Codec for a time-varying signal using coefficient-wise prediction

Abstract

A decoder, encoder and method for decoding and encoding are presented. The decoder is for block-wise decoding a digital time-varying signal from a data stream, configured to decode a signaled residual transform for a currently decoded block from the data stream; sequentially reconstruct coefficients of a residual transform of the currently decoded block by predicting a currently reconstructed coefficient of the residual transform from one or more previously reconstructed coefficients of the residual transform based on filter coefficients to obtain a predictor for the currently reconstructed coefficient, and correcting the predictor using a signaled coefficient value of the signaled residual transform, corresponding to the currently reconstructed coefficient, reconstruct the currently decoded block by correcting a sample predictor for the currently decoded block based on the residual transform, wherein the decoder is configured to, in sequentially reconstructing coefficients of the residual transform of the currently decoded block, adapt the filter coefficients based on 1) a) the predictor for the currently reconstructed coefficient, and the signaled coefficient value of the signaled residual transform, corresponding to the currently reconstructed coefficient, or b) the currently reconstructed coefficient, and 2) the one or more previously reconstructed coefficients of the residual transform so as to be used for a subsequently reconstructed coefficient of the residual transform.

Description

[0001] Codec for a time-varying signal using coefficient-wise prediction

[0002] Description

[0003] Embodiment according to the invention relate to an encoder for encoding a time-varying signal.

[0004] 1. Introduction

[0005] Time-varying signals are commonly used for representation of media and measurement data such as audio signals, biomedical signals, or seismic measurements. With the increase of signal generation, transportation and storage, there is demand for compression of such time-varying signals. There is a need to improve coding efficiency.

[0006] This is achieved by the subject matter of the independent claims of the present application.

[0007] Further embodiments according to the invention are defined by the subject matter of the dependent claims of the present application.

[0008] of the invention

[0009] In accordance with a first aspect, a decoder for block-wise decoding a digital time-varying signal is provided. The decoder is configured to decode a signaled residual transform for a currently decoded block from the data stream. The decoder is configured to sequentially reconstruct coefficients of a residual transform of the currently decoded block by predicting a currently reconstructed coefficient of the residual transform from one or more previously reconstructed coefficients of the residual transform based on filter coefficients to obtain a predictor for the currently reconstructed coefficient, and correcting the predictor using a signaled coefficient value of the signaled residual transform, corresponding to the currently reconstructed coefficient. The decoder is configured to reconstruct the currently decoded block by correcting a sample predictor for the currently decoded block based on the residual transform. The decoder is configured to, in sequentially reconstructing coefficients of the residual transform of the currently decoded block, adapt the filter coefficients based on 1) a) the predictor for the currently reconstructed coefficient, and the signaled coefficient value of the signaled residual transform, corresponding to the currently reconstructed coefficient, or b) the currently reconstructed coefficient, and 2) the one or more previously reconstructed

[0010] FH250307PEP-2026097809fecoefficients of the residual transform so as to be used for a subsequently reconstructed coefficient of the residual transform.

[0011] The decoder uses, in essence, two prediction stages. At a first stage, the residual transform is predicted by predicting coefficients based on already reconstructed coefficients. At a second stage, the decoder corrects a sample predictor using the residual transform. Therefore, correlations and redundancies can be exploited at two prediction stages in order to improve coding efficiency. The filter coefficients allow controlling the prediction of the coefficients. It has been recognized that the prediction can be further improved by adapting the filter coefficients for subsequently reconstructed coefficients of the residual transform. The predictor for the currently reconstructed coefficient is the result of the filter coefficients before the next update and is therefore indicative of how well the coefficients have performed so far. The signaled coefficient value of the signaled residual transform is used for correcting the predictor and therefore indicates how far away the prediction is from the reconstructed coefficient. Consequently, these parameters can both indicate how well the current filter coefficients have performed and how they may need to be adapted. The previously reconstructed coefficients of the residual transform can provide information on how the filter coefficients need to overall adjusted in order to better predict the coefficients. The decoder can use multiple parameters as input for adjusting (e.g., training) the filter coefficients, which can therefore improve overall coefficient prediction.

[0012] In accordance with a second aspect, a decoder is provided for block-wise decoding a digital time-varying signal comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, from a data stream. The decoder configured to reconstruct each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block of a first channel by decoding the signaled residual transform for the currently decoded block from the data stream, sequentially reconstructing coefficients of the residual transform of the currently decoded block by predicting a currently reconstructed coefficient of the residual transform from one or more previously reconstructed coefficients of the residual transform to obtain a predictor for the currently reconstructed coefficient, and correcting the predictor using a signaled coefficient value of the signaled residual transform, corresponding to the currently reconstructed coefficient, reconstructing the currently decoded block by correcting a sample predictor of the currently decoded block

[0013] FH250307PEP-2026097809febased on the residual transform. The decoder is configured to, in sequentially reconstructing coefficients of the residual transform of the currently decoded block, depending on whether the selected prediction modes of one or more blocks of one or more reference channels, which are temporally aligned to the currently decoded block, fulfill a predetermined relationship with the selected prediction mode of the currently decoded block, predict the currently reconstructed coefficient of the residual transform additionally from a corresponding coefficient of the residual transform of the one or more blocks of the one or more reference channels.

[0014] As discussed above for the first aspect, the decoder uses two prediction stages in order to better exploit signal redundancies. It has been recognized that in the case of multiple channels, coefficient prediction may further be improved by using information from other channels. A particularly strong correlation may be found for a corresponding coefficient of the residual transform of the one or more blocks of the one or more reference channels. By using a coefficient of the residual transform, the decoder is not required to perform transformation of the entire one or more blocks themselves, which could increase coding complexity. It has been recognized that the residual transform itself can also comprise correlations (e.g., without a sample predictor for constructing the block of the other channel). The decoder in able to determine whether or not the corresponding coefficient is to be used based on whether the selected prediction modes of one or more blocks of one or more reference channels fulfill a predetermined relationship with the selected prediction mode of the currently decoded block. It has been recognized that an indicator for a potential correlation are the prediction modes of the blocks. Therefore, the decoder is able to determine whether to use other blocks based on the prediction modes without requiring an explicit signalling in the data stream, further improving coding efficiency. Such predetermined relationships may, for example, relate to a type of prediction mode (e.g., a cross-channel prediction mode) and / or prediction parameters (e.g., temporal offset).

[0015] In accordance with a third aspect, a decoder for block-wise decoding a digital time-varying signal from a data stream is provided. The digital time-varying signal comprises multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned. The decoder is configured to reconstruct each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block of a first channel depending

[0016] FH250307PEP-2026097809feon whether the selected prediction modes of one or more blocks of one or more reference channels, which are temporally aligned to the currently decoded block, fulfill a predetermined relationship with the selected prediction mode of the currently decoded block, by decoding the signaled residual transform for the currently decoded block from the data stream; sequentially reconstructing coefficients of the residual transform of the currently decoded block by predicting a currently reconstructed coefficient of the residual transform from one or more previously reconstructed coefficients of the residual transform and a corresponding coefficient of the residual transform of the one or more blocks of the one or more reference channels to obtain a predictor for the currently reconstructed coefficient, and correcting the predictor using a signaled coefficient value of the signaled residual transform, corresponding to the currently reconstructed coefficient, reconstructing the currently decoded block by correcting a sample predictor of the currently decoded block based on the residual transform.

[0017] The third aspect is similar to the second aspect, e.g., in regards to a two-stage prediction process and checking whether the selected prediction modes fulfill a predetermined relationship. However, the decoder of the third aspect is more flexible in whether or not perform coefficient prediction of the currently reconstructed coefficient based on the one or more previously reconstructed coefficients of the residual transform. In case the prediction modes fulfill the predetermined relationship, the decoder performs the coefficient prediction based on both, previously reconstructed coefficients of the residual transform and the corresponding coefficient. However, in case the prediction modes do not fulfill the predetermined relationship, the decoder is still free to perform or not perform prediction based on previously reconstructed coefficients. Therefore, coding flexibility is improved. For example, the decoder may simply not prediction based on previously reconstructed coefficients at all (e.g., in order to reduce coding complexity), may perform prediction based on implicit signalling, or may perform prediction based on signalling (e.g., which may be on a block-by-block bases, e.g., by flag, or may be signalled less frequently, e.g., at each random-access point or once at the start, therefore reducing signaling overhead).

[0018] According to a further aspect is provided an encoder configured to encode the data stream decodeable by the decoders of the first, second, and third aspect.

[0019] According to a further aspect is provided a method for decoding or encoding as described herein.

[0020] FH250307PEP-2026097809feAccording to a further aspect is provided a data stream encoded according to any encoding method (or by any encoder) described herein.

[0021] According to a further aspect is provided a computer program product for implementing any method disclosed herein when being executed on a computer or signal processor.

[0022] Brief Description of the Drawings

[0023] The drawings are not necessarily to scale, emphasis instead generally being placed upon illustrating the principles of the invention. In the following description, various embodiments of the invention are described with reference to the following drawings, in which:

[0024] Fig. 1 shows a schematic view of a decoder for block-wise decoding a digital timevarying signal from a data stream;

[0025] Fig. 2 shows a schematic view of another example of a decoder;

[0026] Fig. 3 shows a schematic view of an example of a block and the currently decoded block with respective temporal offsets;

[0027] Fig. 4 shows a flow diagram of a method for block-wise decoding a digital timevarying signal from a data stream;

[0028] Fig. 5 shows a schematic view of an encoder for block-wise encoding a digital timevarying signal into a data stream;

[0029] Fig. 6 shows a flow diagram of a method for block-wise encoding a digital timevarying signal into a data stream;

[0030] Fig. 7 shows a schematic view of a decoder for block-wise decoding a digital timevarying signal from a data stream;

[0031] Fig. 8 shows a flow diagram of a method for decoding a digital time-varying signal from a data stream;

[0032] FH250307PEP-2026097809feFig. 9 shows a schematic view of an encoder for block-wise encoding a digital timevarying signal into a data stream;

[0033] Fig. 10 shows a flow diagram of a method for block-wise encoding a digital timevarying signal into a data stream;

[0034] Fig. 11 shows a flow diagram of a method for decoding a digital time-varying signal from a data stream;

[0035] Fig. 12 shows a flow diagram of a method for block-wise encoding a digital timevarying signal into a data stream;

[0036] Fig. 13a shows a schematic view of a full channel block with a plurality of channel;

[0037] Fig. 13b shows a schematic view of a coefficient prediction using corresponding coefficients of other channels;

[0038] Fig. 14 shows a schematic view of an example of channels, which do not have the same temporal offset;

[0039] Fig. 15 shows schematic view of an example of channels, which do not have the same temporal offset;

[0040] Fig. 16 shows a schematic view of an example of a plurality of channels, for which cross-channel prediction is performed; and

[0041] Fig. 17 shows a schematic view of an encoder and decoder for coding a datastream.

[0042] Detailed Description of the Embodiments

[0043] Equal or equivalent elements or elements with equal or equivalent functionality are denoted in the following description by equal or equivalent reference numerals even if occurring in different figures.

[0044] FH250307PEP-2026097809feIn the following description, a plurality of details is set forth to provide a more throughout explanation of embodiments of the present invention. However, it will be apparent to those skilled in the art that embodiments of the present invention may be practiced without these specific details. In other instances, well-known structures and devices are shown in block diagram form rather than in detail in order to avoid obscuring embodiments of the present invention. In addition, features of the different embodiments described herein after may be combined with each other, unless specifically noted otherwise.

