter for each time step N.

[0134] The calculation of many comparison parameters, for example, a cross-correlation result, can also be made in parallel if multiple comparators are used. Also, processing of different time frames can be performed in parallel if the necessary devices are available several times. The loop for cross-correlation calculations can also begin at + Λ and can be reduced with each loop until ν &lt;Λ.

[0135] Fig. 11 is a schematic illustration of frequency shifted interpolation 926 of different time frames, time intervals or windows. Fig. 11a shows interpolation 1100 if the time frames do not overlap. The delay p, k is determined for the whole time frame 1110. The easiest interpolation method of the parameter for each time step 1120 can be implemented by setting parameters of all time steps 1120 of the time frame 1110 equal to the respective delay p.sub.i. At the edges of the time frame, the delay of the previous or next time frame can be selected. For example, the parameters λκ ^) to λk (n + 3) are equal to p, k, and the parameters λk (n + 4) to λk (n + 7) are equal to sul.s + 1, k.

[0136] Alternatively, the time frame delays 1110 may be interpolated linearly between time frames. For example:

<img file="PL2359366T3_D0014.tif" />

=

<img file="PL2359366T3_D0015.tif" />

· D ^_k + λ, (" ₊ 3) = ^ Α

<img file="PL2359366T3_D0016.tif" />

[0137] Accordingly, Fig. 11B depicts an example of 1150 overlapping time frames 1110. In this case, one time step 1120 is associated with more than one time frame 1110. Therefore, more than one determined delay may be associated with one step 1120. In this way, the determined delays may be interpolated 926 to obtain one parameter for each time step 1120. Thus, the determined delays may be interpolated 926 to obtain one parameter for each time step 1120. For example, delays corresponding to one time step 1120 can be interpolated linearly. For example, potential interpolations can be:

<img file="PL2359366T3_D0017.tif" />

_λ'(_{η + 1}μίαζζί

<img file="PL2359366T3_D0018.tif" />

[0138] Alternatively, the interpolation can also be performed, for example, by median filtering.

[0139] Alternatively, interpolation can also be performed by means of interpolation. Interpolation centers can be part of the parameter acquisition unit or output interface, or they can be a separate unit.

[0140] On the decoder side, the bandwidth extension can be realized by means of:

μ (η) [0141] After decoding μ (η) and φLF (n) as the output signal from the output encoder. Additionally, ψ ^) can be matched with the parameters previously obtained from the original signal for the level of tonality and / or noise.

[0142] The calculation of the general modulation function in the decoder is carried out according to one of the following * rnax »/ (n) patterns ⁺ foise (n) k = 1 and

yEn) = q&gt;_LF(n) · μ (η) + noise (n) [0143] The imaginary part of the signal can be ignored:

o (n) = ReMM) [0144] Next, as mentioned above, a tonality correction can take place, for example by means of inverse filtering.

[0145] Fig. 12 shows a block diagram of a bandwidth extension decoder 1200 for providing an audio signal 532 with an increased bandwidth based on the input audio signal 502 according to an embodiment of the invention. The bandwidth extension decoder 1200 includes a generator 1210 of the transposed areas, a comparator 1220, a combiner 1230 and an output interface 1240. The transponder 1210 generator 1210 is connected to the comparator 1220, the comparator 1220 is connected to the combiner 1230, and the combiner 1230 is connected to the output interface 1240.

[0146] The translational region generator 1210 generates at least one high frequency bandwidth 1212 signal including a high frequency band based on the audio input 502, wherein the lower frequency of the high frequency bandwidth 1212 of the high bandwidth extension bandwidth is lower than the upper border frequency. 502. Different high frequency bandwidth signals include different frequencies in the high frequency bands if different high frequency bandwidth 1212 signals are generated.

[0147] Comparator 1220 compares many comparison parameters. The comparison parameter is calculated based on the comparison of the audio input signal 502 and the high bandwidth extension signal 1212 generated. Each comparison parameter from a plurality of comparison parameters is calculated based on various frequencies offset between the audio input signal 502 and the high bandwidth extension signal generated 1212. In addition, the comparator determines the comparison parameter from many comparison parameters, where the determined comparison parameter meets the pre-determined criterion.

[0148] The link module 1230 combines the audio input signal 502 and the high bandwidth extension signal 1212 to obtain an audio signal 532 with an increased bandwidth, wherein the high bandwidth extension signal 1212 is based on the offset frequency corresponding to the determined comparison parameter.

[0149] The output interface 1240 provides an audio signal 532 with an increased bandwidth.

[0150] In comparison to the decoder shown in Fig. 5, the decoder 1200 independently determines the offset frequency. Accordingly, it is not necessary to receive this parameter with the audio input signal 502. In this way, the transmission or audio bit rate can be further reduced.

