TR201900906T4 - Multi-channel audio coding using complex prediction and explicit prediction signaling. - Google Patents

Abstract

An audio encoder and an audio decoder are disclosed based on combining two audio channels 201, 202 to obtain a residual signal 205 that can be derived using a first combination signal 204 as a middle signal and a predicted side signal derived from the middle signal. . The first combination signal and the prediction residual signal are encoded 206 and written to a data stream 213 with predictive information 206 derived by an optimizer 207 based on an optimization target 208. A decoder uses the prediction residual signal, the first combination signal, and predictive information to derive the decoded first channel signal and the decoded second channel signal. In an encoder sample or a decoder example, a real-virtual transform can be applied to estimate the imaginary part of the spectrum of the first combination signal. While the real valued first combination signal is multiplied by a real part of the complex predictive information to calculate the prediction signal used in the derivation of the prediction residual signal; The predicted imaginary part of the first combination signal is multiplied by an imaginary part of the complex predictive information.

Description

DESCRIPTION MULTI-CHANNEL AUDIO CODING USING COMPLEX PREDICTION AND OPEN PREDICTION SIGNALING The present invention relates to audio processing, particularly to multi-channel audio processing of a multi-channel signal having two or more channel signals. In the field of multi-channel or stereo processing, the application of a method known as mid/side (M/S) stereo coding is known. In this concept, a combination of the left or first audio channel signal and the right or second audio channel signal is created to obtain a middle or mono signal (M). In addition, a difference is created between the left or first channel signal and the right or second channel signal to obtain the side signal (S). Since the side signal will be quite small when the left and right signals are very similar, this mid/side coding method results in a significant coding gain. In general, as the range of values to be quantized/entropy-coded decreases, the coding gain at the quantizer/entropy encoder stage increases. Consequently, as the side signal decreases, the coding gain for a PCM (Pulse Code Modulation) or Huffman-based or arithmetic entropy encoder increases. However, there are some cases where center/side coding does not result in coding gain. This may occur when the signals in both channels are phase-shifted, for example, by 90°. Then, the center and side signals may have very similar ranges, so encoding the center and side signals using an entropy encoder will not result in coding gain but will instead result in an increased bit rate. Therefore, to disable mid/side coding across bands, frequency-selective mid/side coding can be applied, for example, where the side signal is not significantly reduced relative to the original left signal. While the side signal is zero when the left and right signals are identical, resulting in maximum coding gain due to the elimination of the side signal, the situation is different when the mid and side signals are waveform-identical, with the only difference between them being their overall amplitude. In this case, assuming no phase shift from the side signal to the mid, the side signal increases significantly, while the mid signal does not experience a significant decrease in value across the range. When this occurs in a particular frequency band, mid/side coding can be disabled due to the lack of coding gain. Mid/side coding can be implemented frequency-selectively or, alternatively, in the time domain. Alternative multichannel coding techniques exist that rely not on a waveform approach like center/side coding, but rather on parametric processing based on specific stereophonic cues. These techniques are known as "double-cue coding," "parametric stereo coding," or "MPEG surround coding." In this technique, specific cues are calculated for multiple frequency bands. These cues include inter-channel level differences, inter-channel coherence measurements, inter-channel time differences, and/or inter-channel phase differences. These approaches assume that a multichannel effect experienced by the listener does not necessarily depend on the detailed waveforms of the channels, but rather on accurate frequency-selective cues or inter-channel information. This means that a rendering machine must be careful when rendering multi-channel signals that accurately reflect the cues, but the waveforms are not of decisive importance. This approach can be particularly complex when the decoder must apply decorrelation processing to artificially obtain decorrelated stereo signals from each other, even though all these channels are derived from one or the same downmix channel. Depending on the application, decorrelators for this purpose are complex and can cause artifacts, especially in the interlaced signal portions. Furthermore, the parametric coding approach, unlike waveform coding, is lossy, resulting in unavoidable information loss due to both typical quantization and the consideration of stereophonic cues rather than specific waveforms. This approach is very low. This allows for higher bit rates, but may require compromises in quality. Recent improvements have been made to Unified Speech and Audio Coding (USAC), as shown in Figure 7. A core decoder 700 decodes the encoded stereo signal, which may be center/side encoded on input 701. The core decoder outputs a center signal on line 702 and a side or residual signal on line 703. Both signals are converted to the QMF domain by QMF (Quadrature Mirror Filter) filter banks 704 and 705. Next, a left channel signal 707 and a right channel signal 708 are applied. The low-band signals are then passed to a spectral band replication (SBR) decoder 709, which produces wide-band left and right signals on lines 710 and 711, which are then converted to a time domain by QMF synthesis filter banks 712, 713 to obtain wide-band left and right signals L, R. Figure 7 shows the situation where the MPEG surround decoder 706 performs a center/side decoding. Alternatively, the MPEG surround decoder block 706 performs a stereophonic cue-based parametric decoding to produce stereo signals from a single mono core decoder signal. The MPEG surround decoder 706 also performs inter-channel level differences, Using parametric information such as inter-channel coherence measurements or other such inter-channel information parameters, it can also generate multiple low-band output signals to be input to the SBR decoder block 709. When the MPEG surround decoder block 706 performs the mid/side decoding shown in Figure 7b, a real-gain coefficient g can be applied; DMX/RES and L/R are the downmix/residual and left/right signals, respectively, represented in the complex mixed QMF region. Using a combination of block 706 and block 709 results in only a slight increase in numerical complexity compared to using a basic stereo decoder, because the complex QMF representation of the signal is already available as part of the SBR decoder. In a non-SBR structure, QMF-based stereo coding, as described in the USAC context, in this example, is used for 64-band analysis. This can result in a significant increase in computational complexity due to the required QMF blocks, which require blocks and 64-band synthesis blocks. These filter banks should only be added for stereo coding purposes. The MPEG USAC system under development generally has high bitrate coding modes where SBR is not used. The following two MPEG USAC documents provide examples of multichannel audio coding/decoding schemes in which a signal other than a downmix/sum/mono signal is predicted by means of a complex-valued prediction coefficient: HEIKO PURNHAGEN ET AL.: "Technical description of proposed Unified Stereo Coding in USAC", 90th MPEG MEETING; 26-10- MAX NEUENDORF (EDITOR): "WD5 of USAC", 90th MPEG MEETING; 12-08), page 1-146; The disclosure discloses methods and apparatus for decoding stereo coding using complex prediction in the field. In one embodiment, a decoding method for obtaining an output stereo signal from an input stereo signal encoded with complex prediction coding and comprising first frequency domain representations of two input channels comprises the following upmixing steps: (i) computing a second frequency domain representation of a first input channel; and (ii) computing an output channel based on the first and second frequency domain representations of the first input channel, the first frequency domain representation of the second input channel, and a complex prediction coefficient. The upmixing may be delayed responsively to control data. The disclosure discloses a parametric stereo upmix apparatus that generates a left signal and a right signal from the downmix signal. Said parametric stereo upmix is characterized by comprising means for predicting a difference signal comprising a difference between the left signal and the right signal based on the mono downmix signal scaled by a prediction coefficient. Said prediction coefficient is derived from spatial parameters. Said stereo upmixer further comprises an arithmetic means for deriving the left signal and the right signal based on a sum and a difference of the mono downmix signal and said difference signal. An object of the present invention is to provide an improved audio processing concept that results in good audio quality and/or reduced digital complexity while providing high coding gain. This object is achieved by an audio decoder as in claim 1, an audio coder as in claim 4, an audio decoding method as in claim 7, an audio coding method as in claim 8, or