[0045] Introductory remarks:

[0046] In the following, a general framework of a coding scheme (e.g., for a decoder and / or an encoder) will be described according to a first, second, and third aspect. However, these aspects are freely combinable, e.g., wherein any one or more features of one aspect is freely combinable with any features of any other aspect. Features of these aspects are subsequently described in more detailed examples in sections “General description”, “Role of LMS prediction with respect to the selected block prediction modes”, “Description of block matching prediction mode”, “Description of cross channel prediction mode”, “Description of DC prediction mode”, and “Description of the LMS prediction”. A broader framework is subsequently described with reference to fig. 17. Features described aspects, sections and the framework are freely combinable with any other embodiment disclosed herein.

[0047] Also, further embodiments will be defined by the enclosed claims.

[0048] It should be noted that any embodiments as defined by the claims can be supplemented by any of the details (features and functionalities) described in the above-mentioned sections.

[0049] Also, the embodiments described in the above-mentioned sections can be used individually, and can also be supplemented by any of the features in another section, or by any feature included in the claims.

[0050] Also, it should be noted that individual aspects described herein can be used individually or in combination. Thus, details can be added to each of said individual aspects without adding details to another one of said aspects.

[0051] Moreover, features and functionalities disclosed herein relating to a method can also be used in an apparatus (configured to perform such functionality). Furthermore, any features

[0052] FH250307PEP-2026097809feand functionalities disclosed herein with respect to an apparatus can also be used in a corresponding method. In other words, the methods disclosed herein can be supplemented by any of the features and functionalities described with respect to the apparatuses.

[0053] Moreover, features and functionalities disclosed herein relating to a method, in particular an encoding method can also be used in a data stream or bitstream (e.g. defining a respective data stream or bitstream element). Furthermore, any features and functionalities disclosed herein with respect to an apparatus can also be used in a corresponding data stream, e.g. as a resulting data stream as providing by said encoder. In other words, the data streams disclosed herein can be supplemented by any of the features and functionalities described with respect to the apparatuses and methods.

[0054] Also, any of the features and functionalities described herein can be implemented in hardware or in software, or using a combination of hardware and software, as will be described in the section “Implementation alternatives”.

[0055] Herein are described a first, second and third aspect. The first aspect is largely focused on prediction using filter coefficients and an updating thereof. The second aspect is largely focused on predicting coefficients additionally from one or more reference channels if a prediction mode related criteria is fulfilled. The third aspect relates to the entire coefficient prediction process being performed depending on the prediction mode related criteria. While any of the three aspects may be provided in isolation, it is noted that two or three aspects may be combined entirely or in parts (e.g., including one or more features from one aspect in the other aspect).

[0056] Fig. 1 shows a schematic view of a decoder 12 for block-wise decoding a digital time-varying signal 92 (e.g., a seismic signal, biomedical signal, or audio signal) from a data stream 16.

[0057] The decoder 12 is configured to decode 204 a signaled residual transform 202 (e.g., transformed from a time domain to a frequency domain) fora currently decoded block 140a from the data stream 16.

[0058] The decoder 12 is configured to sequentially reconstruct 206 (e.g., coefficient-wise, e.g., in a coding order from a transform coefficient associated with a highest frequency to a transform coefficient associated with a lowest frequency) coefficients 208 of a residual transform

[0059] FH250307PEP-2026097809fe210 of the currently decoded block 140a by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210 based on filter coefficients 218 (e.g., defining weights, e.g., for a weighted sum) to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202 (e.g., based on as a sum of the predictor 222 and the signaled coefficient value 224), corresponding to the currently reconstructed coefficient 214.

[0060] The decoder 12 is configured to reconstruct 228 the currently decoded block 140a by correcting 306 a sample predictor 308 for the currently decoded block 140a based on the residual transform 210 (e.g., based on a coefficient-wise sum, e.g., wherein a retransformation may be applied before or after the reconstruction, e.g., retransformation from frequency to time domain).

[0061] The decoder 12 is configured to, in sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently decoded block 140a, adapt 230 the filter coefficients 218 (e.g., change values of one or more or all filter coefficients, e.g., so as to reduce a difference between the predictor and the corrected predictor) based on

[0062] 1) a) the predictor 222 for the currently reconstructed coefficient 214, and the signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214, or b) the currently reconstructed coefficient 214, and

[0063] 2) the one or more previously reconstructed coefficients 216 of the residual transform 210

[0064] so as to be used for a subsequently reconstructed coefficient 232 of the residual transform 210.

[0065] Any coding device (e.g., encoder 10 and / or decoder 12) disclosed herein may comprise a processor (e.g., microprocessor) configured to perform the functions (e.g., steps) of the coding device. The coding device may be part of or may comprise a computer, smartphone, tablet, smart-watch, medical device, displaying device, server, or any other form of cloud computing resource. The coding device may be part of or comprise a medical device (e.g., electroencephalograph and / or electrocardiograph), a seismic device (e.g., seismograph or seismometer), or an audio processing device. The digital time-varying signal 92 (e.g., comprising or forming a waveform signal) may be a one channel signal, may be part of a multi-

[0066] FH250307PEP-2026097809fechannel signal (e.g., one channel thereof), or be a multi-channel signal. The digital timevarying signal 92 (e.g., or multi-channel waveform signal) may represent (e.g., may be or may comprise) a biometric signal (e.g., of heart, brain, or eye functions, e.g., body temperature), seismic data (e.g., amplitude of ground motion), or a sound signal (e.g., sound amplitude, e.g., of one or more audio channels). The digital time-varying signal 92 (e.g., or multi-channel digital time-varying signal 92) may be a time varying signal (e.g., in units of ms, e.g., in units of samples). A combination of an encoder and decoder may be referred to as codec. The coding device may be or may be part of a codec. The term “codec” may be used herein to refer to either apparatus and may be considered a term related to a general coding concept (e.g., not necessarily limited to either of a decoder or encoder aspect).

[0067] A block may form (or define) a (one-dimensional) sequence of samples (e.g., sequence of values in a non-transform domain, e.g., in a time domain). Blocks within a channel may have the same length (e.g., sample number, e.g., duration) or may have different lengths. In case the digital time-varying signal 92 has more than one channel, blocks may have the same of different length across all channels or a subset of the channels.

[0068] A transform may form of (define) a (one-dimensional) sequence of coefficients (e.g., sequence of values in a transform domain, e.g., in a frequency domain). Subjecting a block to a transform function may yield a transform and subjecting a transform to a re-transformation function may yield a block. The (re-)transformation function may comprise one or more of a discrete cosine transform (DCT), discrete sine transform (DST), and fast-Fourier transform (FFT).

[0069] The one or more previously reconstructed coefficients 216 of the residual transform 210 may comprise all previously reconstruction coefficients 216 of the residual transform 210. For example, the currently reconstructed coefficient 214 may be an eights coefficient, wherein all seven previously reconstructed coefficients 216 may be used to predict the currently reconstructed coefficient 214. In a different example, a maximum amount (e.g., in a range of 2 to 64) of previously reconstructed coefficients 216 may be used for to predict the currently reconstructed coefficient 214 (e.g., in case of a maximum number of four coefficients, all previously reconstructed coefficients 216 are used for predicting from the second to the fifth coefficient, from which point on, only the four most previously reconstructed coefficients 216 are used). An order of a plurality of previously reconstructed coefficients 216 may be defined by a scan order, wherein the scan order may be defined from higher frequencies to lower frequencies (e.g., from left to right, e.g., starting with a coefficient for the

[0070] FH250307PEP-2026097809fehighest frequency and ending at a coefficient with a lowest frequency).

[0071] The filter coefficients 218 may define (or be) weights (e.g., one dimensional array of weights, e.g., a two-dimensional array of weights in case of using coefficients of more than one channel). The weights (e.g., filter coefficients) may be defined in a weight vector and the previously reconstructed coefficients 216 to be used may be defined in a coefficient vector, wherein, for example, the currently reconstructed coefficient 214 may be predicted based on (or as) a scalar product of the weight vector and the coefficient vector. Values of the weights may (e.g., monotonically) increase in the scan direction (e.g., increase towards the currently reconstructed coefficient 214), e.g., wherein the weights may have fixed values or may be updated (e.g., with each predicted coefficient).

[0072] The predictor 222 may be corrected in a transform domain or non-transform domain (e.g., based on as a sum of the predictor 222 and the signaled coefficient value 224). The sample predictor 308 may be corrected in a transform domain or non-transform domain. The predictor 308 and the residual transform 210 may both be subjected to a re-transformation (e.g., into the time domain) before correcting the predictor 308. In a different example, the predictor 308 may be corrected in a transformation domain and the corrected predictor 308 may be subjected to a re-transformation.

[0073] The decoder 12 may be configured to adapt one or more (e.g., all) filter coefficients 218 so that a difference between the predictor 222 and the currently reconstructed coefficient 214 is reduced. In other words, the decoder 12 may be configured to adapt one or more (e.g., all) filter coefficients 218 so that a difference between the predictor 22 and the currently reconstructed coefficient 214 is smaller (e.g., or zero) using the adapted filter coefficients 218 compared to using the not (e.g., not yet) adapted filter coefficients 218.

[0074] Adapting the filter coefficients 218 may comprise changing a number of filter coefficients 218 (or of weights defined thereby). For example, the number of filter coefficients 218 may be increase by one having predicted the currently reconstructed coefficients, e.g., if there is no upper threshold for the number of the filter coefficients 218 or when such an upper threshold is not used. The filter coefficients 218 may have a fixed association with the previously reconstructed coefficients 216. In other words, the filter coefficients 218 may have a fixed association with coefficient positions. For example, a filter coefficient 218 combined (e.g., multiplied with) with a previously reconstructed coefficient 216 at coefficient position i may also be combined with the same previously reconstructed coefficient 216 at coefficient

[0075] FH250307PEP-2026097809feposition i for the next coefficient 214 to be predicted. Such a fixed association may, for example, be used if all previously reconstructed coefficient 216 of the block are to be used. In a different example, the filter coefficients 218 have a fixed position (e.g., in regards to coefficient positions) relative to the currently to be predicted coefficient 214 (e.g., wherein pre-determined number of filter coefficients 218 are combined with the same-predetermined number of previously reconstructed coefficient 216 immediately preceding the currently to be predicted coefficient 214). In other words, the filter coefficients 218 may be defined in a template that moves with the currently to be predicted coefficient 214. Such an arrangement may be used, for example, if there is an upper threshold for the number of filter coefficients 218. For example, the prediction may use (a maximum of) four previously reconstructed coefficient 216 with four filter coefficients 218 (or any other number), in which case the currently to be predicted coefficient 214 may be predicted by combining (e.g., forming a weighted sum of) the four previously reconstructed coefficient 216 preceding the four previously reconstructed coefficient 216 with the four filter coefficients 218.

[0076] Adapting the filter coefficients 218 may comprise changing (e.g., by multiplication with a rescaling factor, e.g., sum with a difference) of one or more filter coefficients, e.g., rescaling their values using a common factor (e.g., with optional rounding, e.g., so as to adapt to a numeric data type) for one or more (or all) filter coefficients 218 or individual factors (or differences) for one or more (or all) filter coefficients). The filter coefficients 218 may be adapted using gradient update, e.g., and fixed learning. For example, the decoder 12 may be configured to adapt the filter coefficients 218 using a least mean square algorithm (of filter).