[0151] As described for Fig. 1, the transposed region generator 1210 may generate a plurality of high bandwidth magnification signals with different shifted frequencies or only one high bandwidth magnification signal that is shifted by different offset frequencies. Again, a combination of these two possibilities can also be used.

[0152] FIG. 13 is a flow diagram of a method 1300 for providing a high bandwidth extension signal according to an embodiment of the invention. The method 300 includes at least one high frequency bandwidth signal generated 1310, computing 1320 a plurality of comparison parameters, determining a comparison parameter 1330 from a plurality of comparison parameters, combining a 1340 input audio signal and a high bandwidth extension signal signal and providing 1350 an enlarged audio signal. bandwidth.

[0153] The high bandwidth extension signal includes a high frequency band based on the input audio signal. The lower frequency of the high frequency band of the audio signal with the increased bandwidth is lower than the upper frequency of the input audio signal. Different high frequency bandwidth signals include different frequencies shifted in their high frequency bands if different high frequency bandwidth signals are generated.

[0154] The comparison parameter is calculated based on a comparison of the input audio signal and the high frequency bandwidth signal generated. Each comparison parameter from many comparison parameters is calculated based on a different frequency shifted between the input audio signal and the high frequency bandwidth signal generated.

[0155] The determined comparison parameter fulfills the predetermined criterion.

[0156] The high bandwidth extension frequency signal that is connected to the input audio signal to obtain the bandwidth audio signal is based on a shifted frequency corresponding to the determined comparison parameter.

[0157] FIG. 14 is a flowchart of a method 1400 for providing an audio signal with an increased bandwidth according to an embodiment of the invention.

[0158] After receiving the 1402 bit stream including the input audio signal, the core decoder decodes 1410 the audio input signal. Based on the input audio signal, generated 1310 is a high bandwidth extension frequency signal and the computed 1320 there are many comparison parameters in the cross-correlation between the input audio signal and the high frequency bandwidth generated signal with different shifted frequencies. Then, the designated comparison parameter 1330 fulfills the pre-determined criterion, which is also called delay estimation.

[0159] Based on the offset frequency corresponding to the predetermined comparison parameter, the modulator can modulate the audio input signal 1420.

In addition, the parameter can be obtained 1430 from the received bitstream 1402 to match, e.g., the power density of the modulated signal. The modulated signal is then combined 1340 with the input audio signal. In addition, 1440 corrected may be tonality and noise of the audio signal with increased bandwidth. It can also be done before connecting to the input audio signal. Then, the supplied 1350 is audio data in the field of an audio signal with an increased bandwidth, for example, for acoustic reproduction.

[0160] In this way, calculation of the temporally variable modulation at the decoder side is performed.

[0161] Alternatively, for modulation of the input audio signal by the modulator 1420 to generate the transposed area, for example, the previously generated high frequency bandwidth extension signal may be used, or the transposed region generator may generate a high bandwidth extension frequency signal (transposed area) based on o shifted frequency corresponding to the designated comparison parameter.

[0162] In other words, if the low bit rate is more important than the low complexity on the decoder side, determining the frequency modulation of the modulators can also be performed on the decoder side. For this purpose, the algorithm shown in Fig. 9 can be performed in the decoder only with some changes. Since the original signal is not available for calculations of cross-correlation in the decoder, the correlations can be calculated between the original signal (input audio signal) and the shifted original signal (input audio signal) in the tab area. For example, the signal may be shifted between zero and ak, for example, ak divided by 2, a divided by 3 or a divided by 4. ak means the edge of the k-th band, for example, a1 is the crossover frequency of the core coder.

[0163] For example, this can occur in the same manner in the decoder as in the encoder. In the encoder, parameters for spectral shaping, noise correction and / or tonality correction can be acquired and sent to the decoder.

[0164] Accordingly, Fig. 15 is a block diagram of a bandwidth extension encoder 1500 for providing an output signal using an audio input signal. The encoder 1500 corresponds to the encoder shown in Fig. 4. However, the encoder 1500 does not provide an audio output signal 132 with an indication of the parameter based on the same offset frequency. It can only determine the power density parameter and optional parameters for tonality correction and noise correction and includes an indication of the parameter parameter in the output signal 132. However, the power density parameter (as well as other parameters, if determined) is determined based on the offset frequency corresponding to the designated comparison parameter.

[0165] For example, the power density parameter may indicate the ratio between the input audio signal 102 and the high frequency bandwidth signal with the offset frequency corresponding to the determined comparison parameter. Accordingly, the indication of a parameter that relates to a power density parameter and optionally parameters for tonality correction and / or noise correction is based on the offset frequency corresponding to the determined comparison parameter.