a computer program as in claim 9. The present invention is based on the finding that the coding gain in a high-quality waveform coding approach can be significantly increased for a second combination signal using a first combination signal, where both signals are derived from the original channel signals using a combination rule such as the mid/side combination rule. Because the inventive prediction is a waveform-based coding approach rather than a parameter-based dual or multi-channel coding approach, it has been found that this prediction information is calculated by a predictor in the audio encoder such that an optimization objective is achieved, incurring only a small overhead, but resulting in a significant reduction in the bit rate required for the side signal without any loss of audio quality. To reduce computational complexity, it is preferable to implement frequency-domain coding in which the prediction information is derived band-selectively from the frequency-domain input. The transformation algorithm for converting a time domain representation to a spectral representation is preferably a critically oversampled operation such as the modified discrete cosine transform (MDCT) or the modified discrete sine transform (MDST). This differs from the complex transform in that only real and imaginary values are computed, while in the complex transform, the real and complex values of a spectrum are oversampled by a factor of 2, resulting in oversampling. Preferably, a transform based on overlapping presentation and deletion is used. MDCT, in particular, is such a transform, and it allows cross-fading between subsequent blocks without any overhead due to the well-known Time Domain Aliasing Cancellation (TDAC) property achieved through decoder-side overlap-add processing. Preferably, the prediction information computed in the encoder, transmitted to the decoder, and used in the decoder includes an imaginary part that reflects the phase differences between two audio channels, advantageously chosen from an order between 0° and 360°. The computational complexity is significantly reduced when a purely real-valued transform, or in general, a transform that provides only a real spectrum or only an imaginary spectrum, is applied. To utilize this virtual prediction information, which indicates the phase shift between a specific band of the left signal and a corresponding band of the right signal, a real-to-virtual converter, or a virtual-to-real converter depending on the implementation of the conversion, is provided in the decoder to calculate a predictive residual signal from the first combination signal that has been phase-rotated with respect to the original combination signal. This phase-rotated predictive residual signal can then be combined with the predictive residual signal transmitted in the bitstream to reproduce a side signal, which can then be combined with a middle signal to obtain the decoded left channel in a specific band and the decoded right channel in that band. To improve audio quality, the same real-to-virtual or virtual-to-real converter applied at the decoder is also applied at the encoder when the predictive residual signal is calculated in the encoder. The present invention is advantageous in that it provides improved audio quality and a reduced bit rate when compared to systems having the same bit rate and the same audio quality. Advantages are also gained in terms of the computational efficiency of combined stereo coding, which is useful in the MPEG USAC system at high bit rates, where SBR is generally not used. In these approaches, instead of processing the signal in the complex mixed QMF domain, residue-based predictive stereo coding is applied in the local MDCT domain of the underlying stereo transform encoder. In accordance with one aspect of the present invention, the present invention includes an apparatus or method for generating a stereo signal by complex prediction in the MDCT domain, wherein the complex prediction is performed in the MDCT domain using a real-to-complex transform; wherein said stereo signal may be an encoded stereo signal at the encoder side when the stereo signal generating device or method is implemented at the decoder side, or alternatively, a decoded/transmitted stereo signal. In the following, preferred embodiments of the present invention will be described with reference to the attached drawings, of which: Figure 1 is a schematic of an audio decoder; Figure 2 is a block diagram of an audio encoder; Figure 3a shows an implementation of the encoder calculator of Figure 2; Figure 3b shows an alternative implementation of the encoder calculator of Figure 2; Figure 3c shows a mid/side combination rule to be implemented at the encoder side; Figure 4a shows an implementation of the decoder calculator of Figure 1; Figure 4b shows an alternative implementation of the decoder calculator in the form of a matrix calculator; Figure 4c shows a mid/side reverse combination rule corresponding to the combination rule shown in Figure 3c; Figure 5a shows an embodiment of a voice encoder operating in the frequency domain, preferably a real-valued frequency domain; Figure 5b shows an embodiment of a voice decoder operating in the frequency domain; Figure 6a, in accordance with an embodiment of the present invention. Figure 6b shows an alternative implementation of an audio encoder operating in the MDCT domain and using a real-to-virtual transform; Figure 6b shows an audio decoder operating in the MDCT domain and using a real-to-virtual transform in accordance with an embodiment of the present invention; Figure 7a shows an audio postprocessor using a stereo decoder and a downstream SBR decoder; Figure 7b shows a mid/side upmix matrix; Figure 8a shows a detailed view of the MDCT block of Figure 6a; Figure 8b shows a detailed view of the MDCT*1 block of Figure 6b; Figure 9a shows an implementation of an optimizer operating at reduced resolution with respect to the MDCT output; Figure 9b shows a representation of an MDCT spectrum and the corresponding low resolution bands for which predictive information is calculated; Figure 10a shows an implementation of the real-to-virtual converter of Figure 6a or 6b, and Figure 10b shows a possible implementation of the virtual spectrum calculator of Figure 10a. Figure 1 shows an audio decoder for decoding an encoded multi-channel audio signal obtained on an input line 100. The encoded multi-channel audio signal includes an encoded first combination signal, an encoded predictive residual signal, and predictive information generated using a combination rule for combining a first channel signal representing the multi-channel audio signal and a second channel signal. The encoded multi-channel signal may be a data stream, such as a bit stream having three components in multiplexed form. Additional side information may be included in the encoded multi-channel signal on line 100. The signal is input to the input interface 102. The input interface 102 may be implemented as a data stream demultiplexer that extracts the encoded first combination signal on line 104, the encoded residual signal on line 106, and the prediction information on line 108. The prediction information is preferably a factor having a real part that is not equal to zero and/or an imaginary part that is different from zero. The encoded combination signal and the encoded residual signal are input to a signal decoder 110 to decode the first combination signal to obtain a decoded first combination signal on line 112. Additionally, the signal decoder 110 is configured to decode the encoded residual signal to obtain a decoded residual signal on line 114. Depending on the encoding process on the side of an audio encoder, the signal decoder may include an entropy decoder such as a Huffman decoder, an arithmetic decoder, or any other entropy decoder, and a post-coupled dequantization (recovery of lost detail in a quantized image) step to perform a dequantization operation that maps to a quantizer operation in an associated audio encoder. The signals on lines 112 and 114 are input to a decoding calculator 115, which outputs the first channel signal on line 117 and the second channel signal on line 118, where the two signals are two channels of a multi-channel audio signal or stereo signals. For example, if the multi-channel audio signal contains five channels, then the two signals are two channels from the multi-channel signal. To fully encode such a five-channel multichannel signal, two decoders can be implemented, as shown in Figure 1, where the first decoder processes the left channel and right channel, the second decoder processes the left surround channel and right surround channel, and a third mono decoder can be used to apply mono-coding to the center channel. However, other groupings and combinations of waveform and parametric encoders can also be implemented. An alternative way to generalize the prediction scheme to more than two channels is to process three (or more) signals simultaneously, i.e., to predict a 3rd combination signal from signals 1 and 2 using two prediction coefficients, much like the "two-to-three" approach in MPEG Environment. The decoder calculator 116 is configured to calculate a decoded multi-channel signal having a decoded first channel signal 117 and a decoded second channel signal 118 using the decoded residual signal 114, the prediction information 108, and the decoded first combination signal 112. Specifically, the decoded first channel signal and the decoded second channel signal are combined by the combination rule when generating the first combination signal and the prediction residual signal. The encoder is configured to process the multi-channel signal input to the encoder such that it has at least an approximation of a