[0077] Fig. 2 shows a schematic view of another example of a decoder 12. The decoder 12 may comprise or be any other decoder 12 disclosed herein. It is noted that with reference to fig.

[0078] 2, a plurality of features will be described in combination for the sake of conciseness, but is are not limited to this specific combination. Any feature may be provided in isolation or in any combination of features.

[0079] The digital time-varying signal 28 may comprise multiple channels (e.g., M channels, e.g., with M = 2, e.g., M = 3, 4, 5, or larger), each partitioned into consecutive blocks which are mutually temporally aligned (e.g., having aligned sample borders, e.g., having temporally aligned block starts and ends, e.g., having the same block size), from the data stream 16.

[0080] FH250307PEP-2026097809feThe decoder 12 may be configured to reconstruct each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes (e.g., comprising one or more of a DC description mode, a cross-channel prediction mode, an intra-channel copy mode) to obtain a sample predictor (e.g., in the non-transform domain, e.g., time domain) for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream 16 (e.g., by re-transforming the residual transform into the non-transform domain; e.g., by performing a linear combination between the sample predictor and the retransformed residual transform or a dequantized version thereof).

[0081] The currently decoded block 140a may be of a first channel 302 (e.g., wherein “first” does not necessarily indicate a first position in a channel direction, but indicates a specific channel, e.g., a respective channel, e.g., a pre-determined channel) and the decoder 12 may be configured to reconstruct 228 the currently decoded block 140a based on the residual transform 210 by correcting 306 a sample predictor 308 of the currently decoded block 140a based on the residual transform 210.

[0082] The decoder 12 may be configured to, in sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently decoded block 140a, depending on whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship (e.g., having the same prediction mode, e.g., having a same temporal referencing offset, e.g., using one or more same prediction parameter within the same prediction mode, e.g., using the same prediction mode with the same prediction parameters) with the selected prediction mode of the currently decoded block 140a, predict 220 the currently reconstructed coefficient 214 of the residual transform 210 additionally from a corresponding (e.g. “corresponding coefficient” is “spectrally collocated (to the currently reconstructed coefficient) coefficient”, e.g., having the same coefficient index) coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 (e.g., using a rescaled or weighted version of the corresponding coefficient, e.g., based on a weighted sum that includes previously reconstructed coefficients 216 of the reconstructed transform and the corresponding coefficient 309 of the residual transform 310), and adapt 230 the filter coefficients 218 additionally based on the corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306.

[0083] FH250307PEP-2026097809feFig. 3 shows a schematic view of an example of a block 304 and the currently decoded block 140a with respective temporal offsets At1 , At2.

[0084] The decoder 12 may be configured so that the predetermined relationship is fulfilled if one or all of one or more out of the following conditions are true (e.g., at least one is true, at least two are true, e.g., all are true, e.g., only a specific condition is true; e.g. merely conditions a and b might be subject of the dependency, or merely a and c or merely a or merely b or merely c; note that it might by that more than the example conditions set out hereinbelow may cause that the predetermined relationship is fulfilled):

[0085] a) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306 and the selected prediction mode of the currently decoded block 140a are a block matching mode according to which the sample predictor for a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs (e.g., a block of samples having the same samples in the same order as obtained from a sequence of already decoded samples), displaced from the block to be predicted by a temporal offset (e.g., measured in time and / or samples), and the temporal offset is equal among the one or more blocks 304 and the currently decoded block 140a (e.g., in fig. 3, a temporal offset At2 for the currently decoded block 140a and a temporal offset At1 for the block 304 may be equal, thusly fulfilling a corresponding predetermined relationship),

[0086] b) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306 and the selected prediction mode of the currently decoded block 140a are a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels (e.g., in fig. 3, currently decoded block 140a may reference a block in the same vertical column of blocks and block 304 may reference a block in the same vertical column of blocks, e.g., wherein optionally block 140a block 304 reference the same block or reference by a same channel difference),

[0087] c) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306 and the selected prediction mode of the currently decoded block 140a are all of a prediction mode (e.g. possibly mutually different or mutually equal prediction modes) comprised by a set of prediction modes comprising, or consisting of, one or more of:

[0088] FH250307PEP-2026097809fea zero prediction mode according to which a zero signal is used as the sample predictor for a block to be predicted (e.g., a consecutive sequence of zeroes), a DC prediction mode according to which the sample predictor for a block to be predicted is a constant signal of a constant value with predicting the constant value based on a previously reconstructed portion of the digital time-varying signal (e.g., based on an average or weighted average),

[0089] a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels.

[0090] Criteria for the predetermined relationship may comprise one or more of the use of same prediction mode (e.g., DC-prediction, cross-channel prediction, inter-channel prediction), and the use of one or more (e.g., all) same prediction parameters (e.g., same DC-value for DC-prediction, same reference channel for cross-channel prediction, e.g., same channel distance for cross-channel prediction, e.g., same temporal offset for inter-channel prediction). The same prediction mode may be defined by the same flags that define the prediction mode.

[0091] The decoder 12 may be configured to perform the steps of predicting and adapting depending on whether the blocks 304 of all reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship (e.g., all channels using the same prediction mode, e.g., using the same one or more prediction parameters).

[0092] The decoder 12 may be configured to perform the steps of predicting and adapting only using corresponding coefficient 309 of the residual transform 310 of blocks 304 of the reference channels 306 that fulfill the predetermined relationship (e.g., only from reference channels, for which the same prediction mode and / or the same one or more prediction parameters have been used as for the first channel).

[0093] In predicting 220 the currently reconstructed coefficient 214 of the residual transform 210 from one or more previously reconstructed coefficients 216 of the residual transform 210 based on filter coefficients 218 to obtain a predictor 222 for the currently reconstructed coefficient (214), the decoder 12 may be configured to subject the one or more previously reconstructed coefficients 216 of the residual transform 210 to a weighted sum, wherein at least one (e.g., all) filter coefficients 218 may form weights for one or more previously reconstructed coefficients 216.

[0094] FH250307PEP-2026097809feIn adapting 230 the filter coefficients 218, the decoder 12 may be configured to determine adapted filter coefficients (e.g., by increasing or decreasing one or more or all filter coefficients) so that a difference between the corrected predictor 222 and an updated predictor obtained from the one or more previously reconstructed coefficients 216 of the residual transform 210 based on the adapted filter coefficients 218 is smaller than a difference between the corrected predictor 222 and the predictor 222 obtained from the un-adapted filter coefficients 218 (e.g., by using a gradient update which may optionally have a fixed learning rate, e.g., based on a derivative of a difference of squared difference, e.g., multiplied with a learning rate; e.g., by determining a ratio between the corrected predictor 222 and the predictor 222 and rescaling the filter coefficients based on or by the ratio, which may comprise rounding the filter coefficients).

[0095] The decoder 12 may be configured to sequentially reconstruct the coefficients 208 of a residual transform 210 in a coding order 210 (e.g., scan order) from a transform coefficient of the residual transform 210 of the currently decoded block 140a associated with a highest frequency (e.g., right-most transform coefficient) to a transform coefficient of the residual transform 210 of the currently decoded block 140a associated with a lowest frequency (e.g., left-most transform coefficient, e.g., DC-coefficient) (e.g., wherein the coding order is reverse to a coefficient order in the residual transform).

[0096] Fig. 4 shows a flow diagram of a method 400 for block-wise decoding a digital time-varying signal 28; 92) from a data stream 16. The method 400 may be performed by any decoder 12 disclosed herein (e.g., decoder 12 shown in fig. 1 and 2).

[0097] The method 400 comprises, in step 204, decoding a signaled residual transform 202 (e.g., transformed from a time domain to a frequency domain) for a currently decoded block 140a from the data stream 16.

[0098] The method 400 comprises, in step 402, sequentially reconstructing 206 (e.g., coefficientwise, e.g., in a coding orderfrom a transform coefficient associated with a highest frequency to a transform coefficient associated with a lowest frequency) coefficients 208 of a residual transform 210 of the currently decoded block 140a by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210 based on filter coefficients 218 (e.g., defining weights, e.g., fora

[0099] FH250307PEP-2026097809feweighted sum) to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202 (e.g., based on as a sum of the predictor and the signaled coefficient value), corresponding to the currently reconstructed coefficient 214.

[0100] The method 400 comprises, in step 404, reconstructing 228 the currently decoded block 140a by correcting 306 a sample predictor 308 for the currently decoded block 140a based on the residual transform 210 (e.g., based on a coefficient-wise sum, e.g., wherein a retransformation may be applied before orafterthe reconstruction, e.g., retransformation from frequency to time domain).

[0101] The method comprises, in sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently decoded block 140a, in step 406, adapting 230 the filter coefficients 218 (e.g., change values of one or more or all filter coefficients, e.g., so as to reduce a difference between the predictor and the corrected predictor) based on 1) a) the predictor 222 for the currently reconstructed coefficient 214, and the signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214, or b) the currently reconstructed coefficient 214, and 2) the one or more previously reconstructed coefficients 216 of the residual transform 210 so as to be used for a subsequently reconstructed coefficient 232 of the residual transform 210.

[0102] The method 400 may comprise any other function and / or feature of any decoder 12 disclosed herein (e.g., the decoder 12 shown in fig. 2).

[0103] The method 400 may comprise one or more of receiving, transmitting, and storing the data stream 16. The method 410 may comprise playing back (e.g., reproducing) the digital time varying signal 92.

[0104] Fig. 5 shows a schematic view of an encoder 10 for block-wise encoding a digital timevarying signal 28; 92 (e.g., a seismic signal, biomedical signal, or audio signal) into a data stream 16. The encoder 10 may be configured to encode a data stream 16 decodable by any decoder 12 disclosed herein (e.g., with reference to fig. 1 and 2).

[0105] The encoder 10 is configured to encode 204 a signaled residual transform 202 (e.g., transformed from a time domain to a frequency domain) for a currently encoded block 140a into the data stream 16 so that coefficients 208 of a residual transform 210 of the currently encoded block 140a is sequentially reconstructable 206 (e.g., coefficient-wise, e.g., in a coding

[0106] FH250307PEP-2026097809feorder from a transform coefficient associated with a highest frequency to a transform coefficient associated with a lowest frequency) by predicting 222 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210 based on filter coefficients 218 (e.g., defining weights, e.g., for a weighted sum) to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202 (e.g., based on as a sum of the predictor 222 and the signaled coefficient value 224), corresponding to the currently reconstructed coefficient 214.

[0107] The currently encoded block 140a is reconstructable by correcting 306 a sample predictor 308 for the currently encoded block 140a based on the residual transform 210 (e.g., based on a coefficient-wise sum, e.g., wherein a retransformation may be applied before or after the reconstruction, e.g., retransformation from frequency to time domain).

[0108] The sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently encoded block 140a comprises adapting 230 the filter coefficients 218 (e.g., change values of one or more or all filter coefficients, e.g., so as to reduce a difference between the predictor and the corrected predictor) based on 1) a) the predictor 222 for the currently reconstructed coefficient 214, and the signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214, or b) the currently reconstructed coefficient 214, and 2) the one or more previously reconstructed coefficients 216 of the residual transform 210 so as to be used for a subsequently reconstructed coefficient 232 of the residual transform 210.