Another difference between the encoder 1500 and the encoder shown in Fig. 4 is that the transphrane area generator 110 generates a high bandwidth extension frequency signal in the same manner as the generator of the transposed decoder 1400 area. In this way, the encoder 1500 and the decoder. they can obtain the same offset frequencies and, therefore, the parameters acquired by the encoder 1500 are valid for the transposed areas generated by the decoder.

[0167] The apparatus and method for increasing the bandwidth of time domain audio signals may use time-varying modulators. In other words. The transposed area can be generated with a variable limit frequency, for example, for each time step, each time frame, time frame part or for time frame groups.

[0168] The described method of increasing the bandwidth of an audio signal may be used on the encoder side and on the decoder side, as well as only on the decoder side. In contrast to known methods, the described new method can realize so-called harmonics of bandwidth extension without having to obtain information about the fundamental frequency of the audio signal. In addition, in contrast to the so-called harmonic bandwidth extension, as for example shown in US provisional patent application "F. Nagel, S. Disch "Apparatus and method of harmonic bandwidth extension in audio signals" "with application number US 61/025129, which are implemented by phase vocoders, the spectrum may not be stretched, and therefore the density may not change. To ensure harmony, correlations between the plus and basic bands are used. This correlation can be calculated in the encoder as well as in the decoder, depending on the computation needs and complexity of memory and data rate.

[0169] For example, the bandwidth extension itself may be implemented using amplitude modulation (AM) and frequency shift by means of single-side modulation (SSB) with a plurality of free, adaptive single-time, time-varying carriers. The subsequent post-processing according to additional parameters may attempt to approximate the spectral envelope and the noise level as well as other properties of the original signals.

[0170] The new method of signal transformation can avoid problems arising due to simple copying and reflection operations by correct harmonically continuing the spectrum using the XOver time-bound limiting frequency frequency between the low frequency (LF) and high frequency (HF) regions, as well as between the following areas high frequencies, so-called transposed areas. These limit frequencies are selected in such a way that the generated transposed areas match the existing harmonic grid existing in the original signal as best as possible.

[0171] Fig. 16 shows a modulator with 3 time-varying amplitudes and limiting frequencies, by means of which 3 regions transposed by single-band modulation of base bands can be generated. Fig. 16a shows a graph of the spectrum of the signal with enlarged bandwidth using time-varying threshold frequencies 1610. Fig. 16b is a graph 16000 of the spectrum of an audio signal of three tones. In comparison to the spectrogram shown in Fig. 18b, lines 1620 are much less blurred.

[0172] Fig. 17 illustrates the effect using the 1700 period diagram. The power density spectrum of the third tones of the audio signal is shown as the original 1710 with a fixed limit frequency of 1720 and with a variable limiting frequency of 1730. Unlike using the constant frequency of 1720, the harmonic structure is maintained by using the variable frequency limit 1730.

[0173] Due to the harmonic continuation of the spectrum, the problems of the transition points can be avoided both between the base band (core coder) and the extended bandwidth, and between successive transposed areas. Without F0 estimation as requirements for the function of the system, arbitrary signals can be continued harmonically, without the existence of audible artifacts, or as a result of disturbance of harmony or by the existence of transient sound events.

[0174] Some embodiments of the invention relate to a method suitable for all audio applications in which full bandwidth is not available. For example, for broadcasting audio content, such as in the case of a digital radio, an internet stream or an audio communication application, the described method may be used.

[0175] An embodiment of the invention relates to a bandwidth extension decoder for providing an audio signal with an increased bandwidth based on an input audio signal and a parameter signal, as shown in independent claim 3.

[0176] Some further preferred embodiments according to the invention relate to the bandwidth enlargement decoder described before, wherein the generator of the transposed areas is configured to amplify or suppress the high frequency band of the high bandwidth extension signal by a factor equal to the power density parameter value or the inverse value to a power density parameter, wherein the indication of the power density parameter is included in the input audio signal.

[0177] In particular, it should be noted that depending on the conditions, the method according to the invention may also be implemented programmatically. The implementation may be on a 5-digit storage medium, in particular a floppy disk or a CD with electronically readable control signals, capable of cooperating with the programmable computer system, so that an appropriate method is implemented. Generally, the invention also consists of a computer program product with a program code recorded on a machine-readable carrier for carrying out the method according to the invention when the computer program product 10 is executed on a computer. In other words, the invention may thus be implemented as a computer program with a program code for carrying out the method defined in one of claims 11 to 13,

Fraunhofer-Gesellschaft zur Forderung der angewandten Forschung e. V., Germany

Proxy:

EP 2 359 366 B1-15099