first channel signal and a second channel signal. Specifically, the prediction information in line 108 can be a non-zero real-valued part and/or a non-zero imaginary part. The decoder calculator 116 can be implemented in different ways. A first implementation is shown in Figure 4a. This implementation includes a predictor 1160, a combination signal calculator 1161, and a combiner 1162. The predictor receives the decoded first combination signal 112 and the prediction information 108 and outputs a prediction signal 1163. Also specifically, the predictor 1160 combines the prediction information 108 into the decoded first combination signal 112 or a signal derived from the decoded first combination signal. The derivation rule for deriving the signal to which the prediction information 108 is applied may be a real-to-virtual transformation or equivalently a virtual-to-real transformation or a weighting operation, or may also be a phase shift operation or a combined weighting/phase shift operation, depending on the application. The prediction signal 1163 is input to the combination signal calculator 1161 together with the decoded residual signal to calculate the decoded second combination signal 1165. Both signals 112 and 1165 are combined to obtain the decoded multi-channel audio signal having the decoded first channel signal and the decoded second channel signal on lines 1166 and 1167, respectively. The decoder calculator is implemented as a matrix calculator 1168, which takes as input the decoded first combination signal or M signal, the decoded residual signal or D signal, and the prediction information 108. The matrix calculator 1168 applies a transformation matrix, denoted 1169, to the M, D signals to obtain output signals L, R, where L is the decoded first channel signal and R is the decoded second channel signal. The notation in Figure 4b resembles a stereo notation with one left channel (L) and one right channel (R). This notation is implemented for clarity, but those skilled in the art will understand that the L, R signals can be any combination of two channel signals within a multi-channel signal having more than two channel signals. The matrix operation is a kind of "monostable" 4a" to the matrix calculation, where the inputs to the circuit in Figure 4a and the outputs from the circuit in Figure 4a are identical to the inputs to the matrix calculator 1168 and the outputs from the matrix calculator 1168. Figure 4c shows an example of an inverse combination rule implemented by the combiner 1162 in Figure 4a. Specifically, this combination rule is similar to the decoder-side combination rule in well-known middle/side coding, where L = M + S and R = M - S. It should be noted that the signal S used by the inverse combination rule in Figure 4c is the signal calculated by the combination signal calculator, i.e., a combination of the prediction signal on line 1163 and the decoded residual signal on line 114. Again, in this description, the signals on lines It should be noted that in some cases, the lines are named using the reference numbers used for them, and in some cases, the reference numbers used for the lines themselves. Therefore, the notation is that a line having a certain signal represents the signal itself. A line may be a physical line in a hardware-integrated implementation. In a computer-aided implementation, there is no physical line; instead, the signal represented by the line is transmitted from one computational module to another. Figure 2 shows an audio coder for encoding a multi-channel audio signal 200 having two or more channel signals, where the first channel signal is shown at 201 and the second channel at 202. When the prediction residual signal 205 is combined with a prediction signal derived from the first combination signal 204, and the prediction information 206 is combined with a first combination signal 204, When the first combination signal and the second combination signal are input to an encoder calculator 203 for calculating a first combination signal 204 and a prediction residual signal 205 using the first channel signal 201 and the second channel signal 202 and the prediction information 206, so that the first combination signal and the second combination signal can be derived from the first channel signal 201 and the second channel signal 202 using a combination rule. The prediction information is generated by an optimizer 207 for calculating the prediction information 206 such that the prediction residual signal satisfies an optimization goal 208. The first combination signal 204 and the residual signal 205 are input to a signal encoder 209 for coding the first combination signal to obtain an encoded first combination signal 210 and encoding the residual signal 205 to obtain an encoded residual signal 211. Both encoded signals 210, 211 are input to an output interface for combining the encoded first combination signal 201 with the encoded prediction residual signal 211 and the prediction information 206 to obtain an encoded multi-channel signal 213 similar to the encoded multi-channel signal 100 input to the input interface 102 of the audio decoder shown in Fig. 1. Depending on the application, the optimizer 207 either receives the first channel signal 201 and the second channel signal 202, or receives the first combination signal 214 and the second combination signal 215 derived from the combiner 2031 (described later) in Fig. 3a, indicated by lines 214 and 215. A preferred optimization target is shown in Fig. 2, wherein the coding gain is maximized. is removed, i.e., the bit rate is reduced as much as possible. In this optimization objective, the residual signal D is reduced to Hanimum with respect to a. In other words, the predictive information a HS dMHZ is chosen to be minimized. This results in a solution for d shown in Figure 2. The signals S, M are given blockwise and preferably spectral domain signals; where "H..." is the 2-norm of the argument and <... denotes the inner product as usual. When the first channel signal 201 and the second channel signal 202 are input to the optimizer 207, the optimizer can apply the combination rule (an example combination rule is shown in Figure 3c). However, when the first combination signal 214 and the second combination signal 215 are input to the optimizer 207, the optimizer 207 need not automatically apply the combination rule. Other optimization goals may relate to perception quality. The optimization goal may be to achieve maximum perception quality. In this case, the optimizer may require additional information from a perceptual model. Other applications of the optimization goal may relate to achieving a minimum or constant bit rate. The optimizer 207 may then be applied to perform quantization/entropy-coding to determine the required bit rate for particular values of a, such that a is adjusted to meet requirements such as a minimum bit rate or, alternatively, a constant bit rate. Other applications of the optimization goal may relate to minimizing the use of encoder or decoder resources. If such an optimization objective is implemented, information regarding the resources required for a particular optimization may be available at optimizer 207. Additionally, a combination of these optimization objectives or other optimization objectives may be implemented to control optimizer 207, which computes the predictive information 206. While the encoder calculator in Figure 2 can be implemented in various ways, a first example implementation in which an explicit combination rule is implemented in combiner 2031 is shown in Figure 3a. An alternative example implementation using a matrix calculator 2039 is shown in Figure 3b. Combiner 2031 in Figure 3a can be implemented to implement the combination rule shown in Figure 3c, which is a well-known example of a mid/side encoding rule where a weighting factor of 0.5 is applied to all branches. However, depending on the application, other weighting factors may be applied, or it is possible to apply no weighting factor at all. It should also be noted that other combination rules, such as other linear combination rules or nonlinear combination rules, can be applied, as long as a corresponding inverse combination rule is available that can be applied in the decoder combiner 1162 shown in Figure 4a, which implements a combination rule that is opposite to the combination rule implemented by the encoder. Any reversible prediction can be used, as any existing error remains in the transmitted residual signal, since the prediction process performed by the optimizer 207 in combination with the effect on the waveform is a waveform-protecting process. The combiner 2031 outputs the first combination signal 204 and a second combination signal 2032. The first combination signal is input to a predictor 2033, while the second combination signal 2032 is input to the residual calculator 2034. The predictor 2033 calculates a prediction signal 2035, which is finally combined with the second combination signal 2032 to obtain a residual signal 205. The combiner 2031 is specifically configured to combine two channel signals 201 and 202 of the multi-channel audio signal in different ways to obtain the first combination signal 204 and the second combination signal 2032, the two different ways being shown in an example in Figure 3C. The predictor 2033 is configured to apply the prediction information to the first combination signal 204 or a signal derived from the first combination signal to obtain the prediction signal 2035. The signal derived from the combination signal can be derived by any linear or nonlinear operation; in this case, real-to-virtual/virtual-to-real transformation is preferred, which can be implemented using a linear filter such as a FIR (Finite Impulse Response) filter