[0109] The encoder 10 may be configured to perform one or more steps for reconstructing the currently encoded block 140a, e.g., one or more steps of predicting 222 the currently reconstructed coefficient 214 of the residual transform 210 from one or more previously reconstructed coefficients 216 of the residual transform 210 based on the filter coefficients 218 to obtain a predictor 222 for the currently reconstructed coefficient 214, correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202 corresponding to the currently reconstructed coefficient 214, correcting 306 the sample predictor 308 for the currently encoded block 140a based on the residual transform 210, and adapting 230 the filter coefficients 218.

[0110] FH250307PEP-2026097809feIn the following, further optional features for the encoder 10 are described for the sake of conciseness jointly with reference to fig. 2. However, it is noted that any feature may be used in isolation or in combination with any one or more other features.

[0111] The digital time-varying signal 28 may comprise multiple channels (e.g., M channels, e.g., with M = 2, e.g., M = 3, 4, 5, or larger), each partitioned into consecutive blocks (e.g., block 304) which are mutually temporally aligned (e.g., having temporally aligned block starts and ends, e.g., having the same block size), into the data stream 16.

[0112] The encoder 10 may be configured to encode each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes (e.g., comprising one or more of a DC description mode, a cross-channel prediction mode, an intra-channel copy mode) to obtain a sample predictor for the respective block, and encoding a residual transform for the respective block encoded into the data stream 16 for correcting the sample predictor so as to reconstruct the respective block.

[0113] The currently encoded block 140a may be of a first channel 302 (e.g., wherein “first” does not necessarily indicate a first position in a channel direction, but indicates a specific channel, e.g., a respective channel, e.g., a pre-determined channel) and the currently encoded block 140a may be reconstructable 228 based on the residual transform 210 by correcting 306 a sample predictor 308 of the currently encoded block 140a based on the residual transform 210, and in sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently encoded block 140a, depending on whetherthe selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently encoded block 140a, fulfill a predetermined relationship (e.g., having the same prediction mode, e.g., having a same temporal referencing offset, e.g., using one or more same prediction parameter within the same prediction mode, e.g., using the same prediction mode with the same prediction parameters) with the selected prediction mode of the currently encoded block 140a, predicting 220 the currently reconstructed coefficient 214 of the residual transform 210 additionally from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 (e.g., using a rescaled or weighted version of the corresponding coefficient, e.g., based on a weighted sum that includes previously reconstructed coefficients 216 of the reconstructed transform and the corresponding coefficient 309 of the residual transform), and adapting 230 the filter coefficients 21 Sadditionally based on the corresponding

[0114] FH250307PEP-2026097809fecoefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306.

[0115] The encoder 10 may be configured so that the predetermined relationship is fulfilled if one or all of one or more out of the following conditions are true (e.g., at least one is true, at least two are true, e.g., all are true, e.g., only a specific condition is true; e.g. merely conditions a and b might be subject of the dependency, or merely a and c or merely a or merely b or merely c; note that it might by that more than the example conditions set out herein below may cause that the predetermined relationship is fulfilled):

[0116] a) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306 and the selected prediction mode of the currently encoded block 140a are a block matching mode according to which the sample predictor for a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs, displaced from the block to be predicted by a temporal offset, and the temporal offset is equal among the one or more blocks 304 and the currently encoded block 140a (e.g., in fig. 3, a temporal offset At2 for the currently encoded block 140a and a temporal offset At1 for the block 304 may be equal, thusly fulfilling a corresponding predetermined relationship),

[0117] b) the selected prediction modes of the one or more blocks 304of one or more reference channels 306 and the selected prediction mode of the currently encoded block 140aare a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels (e.g., in fig. 3, currently encoded block 140a may reference a block in the same vertical column of blocks and block 304 may reference a block in the same vertical column of blocks, e.g., wherein optionally block 140a block 304 reference the same block or reference by a same channel difference),

[0118] c) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306 and the selected prediction mode of the currently encoded block 140a are all of a prediction mode (e.g. possibly mutually different or mutually equal prediction modes) comprised by a set of prediction modes comprising, or consisting of, one or more of:

[0119] a zero prediction mode according to which a zero signal is used as the sample predictor for a block to be predicted,

[0120] FH250307PEP-2026097809fea DC prediction mode according to which the sample predictor for a block to be predicted a constant signal of a constant value with predicting the constant value based on a previously reconstructed portion of the digital time-varying signal, a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels.

[0121] Criteria for the predetermined relationship may comprise one or more of the use of the same prediction mode (e.g., DC-prediction, cross-channel prediction, inter-channel prediction), and the use of one or more (e.g., all) same prediction parameters (e.g., same DC-value for DC-prediction, same reference channel for cross-channel prediction, e.g., same channel distance for cross-channel prediction, e.g., same temporal offset for inter-channel prediction).

[0122] The encoder 10 may be configured to perform the steps of predicting and adapting depending on whether the blocks 304 of all reference channels 306, which are temporally aligned to the currently encoded block 140a, fulfill a predetermined relationship.

[0123] The encoder 10 may be configured to perform the steps of predicting and adapting only using corresponding coefficient 309 of the residual transform 310 of blocks 304 of the reference channels 306that fulfill the predetermined relationship (e.g., only from reference channels, for which the same prediction mode and / or the same one or more prediction parameters have been used as for the first channel).

[0124] In predicting 220 the currently reconstructed coefficient 214 of the residual transform 210 from one or more previously reconstructed coefficients 216 of the residual transform 210 based on filter coefficients 218 to obtain a predictor 222 for the currently reconstructed coefficient 214, the encoder 10 may be configured to subject the one or more previously reconstructed coefficients 216 of the residual transform 210 to a weighted sum, wherein at least one (e.g., all) filter coefficients form weights for one or more previously reconstructed coefficients 216.

[0125] In adapting 230 the filter coefficients 218, the encoder 10 may be configured to determine adapted filter coefficients (e.g., by increasing or decreasing one or more or all filter coefficients) so that a difference between the corrected predictor 222 and an updated predictor obtained from the one or more previously reconstructed coefficients 216of the residual

[0126] FH250307PEP-2026097809fetransform 210 based on the adapted filter coefficients 218 is smaller than a difference between the corrected predictor 222 and the predictor 222 obtained from the un-adapted filter coefficients 218 (e.g., by determining a ratio between the corrected predictor 222 and the predictor 222 and rescaling the filter coefficients based on or by the ratio, which may comprise rounding the filter coefficients).

[0127] The encoder 10 may be configured to sequentially reconstruct the coefficients 208 of a residual transform 210 in a coding order 210 (e.g., scan order) from a transform coefficient of the residual transform 210 of the currently encoded block 140a associated with a highest frequency (e.g., right-most transform coefficient) to a transform coefficient of the residual transform 210 of the currently encoded block 140a associated with a lowest frequency (e.g., left-most transform coefficient, e.g., DC-coefficient) (e.g., wherein the coding order is reverse to a coefficient order in the residual transform).

[0128] Fig. 6 shows a flow diagram of a method 410 for block-wise encoding a digital time-varying signal 92 (e.g., a seismic signal, biomedical signal, or audio signal) into a data stream 16. The method 410 may be performed by any encoder 10 disclosed herein (e.g., the encoder 10 shown in fig. 5). The method 410 may comprise any encoder-side pendant corresponding to any decoder and / or decoding method disclosed herein.

[0129] The method 410 comprises, in step 412, encoding a signaled residual transform 202 (e.g., transformed from a time domain to a frequency domain) for a currently encoded block 140a into the data stream 16 so that coefficients 208 of a residual transform 210 of the currently encoded block 140a are sequentially reconstructable 206 (e.g., coefficient-wise, e.g., in a coding order from a transform coefficient associated with a highest frequency to a transform coefficient associated with a lowest frequency) by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210 based on filter coefficients 218 (e.g., defining weights, e.g., for a weighted sum) to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202 (e.g., based on as a sum of the predictor and the signaled coefficient value), corresponding to the currently reconstructed coefficient 214, wherein the currently encoded block 140a is reconstructable by correcting 306 a sample predictor 308 for the currently encoded block 140a based on the residual transform 210 (e.g., based on a coef-

[0130] FH250307PEP-2026097809feficient-wise sum, e.g., wherein a retransformation may be applied before or after the reconstruction, e.g., retransformation from frequency to time domain), wherein the sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently encoded block 140a comprises adapting 230 the filter coefficients 218 (e.g., change values of one or more or all filter coefficients, e.g., so as to reduce a difference between the predictor and the corrected predictor) based on 1) a) the predictor 222 for the currently reconstructed coefficient 214, and the signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214, or b) the currently reconstructed coefficient 214, and 2) the one or more previously reconstructed coefficients 216 of the residual transform 210 so as to be used for a subsequently reconstructed coefficient 232 of the residual transform 210.

[0131] The method 410 may comprise one or more steps for reconstructing the currently encoded block 140a, e.g., one or more steps of predicting 222 the currently reconstructed coefficient 214 of the residual transform 210 from one or more previously reconstructed coefficients 216 of the residual transform 210 based on the filter coefficients 218 to obtain a predictor 222 for the currently reconstructed coefficient 214, correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202 corresponding to the currently reconstructed coefficient 214, correcting 306 the sample predictor 308 for the currently encoded block 140a based on the residual transform 210, and adapting 230 the filter coefficients 218.

[0132] The method 410 may comprise one or more of storing, transmitting, receiving, and uploading the data stream 16. The method 410 may comprise recording the digital time varying signal 92.

[0133] Further is provided a data stream 16 (e.g., stored on a computer-readable medium, e.g., non-transitory storage medium) encoded according to any encoding method (and / or encoder 10) disclosed herein (e.g., decodable by any decoding method and / or decoder 12 disclosed herein).

[0134] Further is provided a computer program product (e.g., computer program, e.g., stored on a computer-readable medium, e.g., non-transitory storage medium) for implementing any method disclosed herein (e.g., any encoding method and / or any decoding method disclosed herein) when being executed on a computer or signal processor.

[0135] In the following, the second aspect will be described. Any feature described with respect to the first aspect may be applicable to the second aspect and vice versa.

[0136] FH250307PEP-2026097809feFig. 7 shows a schematic view of a decoder 12 for block-wise decoding a digital time-varying signal 92 from a data stream 16.

[0137] The digital time-varying signal 92 comprises multiple channels 302, 306 (e.g., M channels 231, e.g,. with M = 2, e.g., M = 3, 4, 5, or larger, e.g., 3 channels 302, 306 as exemplarily depicted in fig. 7), each partitioned into consecutive blocks 304 (e.g., temporal blocks 140 or any other block) which are mutually temporally aligned (e.g., having temporally aligned block starts and ends, e.g., having the same block size, e.g., having temporally aligned sample borders).

[0138] The decoder 12 is configured to reconstruct each block 304, 140 of each channel 302, 306 by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes (e.g., comprising one or more of a DC description mode, a cross-channel prediction mode, an intra-channel copy mode) to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block 140a of a first channel 302 by decoding 204 the signaled residual transform 202 for the currently decoded block 140a from the data stream 16; sequentially reconstructing 206 (e.g., coefficient-wise, e.g., in a coding order from a transform coefficient associated with a highest frequency to a transform coefficient associated with a lowest frequency, e.g., according to the first aspect) coefficients 208 of the residual transform 210 of the currently decoded block 140a by predicting 222 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210 to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214 (e.g., based on as a sum of the predictor and the signaled coefficient value), reconstructing 228 the currently decoded block 140a by correcting 306 a sample predictor 308 of the currently decoded block 140a based on the residual transform 210 (e.g., based on a coefficient-wise sum, e.g., wherein a retransformation may be applied before or after the reconstruction, e.g., retransformation from frequency to time domain).