that performs weighted addition of certain values. The residual calculator 2034 in Figure 3a can also perform a subtraction operation such that the prediction signal is subtracted from the second combination signal. However, other operations can also be performed in the residual calculator. In parallel, the combination signal calculator 1161 in Fig. 4a may perform a summation operation in which the decoded residual signal 114 and the prediction signal 1163 are summed to obtain the second combination signal 1165. Fig. 5a shows an implementation of a vocoder. Compared to the vocoder shown in Fig. 3a, the first channel signal 201 is a spectral representation of a time domain first channel signal 55a. Correspondingly, the second channel signal 202 is a spectral representation of a time domain channel signal 55b. The conversion from the time domain to the spectral representation is performed by a time/frequency converter 50 for the first channel signal and another time/frequency signal 51 for the second channel signal. Spectral transformers (50, 51) are preferably implemented as real-valued transformers, but this is not mandatory. The transformation algorithm can be a discrete cosine transform, an FFT (Fast Fourier Transform) transform using only one real part, an MDCT transform, or any other transform that provides real-valued spectral values. Alternatively, both transforms can be implemented as a virtual transform, such as a DST (Discrete Sine Transform), MDST, FFT, or similar transform using only one imaginary part and discarding the real part. Any other transform that provides only imaginary values can also be implemented. One purpose of using a pure real-valued transform or a pure imaginary transform is to increase computational complexity, since for each spectral value, a single value, such as magnitude, must be processed, or alternatively, a phase or imaginary part. Unlike a purely complex transform such as an FFT, two values, one real part and one imaginary part for each spectral line, may need to be processed, with a computational complexity increase of at least a factor of 2. Another reason for using a real-valued transform here is that this type of transform is critically sampled, thus providing a convenient (and commonly used) domain for signal quantization and entropy coding (the standard "perceptual audio coding" paradigm implemented in "MP3," "AAC," or similar audio coding systems). Figure 5a also shows a residual calculator 2034 used as an adder that receives the side signal on its "plus" input and receives the prediction signal output by the predictor 2033 on its "minus" input. In addition, Figure 5a shows the case where the predictive control information is routed from the optimizer to the multiplexer 212, which outputs a multiplexed bit stream representing the encoded multi-channel audio signal. Specifically, the prediction process is performed such that the side signal is predicted from the center signal, as shown by the Equations on the right of Figure 5a. Preferably, the predictive control information 206 is a factor, as shown on the right of Figure 3b. In an example where the predictive control information includes a real part, such as a magnitude of a complex value a or a real part of a complex value a, and this part corresponds to a non-zero factor, a significant coding gain can be achieved when the center signal and the side signal are similar to each other due to their waveform structures but have different amplitudes. In the case where the predictive control information includes only a second part, which can be the imaginary part of a complex value factor or the phase information of a complex value factor, and where the imaginary part or the phase information is non-zero, the present invention achieves a significant coding gain for signals that are phase shifted to each other by a value other than 0° or 180° and have similar waveform characteristics and similar amplitude relationships, as well as phase shift. A predictive control information is preferably complex-valued. This allows significant coding gains to be achieved for signals that differ in amplitude and are phase-shifted. In a case where time/frequency transforms yield complex spectra, the process 2034 will be a complex process in which the real part of the predictive control information is applied to the real part of the complex spectrum M, and the imaginary part of the complex prediction information is applied to the imaginary part of the complex spectrum. Then, in the adder 2034, the results of this prediction process are a predicted real spectrum and a predicted imaginary spectrum. The predicted real spectrum is subtracted (bandwise) from the real spectrum of the side signal S, and the predicted imaginary spectrum is subtracted from the imaginary part of the S spectrum to obtain a complex residual spectrum D. Time domain signals (L and R) are real-valued signals, but frequency domain signals can be real-valued or complex-valued. Frequency domain signals are real-valued. If the frequency domain signals are complex, the transform is also a real-valued transform. This shows that the input time-to-frequency and output frequency-to-time transforms are real-valued, but frequency domain signals can be, for example, complex-valued QME' domain signals. Figure 5b shows an audio decoder corresponding to the audio coder shown in Figure 5a. Similar elements have similar reference numbers with respect to the audio decoder in Figure 1. The bitstream output by the bitstream multiplexer 212 in Figure 5a is input to a bitstream demultiplexer 102 in Figure 5b. The bitstream demultiplexer 102 multiplexes the bitstream into a downmix signal M and a residual signal D. The downmix signal M is input to dequantizer 110a. The residual signal D is input to another dequantizer 110b. In addition, the bitstream demultiplexer 102 multiplexes the predictive control information 108 from the bitstream and inputs it to predictor 1160. Predictor 1160 extracts a predicted side signal d' M, and combiner 1161 combines the predicted side signal and the residual signal output via dequantizer 110b to obtain the reconstructed side signal S in the final stage. The signal is then input to combiner 1162, which performs, for example, a sum/difference processing, as shown in Figure 4c, with respect to mid/side coding. Block 1162 specifically performs a (reverse) mid/side decoding to obtain a frequency domain representation of the left channel and a frequency domain representation of the right channel. The frequency domain representation is then converted to a time domain representation by means of corresponding frequency/time converters 52 and 53. Depending on the system implementation, the frequency/time converters 52, 53 are real-valued frequency/time converters if the frequency domain representation is a real-valued representation, or complex-valued frequency/time converters if the frequency domain representation is a complex-valued representation. However, for increased efficiency, it is preferred to perform a real-valued conversion, as in another embodiment shown for the encoder in Figure 6a in accordance with an embodiment of the present invention and for the decoder in Figure 6b in accordance with an embodiment of the present invention. The real-valued transforms 50 and 51 are implemented with an MDCT. Furthermore, the predictive information is calculated as a complex value having a real part and an imaginary part. Since both the M, 8 spectra are real-valued spectra and therefore the spectrum has no imaginary part, there is a real-to-virtual converter 2070 that calculates a virtual spectrum estimated from the real-valued spectrum of the M signal. This real-to-virtual converter 2070 is part of optimizer 207, and the virtual spectrum 600 estimated by block 2070 is input to optimizer stage 2071 along with the real M spectrum to calculate the predictive information 206, which now has a real-valued factor specified in 2073 and an imaginary factor specified in 2074. In this step, in accordance with the embodiment in question, the real-valued spectrum of the first M combination signal is multiplied by the real part (dR, 2073) to obtain the prediction signal, which is then subtracted from the real-valued side spectrum. Furthermore, the virtual spectrum 600 is multiplied by the virtual part (di) shown at 2074 to obtain the additional prediction signal; this prediction signal is then subtracted from the real-valued side spectrum shown at 2034b. Next, the D prediction residual signal is quantized in the quantizer 209b, and the real-valued M spectrum is quantized in the quantized/coded block 209a. Furthermore, it is preferable to quantize and encode the prediction information a in the quantizer/entropy encoder 2072 to obtain the encoded complex value a, which is routed to the bitstream multiplexer 212 in Figure 5 and input into the bitstream as prediction information in the final stage. Regarding the location of the quantization/coding (Q/C) module for a, it should be noted that the multipliers 2073 and 2074 preferably use the same (quantized) a as that which will be used in the decoder. Therefore, one could proceed directly to output 2071, or one could even consider that the quantization of a is already considered in the optimization process at 2071. Although a complex spectrum can be calculated at the encoder, since all the information is available, it is preferable for the decoder to perform the real-to-complex transformation at block 2070 to produce conditions similar to those shown in Figure 6b. The decoder receives a real-valued coded spectrum of the first combination signal and a real-valued spectral representation of the encoded residual signal. In addition, a complex prediction information encoded at 108 is obtained; entropy decoding and dequantization are performed to obtain the real part dR, shown at 1160b, and the imaginary part dI, shown at 1160c. The intermediate signals extracted by the weighting elements 1160b and 1160c are added to the decoded and dequantized prediction residual signal. Specifically, the spectral values input to weighter 1160c, using the imaginary part of the complex prediction factor as weighting factor, are derived from the real-valued spectrum M by means of the real-to-virtual converter 1160a, which is preferably implemented as in block 2070 for the encoder side of Figure 6a. On the decoder side, unlike the encoder side, there is no complex-valued representation of the mid-signal and side-signal. This is because, due to bit rate and complexity reasons, only the encoded real-valued spectra are transmitted from the encoder to the decoder. The real-to-virtual converter 1160a of Figure 6a may be implemented as published in or patent. Alternatively, any other implementation known from the relevant art may be implemented; A preferred embodiment in this regard is described with reference to Figures 10a, 10b. Specifically, as shown in Figure 10a, the virtual-to-virtual converter 1160a includes a spectral frame selector 1000 connected to a virtual spectrum calculator 1001. The spectral frame selector 1000 receives an indication of the current frame i at input 1002 and, depending on the implementation, control information at a control input 1003. For example, if the indication on line 1002 indicates that a virtual spectrum is to be calculated for the current frame i and the control information 1003 indicates that only the current frame is to be used for that calculation, the spectral frame selector 1000 selects only the current frame i and sends this information to the virtual spectrum calculator. The virtual spectrum calculator then uses only the spectral lines of the current frame i to perform a weighted combination of the lines located in the current frame (block 1008) that are close in frequency to or around the current spectral line k for which a virtual line is to be calculated, as shown at 1004 in Figure 10b. However, when the spectral frame selector 1000 receives control information indicating that the previous frame i-1 and the next frame i+1 are also to be used for computing the virtual spectrum, the virtual spectrum calculator additionally receives the values from frames i-l and i+l and performs a weighted combination in the corresponding frames, as shown at 1005 for frame i-l and at 1006 for frame i+1. The results of the weighting operations are combined with a weighted combination in block 1007 to obtain a virtual line k for frame f in the final stage; this is then multiplied by the imaginary part of the prediction information in element 1160c to obtain the prediction signal for that line, which is then added to the corresponding line of the middle signal in the adder 1161b for the decoder. The same operation is performed in the encoder, but a subtraction operation is performed in element 2034b. It should be noted that the control information 1003 may also specify using more than two surrounding frames, or using only the current frame and exactly one or more of the previous frames, for example, without using "ahead" frames to reduce systematic delay. It should also be noted that a stepwise weighted combination can also be implemented, as shown in Figure 10b, where, in a first step, the lines from a frame are combined, and then the results of these frame-by-frame combination operations are automatically combined. What is meant by "other order" here is that, in a first step, the valid k-frequency lines from a number of adjacent frames, specified by control information 103, are combined with a weighted combination. Depending on the number of adjacent lines to be used to estimate the virtual line, this weighted combination is applied for k, k-1, k-2, k+1, k+2, etc. Then, this step is held to obtain the virtual k line for frame fi. The weights are preferably set to values between -1 and 1, and a direct FIR or IIR (Infinite Impulse Response) filter can be applied in combination, performing linear combination of spectral lines or spectral signals from different frequencies and different frames. As indicated in Figures 6a and 6b, the preferred transform algorithm is the MDCT transform algorithm, which is applied forward in elements 0054 and 51 of Figure 6a, and backward in elements 52 and 53, following the combination operation in combiner 1162 operating in the spectral domain. Figure 8a shows a more detailed implementation of block 50 or 51. Specifically, a sequence of time domain audio samples is input to an analysis windower 500, which performs a windowing operation using an analysis window, specifically performing this operation frame by frame within a frame, but using a 50% overlap or overlap. The analysis windower results, i.e., sequences of frames of windowed samples, are input to an MDCT transform block 501, which extracts sequences of real-valued MDCT frames, where they are subjected to aliasing. For example, the analysis windower applies analysis windows of 2048 samples in length. Next, the MDCT transform block 501 extracts MDCT spectra with 1024 real spectral lines or MDCT values. Preferably, the analysis windower 500 and/or the MDCT converter 501 may be controlled by a window length or transform length control 502, such that the window length/transform length is reduced for transition portions in the signal to achieve better coding results. Figure 8 shows the inverse MDCT operation performed in blocks 0056 and 53. For example, block 52 includes block 250 for performing a frame-by-frame inverse MDCT transformation. For example, if one frame of MDCT values has 1024 values, the output of this MDCT inverse transform would contain 2048 aliased time samples. Such a frame is sent to a synthesis windower 521, which applies a synthesis window to that frame of 2048 samples. The windowed frame is then routed to an overlap/summing processor 522, which applies, for example, a 50% overlap between the two subsequent frames, and then performs sample-by-sample addition of the 2048-sample block, resulting in a final sample of 1024 new non-aliased output signal samples. Again, as noted in 523, it is preferable to implement window/transform length control, for example, using information conveyed in the side information of the encoded multi-channel signal. Prediction values can be calculated for each individual spectral line in an MDCT spectrum. However, it has been found that this is not necessary and that a significant amount of side information can be recovered by computing the prediction information band-wise. In other words, a spectral converter 50, which is an MDCT processor as described in Figure 8a, provides a high-frequency resolution spectrum having certain spectral lines as shown in Figure 9b. This high-frequency resolution spectrum is used by a spectral line selector 90, which provides a spectrum with resolutions B1, B2, B3. This low-frequency resolution spectrum is routed to the optimizer 207 to calculate the predictive information, such that the predictive information is calculated only for each band, not for each spectral line. To this end, the optimizer 207 takes the spectral lines per band and calculates the optimization process based on the assumption that the same d value is used for each spectral line in the band. Preferably, the bands are psychoacoustically shaped so that their bandwidth increases from low to high frequencies, as shown in Figure 9b. Alternatively, although not as desirable as bandwidth expansion, it is also possible to use equally sized frequency bands, where each frequency band has at least two or typically many more frequency lines, for example, at least 30. Generally, for a spectrum with 1024 spectral lines, fewer than 30 complex a values are calculated, and preferably more than 5. For spectra with fewer than 1024 spectral lines (e.g., 128 lines), fewer frequency bands (e.g., 6) are preferably used for a. MDCT spectra are not required for calculating a values. Alternatively, a filter bank with a frequency resolution similar to the resolution required to calculate the a values can be used. When increasing frequency bands are to be applied, this filter bank must have a variable bandwidth. However, if a constant bandwidth from low to high frequencies is sufficient, a conventional filter bank with equal-width subbands can be used. Depending on the application, the signature of the a value indicated in Figure 3b or 4b can be inverted. However, for consistency, this inversion should be used both at the encoder and the decoder. Figure 5a shows a general view of the encoder, compared to Figure 6a; where item 2033 is a predictor controlled by means of predictive control information 206, specified in item 207, which is incorporated into the bit stream as side information. In Figure 5a, a generalized time/frequency transform is used in place of the MDCT used in blocks 50, 51 in Figure 6. As previously mentioned, Figure 6a is an encoder process corresponding to the decoder process in Figure 6b, where L represents the left channel signal; R represents the right channel signal; M represents the middle signal or downmix signal; S represents the side signal; and D represents the residual signal. Alternatively, L may be referred to as the first channel signal 201; R represents the second channel signal 202; M represents the first combination signal 204; and δ may be referred to as the second combination signal 2032. Preferably, modules 2070 in the encoder and 1160a in the decoder should be precisely matched to ensure correct waveform encoding. This also applies preferably to the case where these modules use some approach such as cut-off filters, or where the output consists