[0139] FH250307PEP-2026097809feThe decoder 12 is configured to, in sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently decoded block 140a, depending on whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship (e.g., having the same prediction mode, e.g., having a same temporal referencing offset, e.g., using one or more same prediction parameter within the same prediction mode, e.g., using the same prediction mode with the same prediction parameters) with the selected prediction mode of the currently decoded block 140a, predict 222 the currently reconstructed coefficient 214 of the residual transform 210 additionally from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 (e.g., using a rescaled or weighted version of the corresponding coefficient, e.g., based on a weighted sum that includes previously reconstructed coefficients 216 of the reconstructed transform and the corresponding coefficient 309 of the residual transform) (e.g., wherein the decoder is configured to use a default prediction, wherein, dependent on the predetermined relationship, the currently reconstructed coefficient 214 of the residual transform 210 is additionally predicted from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306.

[0140] In the following, further optional features for the decoder 12 are described for the sake of conciseness jointly with reference to fig. 2. However, it is noted that any feature may be used in isolation or in combination with any one or more other features.

[0141] The decoder 12 may be configured so that the predetermined relationship is fulfilled if one or all of one or more out of the following conditions are true (e.g., at least one is true, at least two are true, e.g., all are true, e.g., only a specific condition is true):

[0142] a) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306 and the selected prediction mode of the currently decoded block 140a are a block matching mode according to which the sample predictor for a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs, displaced from the block to be predicted by a temporal offset, and the temporal offset is equal among the one or more blocks 304 and the currently decoded block 140a (e.g., in fig. 3, a temporal offset At2 for the currently decoded block 140a and a temporal offset At1 for the block 304 may be equal, thusly fulfilling a corresponding predetermined relationship),

[0143] FH250307PEP-2026097809feb) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306) and the selected prediction mode of the currently decoded block 140a are a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels (e.g., in fig. 3, currently decoded block 140a may reference a block in the same vertical column of blocks and block 304 may reference a block in the same vertical column of blocks, e.g., wherein optionally block 140a block 304 reference the same block or reference by a same channel difference),

[0144] c) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306 and the selected prediction mode of the currently decoded block 140a are all of a prediction mode (e.g. possibly mutually different or mutually equal prediction modes) comprised by a set of prediction modes comprising, or consisting of, one or more of:

[0145] a zero prediction mode according to which a zero signal is used as the sample predictor for a block to be predicted,

[0146] a DC prediction mode according to which the sample predictor for a block to be predicted a constant signal of a constant value with predicting the constant value based on a previously reconstructed portion of the digital time-varying signal, a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels.

[0147] Criteria for predetermined relationship may comprise one or more of the use of the same prediction mode (e.g., DC-prediction, cross-channel prediction, inter-channel prediction), and the use of one or more (e.g., all) same prediction parameters (e.g., same DC-value for DC-prediction, same reference channel for cross-channel prediction, e.g., same channel distance for cross-channel prediction, e.g., same temporal offset for inter-channel prediction).

[0148] The decoder 12 may be configured predict 220 the currently reconstructed coefficient 214 of the residual transform 210 additionally from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 depending on whether the blocks 304 of all reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship.

[0149] FH250307PEP-2026097809feThe decoder 12 may be configured to predict 220 the currently reconstructed coefficient 214 of the residual transform 210 additionally from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 only using corresponding coefficient 309 of the residual transform 310 of blocks 304 of the reference channels 306 that fulfill the predetermined relationship (e.g., only from reference channels, for which the same prediction mode and / or the same one or more prediction parameters have been used as for the first channel 302).

[0150] The decoder 12 may be configured to sequentially reconstruct 206 coefficients 208 of the residual transform 210 of the currently decoded block 140a in a coding order 210 (e.g., scan order) from a transform coefficient of the residual transform 210 of the currently decoded block 140a associated with a highest frequency (e.g., right-most transform coefficient) to a transform coefficient of the residual transform 210 of the currently decoded block 140a associated with a lowest frequency (e.g., left-most transform coefficient, e.g., DC-coefficient) (e.g., wherein the coding order is reverse to a coefficient order in the residual transform).

[0151] Fig. 8 shows a flow diagram of a method 420 for decoding a digital time-varying signal 92 from a data stream 16. The method 420 may be performed by any decoder 12 disclosed herein (e.g., decoder 12 shown in fig. 7).

[0152] The digital time-varying signal 92 comprises multiple channels 302, 306 (e.g., M channels, e.g., with M = 2, e.g., M = 3, 4, 5, or larger), each partitioned into consecutive blocks (e.g., temporal blocks 140) which are mutually temporally aligned (e.g., having temporally aligned block starts and ends, e.g., having the same block size).

[0153] The method 420 comprises, in step 422, reconstructing each block 140a, 304 of each channel 302, 306 by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes (e.g., comprising one or more of a DC description mode, a cross-channel prediction mode, an intra-channel copy mode) to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block 140a of a first channel 302 by decoding 204 the signaled residual transform 202 for the currently decoded block 140a from the data stream 16; sequentially reconstructing 206 (e.g., coefficient-wise, e.g., in a coding order from a transform coefficient associated with a highest frequency to a transform coefficient associated with a

[0154] FH250307PEP-2026097809felowest frequency) coefficients 208 of the residual transform 210 of the currently decoded block 140a by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210 to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214 (e.g., based on as a sum of the predictor and the signaled coefficient value), reconstructing 228 the currently decoded block 140a by correcting 306 a sample predictor 308 of the currently decoded block 140a based on the residual transform 210 (e.g., based on a coefficient-wise sum, e.g., wherein a retransformation may be applied before orafterthe reconstruction, e.g., retransformation from frequency to time domain).

[0155] The method comprises in step 424, in sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently decoded block 140a, depending on whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship (e.g., having the same prediction mode, e.g., having a same temporal referencing offset, e.g., using one or more same prediction parameter within the same prediction mode, e.g., using the same prediction mode with the same prediction parameters) with the selected prediction mode of the currently decoded block 140a, predict 220 the currently reconstructed coefficient 214 of the residual transform 210 additionally from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 (e.g., using a rescaled or weighted version of the corresponding coefficient, e.g., based on a weighted sum that includes previously reconstructed coefficients 216 of the reconstructed transform and the corresponding coefficient 309 of the residual transform) (e.g., wherein the decoder is configured to use a default prediction, wherein, dependent on the predetermined relationship, the currently reconstructed coefficient 214 of the residual transform 210 is additionally predicted from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306).

[0156] The method 420 may comprise any other function and / or feature of any decoder 12 disclosed herein (e.g., the decoder 12 shown in fig. 2). The method 420 may comprise one or more of receiving, transmitting, and storing the data stream 16. The method 420 may comprise playing back (e.g., reproducing) the digital time varying signal 92.

[0157] FH250307PEP-2026097809feFig. 9 shows a schematic view of an encoder 10 for block-wise encoding a digital timevarying signal 92 into a data stream 16. The encoder 10 may be configured to encode a data stream 16 decodable by any decoder 12 disclosed herein (e.g., with reference to fig.

[0158] 7).

[0159] The digital time-varying signal 92 comprises multiple channels 302, 306 (e.g., M channels, e.g., with M = 2, e.g., M = 3, 4, 5, or larger), each partitioned into consecutive blocks (e.g., temporal blocks 140) which are mutually temporally aligned (e.g., having temporally aligned block starts and ends, e.g., having the same block size).

[0160] The encoder 10 is configured to encode each block 140a, 304 of each channel 302, 306 by predicting the respective block 140a using a selected prediction mode selected for the respective block out of supported prediction modes (e.g., comprising one or more of a DC description mode, a cross-channel prediction mode, an intra-channel copy mode) to obtain a sample predictor for the respective block 140a, and encode a residual transform 202 encoded for the respective block 140a into the data stream 16 for correcting the sample predictor to reconstruct the respective block 140a, with encoding a currently encoded block 140a of a first channel 302 by encoding 204 the signaled residual transform 202 for the currently encoded block 140a into the data stream 16 from which the currently encoded block 140a is reconstructable by sequentially reconstructing 206 (e.g., coefficient-wise, e.g., in a coding order from a transform coefficient associated with a highest frequency to a transform coefficient associated with a lowest frequency) coefficients 208 of the residual transform 210 of the currently encoded block 140a by predicting 222 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210 to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using an encoded coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214 (e.g., based on as a sum of the predictor and the signaled coefficient value), reconstructing 228 the currently encoded block 140a by correcting 306 a sample predictor 308 of the currently encoded block 140a based on the residual transform 210 (e.g., based on a coefficient-wise sum, e.g., wherein a retransformation may be applied before or after the reconstruction, e.g., retransformation from frequency to time domain).

[0161] FH250307PEP-2026097809feIn sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently encoded block 140a, depending on whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently encoded block 140a, fulfill a predetermined relationship (e.g., having the same prediction mode, e.g., having a same temporal referencing offset, e.g., using one or more same prediction parameter within the same prediction mode, e.g., using the same prediction mode with the same prediction parameters) with the selected prediction mode of the currently encoded block 140a, the currently reconstructed coefficient 214 of the residual transform 210 is to be predicted additionally from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 (e.g., using a rescaled or weighted version of the corresponding coefficient, e.g., based on a weighted sum that includes previously reconstructed coefficients 216 of the reconstructed transform and the corresponding coefficient 309 of the residual transform) (e.g., wherein the encoder is configured to use a default prediction, wherein, dependent on the predetermined relationship, the currently reconstructed coefficient 214 of the residual transform 210 is additionally predicted from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306).

[0162] The encoder 10 may be configured to perform one or more steps for reconstructing the currently encoded block 140a, e.g., one or more steps of predicting 222 the currently reconstructed coefficient 214 of the residual transform 210, correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202 corresponding to the currently reconstructed coefficient 214, correcting 306 the sample predictor 308 for the currently encoded block 140a based on the residual transform 210, and predicting the currently reconstructed coefficient 214 of the residual transform 210.

[0163] The encoder 10 may be configured to perform and / or support any encoder-side functionality corresponding to any decoder-side functions described herein (e.g., with reference to fig.

[0164] 7).

[0165] Fig. 10 shows a flow diagram of a method 430 for block-wise encoding a digital time-varying signal 92 into a data stream 16. The method 430 may be performed by any encoder 10 disclosed herein (e.g., the encoder 10 shown in fig. 9). The method 430 may comprise any encoder-side pendant corresponding to any decoder and / or decoding method disclosed herein.

[0166] FH250307PEP-2026097809feThe digital time-varying signal 92 comprises multiple channels 302, 306, each partitioned into consecutive blocks 140, 304 which are mutually temporally aligned.

[0167] The method 430 comprises, in step 432, encoding each block of each channel 302, 306 by predicting the respective block 140a, 304 using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and encode a residual transform 202 encoded for the respective block into the data stream 16 for correcting the sample predictor to reconstruct the respective block 140a, 304, with encoding a currently encoded block 140a of a first channel 302 by encoding 204 the signaled residual transform 202 for the currently encoded block 140a into the data stream 16 from which the currently encoded block 140a is reconstructable by sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently encoded block 140a by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more previously reconstructed coefficients 216 of the residual transform 210 to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using an encoded coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214, reconstructing 228 the currently encoded block 140a by correcting 306 a sample predictor 308 of the currently encoded block 140a based on the residual transform 210.