of only one or two MDCT frames (i.e., the current MDCT frame on line 60), the previous MDCT frame on line 61, and the next MDCT frame on line 62, rather than three MDCT frames. In addition, it is preferred that module 2070 in the encoder of Figure 6a use the non-quantized MDCT spectrum M as input, while module 1160a in the decoder has only the quantized MDCT spectrum as input. Alternatively, it is possible to use an implementation in which the encoder uses quantized MDCT coefficients as input to module 2070. However, from a perceptual perspective, using the unquantized MDCT spectrum as input to module 2070 is a preferred approach. Some aspects of the embodiments of the present invention will be described in more detail later. Standard parametric stereo coding relies on the ability of the oversampled complex (mixed) QMF domain to allow perceptually oriented signal processing that varies in time and frequency without introducing aliasing artifacts. However, in the case of downmix/backmix coding (as used at the high bit rates considered here), the resulting combined stereo encoder functions as a waveform encoder. This allows processing in a critically sampled region, such as the MDCT region, because the waveform coding paradigm (MDCT - modified discrete cosine transform) ensures that the aliasing removal property of the processing chain is sufficiently well preserved. However, to benefit from the coding efficiency achievable with stereo signals with inter-channel time or phase differences via a complex-valued prediction coefficient d, a complex-valued frequency domain representation of the downmix signal (DMX) is required as input to the complex-valued upmix matrix. This can be achieved by using an MDST transform in addition to the MDCT transform for the DMX signal. The MDST spectrum can be calculated (exactly or approximately) from the MDCT spectrum. Furthermore, the parameterization of the upmix matrix can be simplified by passing the complex prediction coefficient d instead of the MPS (MPEG Surround) parameters. Therefore, instead of three parameters (ICC (Inter-Channel Coherence), CLD (Channel Level Difference), and IPD (Inter-Channel Phase Difference)), only two parameters (the real and imaginary parts of d) are transmitted. This is possible due to the redundancy of the MPS parameterization in the case of downmix/residual coding. The MPS parameterization includes information on the relative decorrelation amount to be added to the decoder (i.e., the energy ratio between the RES (Residual) and DMX (Downmix) signals), and this information is more than necessary when actual DMX and RES signals are transmitted. For the same reason, the gain coefficient (g), shown in the upmix matrix above, is no longer used in the case of downmix/residual coding. Therefore, for complex predictive downmix/residual coding, the upmix matrix is now given by rL _ 1 ~ a 1 [DMX] Compared to Equation 1169 in Figure 4b, in this equation the alpha sign is inverted; DMX = M and RES = DT. Therefore, this implementation is an alternative implementation/writing based on Figure 4b. There are two options for calculating the predictive residual signal in the encoder. One is to use the quantized MDCT spectral values of the downmix. Since the encoder and decoder use the same values to generate the prediction, this option results in the same quantization error distribution as in M/S coding. The other option is to use the non-quantized MDCT spectral values. This means that the encoder and decoder do not use the same data to generate the prediction, which allows the coding error to be spatially redistributed according to the instantaneous masking properties of the signal, at the expense of a somewhat reduced coding gain. It is preferable to directly calculate the MDST spectrum in the frequency domain using a two-dimensional FIR filter for three adjacent MDCT frames described. The latter can be thought of as a "real-to-virtual" (R2I) transformation. The complexity of the frequency domain calculation of the MDST can be reduced in several ways, i.e., only an approximation of the MDST spectrum can be calculated: - Limiting the number of FIR filter taps - Estimating the MDST from only the current MDCT frame. Waveform coding properties are not affected as long as the same approximation is used in the encoder and decoder. However, these approximations to the MDST spectrum may result in a reduction in the coding gain achieved with complex prediction. If the underlying MDCT encoder supports window-shaped transitions, the coefficients of the two-dimensional FIR filter used to calculate the MDST spectrum must be adapted to the actual window shapes. The filter coefficients applied to the MDCT spectrum of the current frame depend on the full window, meaning a set of coefficients is required for each window type and each window transition. The filter coefficients applied to the MDCT spectrum of the previous/next frame depend only on the window that overlaps half of the current frame, meaning they only require a set of coefficients for each window type (no additional coefficients for each transition). If the underlying MDCT encoder uses transform-length transitions, a more complex situation arises for transitions between different transform lengths, including the previous and/or next MDCT frame in the approximation. Two-dimensional filtering is more complex in this case due to the different numbers of MDCT coefficients in the current and previous/next frames. To address the increased numerical and structural complexity, the previous/next frame can be omitted from the filtering process during transform length transitions, at the expense of decreasing approximation accuracy for specific frames. Furthermore, special attention should be paid to the lowest and highest parts of the MDST spectrum (near DC and fs/2), where the surrounding MDCT coefficients are less than those required for FIR filtering. Here, the filtering process must be adapted to accurately calculate the MDST spectrum. This can be achieved by symmetrically stretching/expanding the MDCT spectrum for omitted coefficients (according to the periodicity of the spectra of time-discontinuous signals) or by appropriately adapting the filter coefficients. These special cases can be handled in a simplified manner, at the expense of reduced accuracy near the boundaries of the MDST spectrum. Computing the full MDST spectrum from the transmitted MDCT spectra at the decoder reduces decoder latency by one frame (assuming 1024 samples here). Additional latency can be avoided by using an approximation of the MDST spectrum that does not require the MDCT spectrum of the next frame as input. The advantages of MDCT-based combined stereo coding over QMF-based combined stereo coding are listed below: - Only a small increase in computational complexity (when SBR is not used). - Scales up to perfect reconstruction when the MDCT spectra are not quantized, which is not the case with QMF-based combined stereo coding. - Natural expansion of M/S coding and loudness stereo coding. A cleaner architecture that simplifies encoder matching, as stereo signal processing and quantization/coding can be tightly coupled. It should be noted here that in QMF-based combined stereo coding, MPEG Surround frames and MDCT frames are not aligned, and the scale factor bands do not map to the parameter bands. - Efficient encoding of stereo parameters, as only two parameters (complex d) need to be transmitted, instead of three as in MPEG Surround (ICC, CLD, IPD). If the MDST spectrum is calculated as an approximation (without using the next frame), no additional decoder delay is introduced. The important features of an implementation can be summarized as follows: a) MDST spectra are calculated from the current, previous, and next MDCT spectra through a two-dimensional FIR filter. Various complexity/quality trade-offs are possible for the MDST calculation (approximation) by reducing the number of FIR filter taps and/or the number of MDCT frames used. If a contiguous frame is not available due to a frame loss during the transmission or transform length traversal, that frame is excluded from the MDST estimation (exclusion). In the transform length traversal, an exclusion signal is signaled in the bitstream. Instead of MICC, CLD, and IPD, only two parameters (i.e., the real and imaginary parts of the complex prediction coefficient d) are transmitted. The real and imaginary parts of A are processed independently, with a limited range. If a particular parameter (real or imaginary part of d) is not used in a given frame, its signal is signaled in the bitstream, and the irrelevant parameter is not transmitted. The parameters are encoded either time- or frequency-differentially, and finally, Huffman coding is applied using the scale factor codebook. The prediction coefficients are updated in the scale factor band every second, resulting in a frequency resolution similar to that of MPEG Environment. This quantization and coding scheme results in an average bit rate of approximately 2 kb/s for stereo side information in a typical configuration with a target bit rate of 96 kb/s. Preferred additional or alternative implementation details include the following: non-differential (PCM) or differential (DPCM) coding, which operates on a per-frame or per-stream basis, signaled by a corresponding bit; time or frequency differential coding is possible for DPCM coding. This can also be signaled using a one-bit signal. d) instead of reusing a predefined codebook such as the AAC (Advanced Audio Coding) scale factor book, a dedicated invariant or signal-adaptive codebook can be used to encode the d parameter values, or a fixed-length (e.g. 4-bit) unsigned or two's complement codeword can be switched. e) the range of d parameter values, as well as the parameter quantization step size, can be arbitrarily selected and optimized for the signal properties at hand. üThe number and spectral and/or temporal width of the active parameter bands can be arbitrarily selected and optimized for specific signal characteristics. In particular, band structuring can be done on a per-frame or per-stream basis. g) In addition to or as an alternative to the mechanisms described in a) above, it can be explicitly signaled that only the MDCT spectrum of the current frame is used to calculate the MDST spectrum approximation via bits per frame in the bitstream, i.e., adjacent MDCT frames are not taken into account. The embodiments relate to the inventive system for combined stereo coding in the MDCT domain. This allows to utilize the advantages of combined stereo coding in the MPEG USAC system even at higher bit rates without a significant increase in the computational complexity introduced by the QMF-based approach (SBR is not used here). The following two lists describe the preferred construction aspects described previously, which can be used alternatively or in addition to other aspects: 1a) general concept: complex prediction of side MDCT from middle MDCT and MDST; 1b) calculate/approximate MDST from MDCT in the frequency domain using 1 or more frames (leading to a 3-frame delay); 1c) filter truncation to reduce computational complexity (with the l-frame reduced to 2-tap, i.e. [-101]); ld) proper handling of DC and fs/2; le) proper handling of window shape transition; lf) use previous/next frame if it has a different transform size; lg) prediction based on unquantized or quantized MDCT coefficients in the encoder; Za) directly quantize and encode the real and imaginary part of the complex prediction coefficient (i.e. no MPEG Environment parameterization); 2b) use uniform quantizer for this (e.g. step 2c) use appropriate frequency resolution for the prediction coefficients (e.g. 1 coefficient for 2 Scale Factor Bands): 2d) cheap signaling if all prediction coefficients are real; 2e) clear bit per frame to start the l-frame RZI process. In one embodiment, the encoder further comprises a spectral converter (50, 51) for converting the time domain representation of the two channel signals into a spectral representation of the two channel signals having subband signals for said two channels, wherein the combiner (2031), predictor (2033) and residual signal calculator (2034) are configured to process each subband signal separately to obtain the first combined signal and the residual signal for multiple subbands; wherein the output interface (212) is configured to combine the first combined signal encoded for multiple subbands and the encoded residual signal. While some aspects are defined in the context of a device, it is clear that these aspects represent the definition of the corresponding method, where a block or device corresponds to a method step or a feature of the method step. Similarly, aspects described in the context of a method step also include the description of a corresponding block or element or a feature of the corresponding device. In one example, a smooth processing of window shape transition is implemented. Referring to Figure 10a, window shape information 109 can be entered into the virtual spectrum calculator 1001. Specifically, the virtual spectrum calculator, which performs the real-to-virtual transformation of the MDCT spectrum (Figure 1001), can be implemented as a FIR or IIR filter. The FIR or IIR coefficients in this real-to-virtual module depend on the window shape of the left half and the right half of the current frame. This window shape can be different for a sine window or a Kaiser Bessel Derived (KBD) window and can be a long window, a start window, a stop window, a stop-start window or a short window, depending on the respective window sequence configuration. The real-to-virtual module can include a two-dimensional FIR filter; where the first dimension is a time dimension in which two consecutive MDCT frames are input into the FIR filter, and the second dimension is a frequency dimension in which the frequency coefficients of a frame are input. The table below shows the different window shapes and the left half of the window. Different MDST filter coefficients for a valid window sequence are given for different implementations of the right and left halves. Table A - MDST Filter Parameters for the Current Window Current Window Left Half: Sine Shape RightLeft Half: KBD Shape Right Sequence Half: Sine Shape Half: KBD Shape Current Window Left Half: Sine Shape RightLeft Half: KBD Shape Right Sequence Half: KBD Shape Half: Sine Shape (continued) Left Half: Sine Shape RightLeft Half: KBD Shape Right Current Window Half: KBD Shape Half: Sine Shape In addition, when the previous window is used to calculate the MDST spectrum from the MDCT spectrum, the information (109) for the previous window is used to calculate the MDST spectrum from the corresponding MDST Filter for the previous window. The coefficients are given in the table below. Table B - MDST Filter Parameters for the Previous Window Current Window Sequence Current WindowLeft Half of the Current Window: SineLeft Half: KBD Consequently, depending on the window shape information (109), the virtual spectrum calculator (1001) in Figure 10a can be adapted by applying different groups of filter coefficients. The window shape information used on the decoder side is calculated on the encoder side and transmitted as side information with the encoder output signal. On the decoder side, the window shape information (109) is extracted from the bit stream by means of a bitstream demultiplexer (e.g., 102 in Figure 5b) and the virtual spectrum is extracted from the bit stream by means of a bitstream demultiplexer (e.g., 102 in Figure 5b) as shown in Figure 10a. The window shape information (109) signals that the previous frame has a different transform size. If it does, it is preferred not to use the previous frame to calculate the virtual spectrum from the real-valued spectrum. The same applies when the window shape information 109 is interpreted and it is found that the next frame has a different transform size. In this case, the next frame is not used to calculate the virtual spectrum from the real-valued spectrum. In such a case, for example, when the previous frame has a different transform size than the current frame or when the next frame has a different transform size than the current frame, only the spectral values of the current frame, i.e., the current window, are used to estimate the virtual spectrum. The prediction in the encoder is based on unquantized or quantized frequency coefficients, such as the MDCT coefficients. When the prediction, shown by element 2033 in Figure 3a, is based on unquantized data, for example, the calculator 2034 preferentially operates on the unquantized data as well, and the calculator output signal is, i.e., the residual signal 205 is quantized before entropy coding and transmitted to a decoder. In an alternative embodiment, it is preferred that the prediction be based on quantized MDCT coefficients. In this case, quantization can occur before the combiner 2031 in Figure 3a, so that a first quantized channel and a second quantized channel are the basis for calculating the residual signal. Alternatively, the quantization process can occur after the combiner 2031, so that the first combination signal and the second combination signal are calculated in a non-quantized form and quantized before calculating the residual signal. Alternatively, the predictor 2033 can operate in the non-quantized region*, and the prediction signal 2035 is quantized before being input to the residual calculator. In this case, the second combination signal 2032, another signal input to the residual calculator 2034, is also input to the residual calculator. It is beneficial to quantize the residual signal 1070 in Figure 6a before calculating it, which can be implemented within the predictor 2033 in Figure 3a and operates with the same quantized data available at the decoder. This ensures that the MDST spectrum estimated at the encoder for the purpose of performing the calculation of the residual signal is exactly the same as the MDST spectrum at the decoder side used to perform the inverse prediction, i.e., to calculate the side signal from the residual signal. For this purpose, the first combination signal, such as the M signal on line 204 in Figure 6a, is quantized before being input to block 2070. Then, the MDST spectrum calculated using the quantized MDCT spectrum of the current frame and, depending on the control information, the quantized MDCT spectrum of the previous or next frame are input to the multiplier 2074; in this case, the output of the multiplier 2074 in Figure a is also will be an unquantized spectrum. This unquantized spectrum will be extracted from the spectrum input to the adder 2034b and finally quantized in the quantizer 209b. In one embodiment, the real part and the imaginary part of the complex prediction coefficient per prediction band are quantized and encoded directly, i.e., without, for example, MPEG Environment parameterization. The quantization process can be performed using a uniform quantizer with, for example, a 0.1 step size. This means that no logarithmic quantization step size and the like are applied, but any linear step size is applied. In one embodiment, the value range of the real part and the imaginary part of the complex prediction coefficient varies from -3 to 3, which means that 60 or 61 quantization steps for the real part and the imaginary part of the complex prediction coefficient, depending on the implementation details. Preferably, the real part applied in the multiplier 2073 in Figure 6a and the imaginary part 2074 applied