[0168] The method comprises in step 434, in sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently encoded block 140a, depending on whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently encoded block 140a, fulfill a predetermined relationship with the selected prediction mode of the currently encoded block 140a, the currently reconstructed coefficient 214 of the residual transform 210 is to be predicted additionally from a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306.

[0169] The method 430 may comprise one or more of storing, transmitting, receiving, and uploading the data stream 16. The method 430 may comprise recording the digital time varying signal 92.

[0170] FH250307PEP-2026097809feFurther is provided a data stream 16 (e.g., stored on a computer-readable medium, e.g., non-transitory storage medium) encoded according to any encoding method (and / or encoder 10) disclosed herein (e.g., decodable by any decoding method and / or decoder 12 disclosed herein).

[0171] Further is provided a computer program product (e.g., computer program, e.g., stored on a computer-readable medium, e.g., non-transitory storage medium) for implementing any method disclosed herein (e.g., any encoding method and / or any decoding method disclosed herein) when being executed on a computer or signal processor.

[0172] In the following, the third aspect is described. Since the third aspect relates to how the prediction criteria affects the coefficient prediction, the third aspects is also described with reference to fig. 7 and 9 (e.g., drawings for the second aspect can be used to describe the third aspect).

[0173] Provided is a decoder 12 for block-wise decoding a digital time-varying signal 92 from a data stream. The decoder 12 may be any decoder 12 disclosed herein and may comprise any decoder feature or decoder-pendant feature disclosed herein.

[0174] The digital time-varying signal 92 comprises multiple channels (e.g., M channels, e.g., with M = 2, e.g., M = 3, 4, 5, or larger), each partitioned into consecutive blocks (e.g., temporal blocks 140) which are mutually temporally aligned (e.g., having temporally aligned block starts and ends, e.g., having the same block size).

[0175] The decoder 12 is configured to reconstruct each block 140a, 304 of each channel 302, 306 by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes (e.g., comprising one or more of a DC description mode, a cross-channel prediction mode, an intra-channel copy mode) to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block 140a of a first channel 302 depending on (e.g. the coefficient wise prediction based on both, previous coefficients of the currently decoded block’s 140a residual transform as well as the one or more corresponding coefficients of one or more aligned blocks of reference channels might be a mode which is only signable or available for the currently decoded block 140a if the following is true; if not true, the coefficient wise prediction might be performed without the one or more corresponding coefficients of one or more

[0176] FH250307PEP-2026097809fealigned blocks of reference channels 306 (e.g., as done the second aspect), or without use of the coefficient wise prediction at all) whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship (e.g., having the same prediction mode, e.g., having a same temporal referencing offset, e.g., using one or more same prediction parameter within the same prediction mode, e.g., using the same prediction mode with the same prediction parameters) with the selected prediction mode of the currently decoded block 140a, by decoding 204 the signaled residual transform 202 for the currently decoded block 140a from the data stream 16; sequentially reconstructing 206 (e.g., coefficient-wise, e.g., in a coding order from a transform coefficient associated with a highest frequency to a transform coefficient associated with a lowest frequency) coefficients 208 of the residual transform 210 of the currently decoded block 140a by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210 and a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 (e.g., using a rescaled or weighted version of the corresponding coefficient, e.g., based on a weighted sum that includes previously reconstructed coefficients 216 of the reconstructed transform and the corresponding coefficient 309 of the residual transform) to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214 (e.g., based on as a sum of the predictor and the signaled coefficient value), reconstructing 228 the currently decoded block 140a by correcting 306 a sample predictor 308 of the currently decoded block 140a based on the residual transform 210 (e.g., based on a coefficient-wise sum, e.g., wherein a retransformation may be applied before orafterthe reconstruction, e.g., retransformation from frequency to time domain).

[0177] The decoder 12 may be configured so that the predetermined relationship is fulfilled if one or all of one or more out of the following conditions are true (e.g., at least one is true, at least two are true, e.g., all are true, e.g., only a specific condition is true): a) the selected prediction modes of the one or more blocks 304 of one or more reference channels 306 and the selected prediction mode of the currently decoded block 140a are a block matching mode according to which the sample predictor for a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs, displaced from the block to be predicted by a temporal offset, and the

[0178] FH250307PEP-2026097809fetemporal offset is equal among the one or more blocks 304 and the currently decoded block 140a (e.g., in fig. 3, a temporal offset At2 for the currently decoded block 140a and a temporal offset At1 for the block 304 may be equal, thusly fulfilling a corresponding predetermined relationship), b) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels (e.g., in fig. 3, currently decoded block 140a may reference a block in the same vertical column of blocks and block 304 may reference a block in the same vertical column of blocks, e.g., wherein optionally block 140a block 304 reference the same block or reference by a same channel difference), c) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are all of a prediction mode (e.g. possibly mutually different or mutually equal prediction modes) comprised by a set of prediction modes comprising, or consisting of, one or more of: (i) a zero prediction mode according to which a zero signal is used as the sample predictor for a block to be predicted, (ii) a DC prediction mode according to which the sample predictor for a block to be predicted a constant signal of a constant value with predicting the constant value based on a previously reconstructed portion of the digital time-varying signal, and (iii) a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels.

[0179] Criteria for predetermined relationship may comprise one or more of the use of the same prediction mode (e.g., DC-prediction, cross-channel prediction, inter-channel prediction), the use of one or more (e.g., all) same prediction parameters (e.g., same DC-value for DC-prediction, same reference channel for cross-channel prediction, e.g., same channel distance for cross-channel prediction, e.g., same temporal offset for inter-channel prediction).

[0180] The decoder 12 may be configured to perform the steps of reconstructing and correcting depending on whether the blocks 304 of all reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship.

[0181] The decoder 12 may be configured predict 220 the currently reconstructed coefficient 214 of the residual transform 210from the one or more corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels

[0182] FH250307PEP-2026097809feonly using corresponding coefficient 309 of the residual transform 310 of blocks 304 of the reference channels 306 that fulfill the predetermined relationship (e.g., only from reference channels, for which the same prediction mode and / or the same one or more prediction parameters have been used as for the first channel).

[0183] The decoder 12 may be configured to sequentially reconstruct 206coefficients 208 of the residual transform 210of the currently decoded block 140a in acoding order210 (e.g., scan order) from a transform coefficient of the residual transform 210 of the currently decoded block 140a associated with a highest frequency (e.g., right-most transform coefficient) to a transform coefficient of the residual transform 210 of the currently decoded block 140a associated with a lowest frequency (e.g., left-most transform coefficient, e.g., DC-coefficient) (e.g., wherein the coding order is reverse to a coefficient order in the residual transform).

[0184] Fig. 11 shows a flow diagram of a method 440 for decoding a digital time-varying signal 92 from a data stream 16. The method 440 may be performed by any decoder 12 disclosed herein.

[0185] The digital time-varying signal 92 comprises multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned.

[0186] The method 440 comprises, in step 442, reconstructing each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and, in step 444, correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block 140a of a first channel 302 depending on whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship with the selected prediction mode of the currently decoded block 140a, by decoding 204 the signaled residual transform 202 for the currently decoded block 140a from the data stream 16; sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently decoded block 140a by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more previously reconstructed coefficients 216 of the residual transform 210 and a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a

[0187] FH250307PEP-2026097809fesignaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214, reconstructing 228 the currently decoded block 140a by correcting 306 a sample predictor 308 of the currently decoded block 140a based on the residual transform 210.

[0188] For is provided is an encoder 10 for block-wise encoding a digital time-varying signal 92 into a data stream 16. The encoder 10 may be any encoder 10 disclosed herein and may comprise any decoder feature or decoder-pendant feature disclosed herein.

[0189] The digital time-varying signal 92 comprises multiple channels (e.g., M channels, e.g., with M = 2, e.g., M = 3, 4, 5, or larger), each partitioned into consecutive blocks (e.g., temporal blocks 140) which are mutually temporally aligned (e.g., having temporally aligned block starts and ends, e.g., having the same block size).

[0190] The encoder 10 is configured to encode each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes (e.g., comprising one or more of a DC description mode, a cross-channel prediction mode, an intra-channel copy mode) to obtain a sample predictor for the respective block, and encode a residual transform encoded for the respective block into the data stream for correcting the sample predictor to reconstruct the respective block, with encoding a currently encoded block 140a of a first channel 302 depending on (e.g. the coefficient wise prediction based on both, previous coefficients of the currently encoded block’s 140a residual transform as well as the one or more corresponding coefficients of one or more aligned blocks of reference channels might be a mode which is only signable or available for the currently encoded block 140a if the following is true; if not true, the coefficient wise prediction might be performed without the one or more corresponding coefficients of one or more aligned blocks of reference channels (e.g., as done with the encoder described above with reference to fig. 9), or without use of the coefficient wise prediction at all) whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently encoded block 140a, fulfill a predetermined relationship (e.g., having the same prediction mode, e.g., having a same temporal referencing offset, e.g., using one or more same prediction parameter within the same prediction mode, e.g., using the same prediction mode with the same prediction parameters) with the selected prediction mode of the currently encoded block 140a, by encoding 204 the signaled residual transform 202 for the currently encoded block 140a into the data stream 16 from which the currently encoded block 140a is reconstructable by;

[0191] FH250307PEP-2026097809fesequentially reconstructing 206 (e.g., coefficient-wise, e.g., in a coding order into a transform coefficient associated with a highest frequency to a transform coefficient associated with a lowest frequency) coefficients 208 of the residual transform 210 of the currently encoded block 140a by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more (e.g., two, three, or more, e.g., pre-determined number of, e.g., all) previously reconstructed coefficients 216 of the residual transform 210and a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 (e.g., using a rescaled or weighted version of the corresponding coefficient, e.g., based on a weighted sum that includes previously reconstructed coefficients 216 of the reconstructed transform and the corresponding coefficient 309 of the residual transform) to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222using a signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214 (e.g., based on as a sum of the predictor and the signaled coefficient value), reconstructing 228 the currently encoded block 140a by correcting 306 a sample predictor 308 of the currently encoded block 140a based on the residual transform 210 (e.g., based on a coefficient-wise sum, e.g., wherein a retransformation may be applied before or after the reconstruction, e.g., retransformation from frequency to time domain).

[0192] Fig. 12 shows a flow diagram of a method 450 for block-wise encoding a digital time-varying signal 92 into a data stream 16. The method 450 may be performed by any encoder 10 disclosed herein.

[0193] The digital time-varying signal 92 comprises multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned.

[0194] The method 450 comprises, in step 452, encoding each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block.

[0195] The method 450 comprises, in step 454, encoding a residual transform encoded for the respective block into the data stream for correcting the sample predictor to reconstruct the respective block, with encoding a currently encoded block 140a of a first channel 302 depending on whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently encoded block 140a, fulfill a predetermined relationship with the selected prediction mode of the currently

[0196] FH250307PEP-2026097809feencoded block 140a, by encoding 204 the signaled residual transform 202 for the currently encoded block 140a into the data stream 16 from which the currently encoded block 140a is reconstructable by; sequentially reconstructing 206 coefficients 208 of the residual transform 210 of the currently encoded block 140a by predicting 220 a currently reconstructed coefficient 214 of the residual transform 210 from one or more previously reconstructed coefficients 216 of the residual transform 210 and a corresponding coefficient 309 of the residual transform 310 of the one or more blocks 304 of the one or more reference channels 306 to obtain a predictor 222 for the currently reconstructed coefficient 214, and correcting 226 the predictor 222 using a signaled coefficient value 224 of the signaled residual transform 202, corresponding to the currently reconstructed coefficient 214, reconstructing 228 the currently encoded block 140a by correcting 306 a sample predictor 308 of the currently encoded block 140a based on the residual transform 210.

[0197] Further is provided a data stream 16 (e.g., stored on a computer-readable medium, e.g., non-transitory storage medium) encoded according to any encoding method (and / or encoder 10) disclosed herein (e.g., decodable by any decoding method and / or decoder 12 disclosed herein).

[0198] Further is provided a computer program product (e.g., computer program, e.g., stored on a computer-readable medium, e.g., non-transitory storage medium) for implementing any method disclosed herein (e.g., any encoding method and / or any decoding method disclosed herein) when being executed on a computer or signal processor.

[0199] In the following, various examples of the first, second, and third aspect are described. Any feature of the three aspects may be implemented in any example below and vice versa.

[0200] General description

[0201] The present disclosure deals with the coding of time varying signals (e.g., digital time-varying signal 92), e.g., biomedical waveform signals and with the coding of general waveform data. The setup of this application is, for example, as follows. A biomedical waveform (or any other time-varying signal) with possibly multiple channels (e.g., or only one channel) is to be coded. In this example, multiple channels are provided. However, a single channel is also possible. For the coding process (both encoding and decoding), the waveform is, for example, first partitioned into “full-channel-blocks”, where each “full-channel-block” consists (or comprise) of temporally collocated sample values for all channels (e.g., three channels

[0202] FH250307PEP-2026097809fe302, 306 with three temporally aligned blocks 140a, 304, or any other number of channels 302, 306), e.g., over a fixed time interval, see Figure 13a.

[0203] Fig. 13a shows a schematic view of a full channel block 340 with a plurality of channel 302, 306. A horizontal direction indicates a time dimension and a vertical direction indicates a channel direction.

[0204] The channel 302 comprises a block 140a (e.g., block B) and a channel 306 comprises a block 304 (e.g., block A). The full channel block 340 (e.g., channel group) may comprise all channels of the digital time-varying signal 92 or all channels of a channel group of the digital time-varying signal 92.

[0205] Fig. 13b shows a schematic view of a coefficient prediction using corresponding coefficients of other channels. In other words, shows an example of LMS-prediction.

[0206] A currently reconstructed coefficient 214 (e.g., k-th residual DCT coefficient of channel B) is predicted based on previously reconstructed coefficients 216 (e.g., previously reconstructed residual DCT coefficients of channel B, e.g., higher frequencies) and corresponding coefficient 309 of other channels (e.g., previously reconstructed residual DCT coefficients of channels A). In the example show in fig. 13b, two corresponding coefficients 309 (of two channels) and three previously reconstructed coefficients 216 are used for the prediction. However, any other number of corresponding coefficients 309 and previously reconstructed coefficients 216 may be used.

[0207] The individual components of a “full channel block” for a given channel shall be referred to as blocks (e.g., temporal blocks 140) from now on, see Fig. 13a and Fig. 13b. On each block, e.g., predictive transform coding is applied. This means that, e.g., first, a prediction signal out of already reconstructed samples is generated (e.g., in form of set of samples in the time domain). Here, it is possible, e.g., that the prediction is selected out of a plurality of supported prediction modes which may, for instance, contain block-matching, cross channel or a DC prediction mode, described below. However, the supported prediction modes may comprise less prediction modes (e.g., any combination of prediction modes disclosed herein) or more prediction modes (e.g., with any combination of prediction modes disclosed herein). The selection of the applicable prediction modes may, e.g., be either signaled in the bitstream (e.g., data stream 16) or be derived implicitly (e.g., in form of one or more of a default prediction mode, a prediction mode indirectly derivable from signaled information,

[0208] FH250307PEP-2026097809feand from previously reconstructed blocks). Additionally, e.g., a “zero prediction mode” is also supported, which means that the prediction can be skipped (e.g., a prediction signal of only zeroes). It might, e.g., be possible that a high-level control flag limits the number of applicable prediction modes. Second, a DCT-II transform (or any other transform) may be applied to the prediction residual. One core of the present disclosure is that now:

[0209] 1) A least mean squared error (LMS) prediction process (or any other coefficient prediction process), described below, is, for example, applied to the residual coefficients (e.g., coefficients 208) to obtain still further residual coefficients (e.g., in order to obtain predictor 222). Here, the LMS prediction process proceeds sample by sample (e.g., coefficientwise, e.g., from highest frequency to lowest frequency) in a backward scan order (e.g., reverse to a coding order), i.e., it starts with the highest frequency and moves down until it reaches some lowest frequency which might, for example, be the first AC frequency or which might be the DC frequency. For the LMS prediction process, the k-th residual coefficient (e.g., with k forming an index from a lowest to a highest frequency, e.g., currently reconstructed coefficient 214) is predicted out of the k+1st until the k+Lth (e.g., previously reconstructed coefficients 216), some prediction order L (e.g., all previously reconstructed coefficients of the same block 140a or an upper threshold of previously reconstructed coefficients of the same block 140a), reconstructed (e.g., quantized and inverse quantized) residual coefficients of the same block, and, optionally, additionally out of reconstructed resdiual coefficients of previously coded channels of the same block (e.g., according to the second or third aspect). Here, the prediction is a linear prediction (e.g., a weights sum) whose coefficients (e.g., filter coefficients 218 that form or allow obtaining weights) are not transmitted and, e.g., are also not fixed in advance but, e.g., are determined in a backward-adaptive manner by updating (e.g., after having formed the predictor 222) them with a gradient update and some fixed learning rate that, e.g., typically depends on the previous error energy (e.g., based on or defined by a difference between predicted and reconstructed coefficients). For this updating process, an initialization of the coefficients (e.g., filter coefficients 218) may need to be specified, which might, typically, be realized by just initializing the coefficients with 0 (or any other value).

[0210] 2) The difference between the k-th residual transform coefficient and the LMS-prediction value for this coefficient (e.g., difference between the reconstructed coefficient 214 and predictor 222) is, e.g., quantized. Here, e.g., either a scalar uniform reconstruction quan- tizier or a low-complexity vector quantizer like a dependent quantizer that combines two scalar quantiziers, both containing zero but one containing only even multiples of a

[0211] FH250307PEP-2026097809fequantization stepsize and the other only containing odd multiples of a quantization stepsize, and the selection of the quantizer being determined by a state machine that invokes previously quantized coefficients, is used. However, any other form of quantization (including no quantization) may be used.

[0212] 3) The resulting quantization index is (e.g., of the coefficient value 224, optionally without quantization), e.g., entropy coded using Context based Adaptive Binary Arithmetic Coding (CABAC) or any other coding scheme. Here, the context modeling has the property that the binarization contains, e.g., a significance flag for each quantization index and that this flag is, e.g., context coded where the context model selection invokes, e.g., absolute values of previously coded quantization indices of the same block, the latter being LMS-residual coefficients for higher order frequencies and / or the position k of the transform coefficient within the block.

[0213] Role of LMS prediction with respect to the selected block prediction modes

[0214] In multiple embodiments, the LMS prediction mode may be selectively / adaptively combined with the various aforementioned block-based prediction modes as follows (e.g., as described with reference to the second and third aspect).

[0215] First, it might be the case that the LMS prediction mode (e.g., prediction of a currently reconstructed coefficient 214 based on one or more previously reconstructed coefficients 216) is only selectively used, e.g., where the selection is signaled per block (e.g., per block 140a, 304), and thus, separately for each channel. Thus, samples of the same temporal portion of the waveform, i.e. samples of the “full channel block” that belong to channel A (e.g., channel 306), henceforth referred to as block A (e.g., block 304), might use the LMS prediction mode while samples belonging to this temporal portion but to channel B (e.g., channel 302), henceforth referred to as block B (e.g., block 140a), block B being coded after block A, might not use the LMS prediction mode, the applicability being indicated, e.g., by a flag that is signaled separately per channel (or any other syntax element, e.g., a syntax element for larger coding entities such as the full channel block).

[0216] For the second item, it is generally noted that invoking DCT-II coefficients of previous channels is reasonable to improve the prediction quality of the LMS prediction (e.g., predicting coefficients based on previously reconstructed coefficients of other channels) since one can often assume some correlation between the channels. However, this correlation can merely be assumed between the original signals (e.g., in a non-transform domain) -not necessarily

[0217] FH250307PEP-2026097809febetween the residuals (e.g., in a transform domain). Therefore, a design is required which predicts between the residuals only in those cases where the correlation between the original signals can be assume to carry over to a correlation between the residuals. It is noted that an alternative would be to not use the residual of a previous channel but rather to use the full reconstructed sample values. However, this would put a significant computational burden on the process: The LMS prediction is conducted, e.g., in the DCT-II domain in the present context. The encoder and decoder may have the reconstructed residuals of previous channels for the same “full channel block” at their disposal in the DCT-domain, since these residuals may be transform coded, but they may not have the reconstructed full signals (e.g., a corrected sample predictor 306, e.g., in form of a sum of a re-transformed residual transform 210 and a sample predictor 308, e.g., in the time domain) at their disposal in the DCT domain - thus, an extra forward transform step (e.g., into a frequency domain) for the reconstructed signals would be required for the case that one invokes the full reconstructed signals in the DCT domain for the LMS prediction process. However, it is argued that this yields a significant computational burden (an extra forward transform) which is undesirable in many application scenarios of the envisioned coding system, for example, in the case the codec is used to store biomedical waveform data on a wearable device, the latter having limited computational and power resources.

[0218] Staying in the residual domain (e.g., non-transform domain, e.g., frequency domain) for the aforementioned reasons, it is pointed out that invoking samples of previously reconstructed channels (e.g., perform coefficient prediction based on previously reconstructed coefficients of other channels) of the same “full-channel-block” means to predict the prediction residual of the current block B by invoking the prediction residual of the previously coded block A. Thus, since the prediction mode may be signaled per block in each channel and can thus vary between block A and block B, it is therefore desirable that the prediction residual on the current block B should be correlated to the prediction residual of block A, if the block-A-residual is used for the LMS prediction of the block B residual. This can be realized, e.g., by invoking selectively into the LMS prediction process, e.g., only those channels which have a block prediction residual characteristic that is correlated to the block prediction residual characteristics of the current block B. In other words, the block A residual should be selectively invoked in the LMS prediction of block B only under specific circumstances. Instead of forward signaling such a correlation, which would put a lot of burden on an encoder since the latter would have to suitably search over all channel combinations that might be used for the LMS prediction of block B and which would result in a significant signaling overhead to signal the specific channels that are required for the block B LMS-prediction,

[0219] FH250307PEP-2026097809fein this application, it is proposed to instead determine this correlation implicitly by simply making it dependent on the respective block-prediction modes used on the channels, e.g., by comparing the block-prediction mode of the current block B with the used block prediction mode of the previous block A.