in Figure 6a are also quantized before being applied, so that the same prediction value is used at the encoder side as at the decoder side. This ensures that the prediction residual signal covers not only the introduced quantization error but also all errors that may occur when a non-quantized prediction coefficient is applied at the encoder side while a quantized prediction coefficient is applied at the decoder side. Preferably, the quantization process is applied in such a way that the same state and the same signals are present at both the encoder side and the decoder side as much as possible. Therefore, it is preferred that the quantization of the input from the real to the virtual calculator 2070 is performed using the same quantization as that applied at the quantizer 209a. Also, To perform the multiplications in items 2073 and 2074, it is preferable to quantize the real part and the imaginary part of the prediction coefficient. This quantization operation is the same as that implemented in quantizer 2072. In addition, the side signal extracted by block 2031 in Figure 6a can also be quantized before the adders 2034a and 2034b. However, it is not problematic to perform the quantization performed by quantizer 209b after the addition operation, where the addition performed by the adders in question is performed with an unquantized side signal. In another example, cheap signaling is applied if all prediction coefficients are real. There may be a situation where all prediction coefficients are calculated as real for a particular frame, i.e., for the same time portion of the audio signal. Such a situation can be the case for the full mid-signal or full This can happen when there is no or very little phase shift from one side signal to the other. To reduce the amount of bits used, this is indicated by a single real indicator. In this case, the imaginary part of the prediction coefficient does not need to be signaled in the bitstream with a codeword representing a zero value. On the decoder side, the bitstream decoder interface, for example, the bitstream demultiplexer, will interpret this new indicator and will not search for codewords for the imaginary part, but instead will assume that all bits are in the corresponding part of the bitstream for the real-valued prediction coefficients. Furthermore, when the predictor 2033 receives an indication that all imaginary parts of the prediction coefficients in the frame are zero, it will not need to calculate an MDST spectrum, or an imaginary spectrum in general, from the real-valued MDCT spectrum. Therefore, element 1160a in the decoder in Figure 6b will be deactivated. and inverse prediction will be performed using only the real-valued prediction coefficient implemented in multiplier 1160b in Figure 6b. The same applies to the encoder side, where element 2070 will be deactivated and prediction will be performed using only multiplier 2073. This side information is preferably used as an additional bit per frame, in which case the decoder will read this bit frame by frame to decide whether the real-to-virtual converter 1160a will be active for a frame. Therefore, providing this information results in a reduction of the bit stream size due to more efficient signaling of all virtual parts of the prediction coefficient that are zero for a frame, and additionally, it provides less complexity in the decoder for such a frame, which will quickly lead to a reduction in the battery consumption of such a processor implemented in, for example, a battery-powered mobile device. The complex stereo prediction according to preferred embodiments of the present invention is a tool for efficient coding of channel pairs with level and/or phase differences between channels. Using the complex valued parameter a, the left and right channels are reconstructed with the following matrix. Here, dmxm" denotes the MDST corresponding to the MDCT of the downmix channels (dmxmg). l' 1+aRe alin _i i The above equation is another representation of an equation for a prediction/combination process separated and combined in terms of the real part and the imaginary part of a, where the calculation of the predicted signal S is not required. The following data items are preferably used for this tool: cplx_pred_all 0: Some bands use L/R coding, the signal for which is given by cplx_pred_used[] 1: All bands use complex stereo prediction (continued) cplat_pred;used[g][sfb] (after matching from prediction bands) one-bit sign per window group (g) and scale factor band (sfb) indicates O: complex prediction is not used, L/R coding is used 1: complex prediction is used complex_coef 0: for all prediction bands QH1= O 1: for all prediction bands ahi is transmitted use_prev_frame O: use only current frame for MDST estimation 1: use current and previous frame for MDST estimation delta_code_time O: Frequency differential coding of prediction coefficients l: Time differential coding of prediction coefficients hcod;alpha_g_re d Re Huffman code hcod;alpha_g_im d Im Huffman code These data items are calculated in an encoder for a multichannel audio signal The elements are extracted from the side information at the decoder side using a side information extractor and are used to control the decoder calculator to perform a corresponding operation. Complex stereo prediction requires the downmix MDCT spectrum of the current channel pair; in the case complex_coef == 1, the downmix MDST spectrum of the current channel pair, i.e., the virtual counterpart of the MDCT spectrum. The downmix MDST estimation is calculated from the MDCT downmix of the current frame; in the case use_prev_frame == ]., the MDCT downmix of the previous frame belonging to window group (g) and group window (b) is obtained from the reconstructed left and right spectra of that frame. In the calculation of the downmix MDST estimation, the double-valued MDCT transform based on window_sequence, as well as filter_coefs and filter_coefs_prev is used. is used, which are arrays containing the filter kernels and derived according to the previous tables. For all prediction coefficients (in time or frequency), the previous value is encoded using a Huffman codebook. When cplx_pred_used = 0, the prediction coefficients are not transmitted for the prediction bands. The prediction coefficients are not transmitted for the prediction bands for which cplx_pred_used = 0. The inverse quantized prediction coefficients alpha_re and alpha_im are given by: alpha_re = alpha___q_re*0.1 alpha_im = alpha_q`im*0.1 It should be noted that the present invention is applicable not only to stereo signals, i.e. multi-channel signals with only two channels, but also to two channels of multi-channel signals with three or more channels, such as 5.1 or 7.1 signals. Depending on particular application requirements, embodiments of the invention may be implemented in hardware or software. The implementation may be implemented in a digital storage medium such as a floppy disk, DVD, CD, ROM, PROM, EPROM, EEPROM, or FLASH memory, which stores electronically readable control signals that are interoperable (or capable of interoperating) with a programmable computer system to enable the method to be performed. Some embodiments of the invention include a carrier carrying permanent or tangible data with electronically readable control signals that are interoperable with a programmable computer system to enable one of the methods described herein. In general, embodiments of the present invention may be implemented as a computer program product with program code that executes to perform one of the methods when the computer program product is running on a computer. The program code may be stored, for example, on a machine-readable carrier. Other embodiments have the computer program for performing one of the methods described herein stored on the machine-readable carrier. In other words, one embodiment of the inventive step is therefore a computer program having program code for performing one of the methods described herein when the computer program is running on a computer. Another embodiment of the inventive methods is therefore a data carrier (or a digital storage medium, or a computer-readable medium) having the computer program recorded thereon for performing one of the methods described herein. A further embodiment of the inventive method is therefore a data stream or a sequence of signals representing the computer program for performing one of the methods described herein. The data stream or sequence of signals may be configured to be transmitted via a data communications connection, such as the Internet, for example. Another embodiment includes a processing means, such as a computer or a programmable logic device, configured or adapted to perform one of the methods described herein. Another embodiment includes a computer loaded with a computer program to perform one of the methods described herein. In some embodiments, a programmable logic device (e.g., a field programmable gate array) may be used to perform some or all of the functionality of the methods described herein. In some embodiments, the field programmable gate array may cooperate with a microprocessor to perform one of the methods described herein. The methods may generally be implemented by any hardware device, preferably any hardware device. The embodiments described above are merely illustrative of the principles of the present invention. It will be apparent to those skilled in the art that various modifications and variations can be made to the embodiments and details described herein. Therefore, the intention is to be limited only by the scope of the attached patent claims and not by the specific details presented by way of description and explanation of the embodiments herein.TR TR TR TR TR

Claims (1)

1.