[0220] It is noted that even though the LMS prediction works in a backward-adaptive way and thus determines the prediction coefficients based on the actual values of the residuals on block A and block B, if these residuals are not correlated, the LMS prediction might become unstable and the resulting LMS-residual energy might become higher than the original residual energy on residual block B, or, at least, higher than the LMS-residual energy that would have been obtained if the residual samples of block A had been left out of the LMS prediction process on block B, the latter being thus conducted only based on residual DCT coefficients of higher order of the block B itself, and, optionally, out of DCT residual coefficients at the same frequency locations of some other (corresponding coefficient 309, e.g., having the same coefficient index, e.g., having the same frequency; e.g., using only one coefficient of block A), temporally collocated and previously coded, blocks.

[0221] For example, the following specific solutions are proposed: If the channel B block is given, it uses the reconstructed transform residual coefficients of the previously coded temporally collocated samples of block A only if the prediction modes on the corresponding blocks A and B match each other (e.g., depending on whether the selected prediction modes of one or more blocks 304 of one or more reference channels 306, which are temporally aligned to the currently decoded block 140a, fulfill a predetermined relationship), e.g., in one of the following senses: 1) Both block A and block B use block matching prediction (e.g., a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs) as block-based prediction but with the same temporal offset (e.g., At2=At1, see fig. 3). 2) Both block A and block B use cross channel prediction as block-based prediction (e.g., cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels). 3) Both channel A and channel B use either no prediction, i.e. the zero prediction, the DC-prediction or the cross channel prediction.

[0222] Fig. 14 shows a schematic view of an example of channels A and B, which do not have the same temporal offset. In other words, fig. 14 shows an example of blockmatching, in which no cross channel LMS is performed.

[0223] FH250307PEP-2026097809feIn the example shown in fig. 14, blocks A and B are both predicted by blockmatching. Block A is predicted using a previously coded block 342 (e.g., blockA-Previous) of the same channel having a temporal offset A (e.g., At1) relative to block A. Block B is predicted using a previously coded block 344 (e.g., blockB-Previous) of the same channel having a temporal offset B (e.g., At21) relative to block B. The temporal offset A and temporal offset B are not equal (e.g., not equal within a tolerance range, e.g., of 5% of temporal offset B). In such a case, the residuals may not be correlated and the residual of block A is not used for block B-residual-LMS.

[0224] In other words, a selected prediction mode of a block 304 (e.g., block A) of a reference channel 306, which is temporally aligned to the currently decoded block 140a, does not fulfill a predetermined relationship (e.g., by not having a same temporal referencing offset) with the selected prediction mode of the currently decoded block 140a (e.g., block B), wherein the currently reconstructed coefficient 214 of the residual transform 210 is not predicted from a corresponding coefficient 309 of the residual transform 310 of the block 304 of the reference channels 306 (e.g., wherein, however, the currently reconstructed coefficient 214 may optionally be predicted based on previously reconstructed coefficients 216 of the currently decoded block 140a (e.g., block B).

[0225] Fig. 15 shows schematic view of an example of channels A and B, which do not have the same temporal offset. In other words, fig. 15 shows an example of blockmatching, in which cross channel LMS is performed.

[0226] Similar to the example shown in fig. 14, blocks A and B are predicted by block matching using temporal offset A and temporal offset B, respectively. However, in this case, the temporal offset A and temporal offset are the same (e.g., exactly the same, e.g., within a tolerance range).

[0227] In such a case, residuals may be correlated and residuals of block A may be used for block B residual LMS. In other words, a selected prediction mode of a block 304 (e.g., block A) of a reference channel 306, fulfills a predetermined relationship (e.g., by having a same temporal referencing offset) with the selected prediction mode of the currently decoded block 140a (e.g., block B), wherein the currently reconstructed coefficient 214 of the residual transform 210 is predicted from a corresponding coefficient 309 of the residual transform

[0228] FH250307PEP-2026097809fe310 of the block 304 of the reference channels 306 (e.g., wherein, the currently reconstructed coefficient 214 may optionally be additionally predicted based on previously reconstructed coefficient 216 of the currently decoded block 140a (e.g., block B)).

[0229] The reasoning for case 1) is that (e.g., one reason the same temporal offset may result in coefficient prediction using a corresponding coefficient from a different channel is that), even if the original samples of blocks A and B are assumed to be correlated (which is often the case for e.g. biomedical waveforms), this correlation might no longer exist for the prediction residuals on block A and B, e.g., if these residuals result from block-matching predictions of blocks, referred to as blockA-Previous and blockB-Previous, where blockA-Pre-vious and blockP-Previous are no longer temporally adjacent and are thus not correlated. This is outlined in Fig. 14 and Fig. 15.

[0230] Fig. 16 shows a schematic view of an example of a plurality of channels, for which crosschannel prediction is performed. The example shows LMS for cross channel prediction residuals.

[0231] Fig. 16 shows six channels with six corresponding temporally aligned blocks A, B, V, W, X, an Y. Block B is cross-channel predicted using the blocks V and W and block A is crosschannel predicted using the blocks X and Y. However, any other number of channels may be used for cross-channel prediction such as one, three, four, or more channels. Block A may be reconstructed by a linear combination (e.g., weighted sum; e.g., sample-wise) of blocks X and Y that forms a sample predictor, as well as residual samples (e.g., re-trans-formed from a residual transform). Similarly, block B may be reconstructed using a sample predictor (e.g., from a linear combination of blocks V and W) residual samples (e.g., retransformed from a residual transform). Since block A and B are (partially) predictable from other channels, it is not unlikely that a correlation between various channels is present in the plurality of channels. It is noted that in the example shown in fig. 16, block B does not use the same blocks for cross-channel prediction than block A and also does not use block A for cross-channel prediction. Still, considering that at least two blocks A and B appear to correlate with other channels (as they are cross-channel predicted) is an indicator that blocks A and B may also exhibit correlations. An even stronger correlation may be achieved if blocks A and B are cross-channel predicted from the same channel (e.g., in case blocks A and B were both cross-channel predicted from block W, one may likely find also a correlation between blocks A and B). Furthermore, another indicator for correlation may be that block B is cross-channel predicted from channel A. Therefore, in different variations, the

[0232] FH250307PEP-2026097809feselected prediction modes of one or more blocks of one or more reference channels fulfill a predetermined relationship with the selected prediction mode of the currently encoded block 140a, if the one block and currently encoded block 140a are both cross-channel predicted, if the one block and currently encoded block 140a are both cross-channel predicted based on the same (third) block, or if the currently encoded block 140a is cross-channel predicted based on the one block. Which variation is to be used for the predetermined relationship may, for example, be pre-determined or signalled in the data stream 16.

[0233] In case the predetermined relationship is fulfilled (e.g., in fig. 16 the predetermined relationship may be fulfilled by blocks A and B both being cross-channel predicted), the currently reconstructed coefficient 214 of the residual transform 210 (e.g., of block B) may be predicted from a corresponding coefficient 309 of the residual transform 310 of the block 304 (e.g., block A) of the reference channels 306 (e.g., wherein, the currently reconstructed coefficient 214 may optionally be additionally predicted based on previously reconstructed coefficient 216 of the currently decoded block 140a (e.g., block B)).

[0234] The reasoning for case 2) (e.g., the use of cross channel prediction may result in coefficient prediction using a corresponding coefficient from a different channel) is as follows: If block A uses cross channel prediction, this means, for example, that the prediction signal of block A is generated as a linear combination (e.g., additionally with an offset, but the latter only affecting the DC-component in the DCT-II domain) of reconstructed signals of (e.g., possibly two, e.g., or any other number of blocks) blocks block X and block Y that are temporally collocated both to block A and to block B, i.e. that belong to the same “full channel block”, but that have been coded before block A. The weights of this linear combination are determined in a process that is different to the LMS-weight derivation (e.g., one or more of the weights may be signalled in the data stream 16, weights may be derivable based on a channel distance, and weights may be pre-determined), see below, and which rather uses temporally previously coded and reconstructed sample values. In another embodiment, the weights of this linear combination might also be forward signaled. Thus, the reconstructed residual DCT-coefficients on block A maybe nothing than (e.g., determined based only on) a linear combination of the reconstructed DCT coefficients of the ‘full signals’ (meaning not the residuals but the original signals resp. reconstructed versions of them) in block A, block X and (optionally) block Y. Exactly the same reasoning holds for block B and the resulting residual, since it was assumed that cross channel prediction was also used on block B: The reconstructed residual DCT- coefficients on block B are nothing that a linear combination of the reconstructed DCT coefficients of the ‘full signals’ in block B and some blocks V and W

[0235] FH250307PEP-2026097809fe(or only one of them) that are temporally collocated and belong to previous channels. Thus, since it is generally assumed that the ‘full signals’ are (likely) correlated over the channels, it is reasonable to invoke the residual of block A for the LMS residual prediction on block B if both block A and block B were using cross channel prediction as block wise prediction, i.e. if the residuals on block A and block B result from cross channel prediction. This is exemplarily outlined in Fig. 16.

[0236] The reasoning for case 3) (e.g., block A and B either use no prediction, DC-prediction, or cross channel prediction) is basically the same as the reasoning for case 2) and case 3) can be seen as an extension of case 2), since the DC- and the zero prediction alter the ‘full’ signals only in the first component or not at all and thus, correlations that exist in the original signal domain also pertain in the residual domain if these types of prediction are used.

[0237] Description of block matching prediction mode

[0238] In the block matching prediction mode, the prediction signal may be generated by copying the sample values (e.g., in the time domain) of already reconstructed reference blocks of the same channel. The locations (e.g., information that allows deriving the location, e.g., a temporal offset, e.g., in units of samples or in a time duration) of these blocks are transmitted. Block matching prediction with a single hypothesis (e.g., predicting the current block from only one other block) and block matching prediction with two hypotheses (e.g., predicting the current block from two other blocks, e.g., which are linearly combined) are proposed.

[0239] Additionally, it is proposed that an offset can be derived at the decoder from adjacent already reconstructed sample values on a template of size tSize=16 (e.g., the decoder 12 may be configured to decode information indicative of a temporal offset and to copy already reconstructed samples within a template, e.g., the template having a size of 16 samples or any other size). A half-pel accurate representation of the location of the reference block (e.g., a temporal offset in units smaller than a sample length) and a smoothing prediction filter are also proposed.

[0240] The block matching prediction may be performed dependent on one or more syntax elements (e.g., one or more flags). For example, the block matching prediction mode is invoked if the value of the syntax element block_matching_or_cross_channel_pred_flag (e.g., a flag indicating whether the prediction mode is one of block matching of cross channel) is equal

[0241] FH250307PEP-2026097809fe1 and if the value of the syntax element cross_channel_pred_flag (e.g., a flag indicating that the prediction mode is cross channel prediction of the flag is one) is not equal to one. If bm_pred_mult_hyp_flag (e.g., a flag indicating whether cross channel prediction is performed using only one block or multiple blocks) is equal to 0, block matching prediction with a single hypothesis is used. If bm_pred_mult_hyp_flag is equal to 1, block matching prediction with two hypothesis is used.

[0242] The values of the block matching offsets may be coded predictively by invoking the corresponding values from previous blocks as a prediction. Here, the values of the current channel are invoked for prediction if the value of the syntax element bm_pred_off_pred_prev_ch_flag is equal to 0 and the values of the preceding channel are invoked if the value of the syntax element bm_pred_off_pred_prev_ch_flag is equal to 1.

[0243] Description of cross channel prediction mode

[0244] In the cross channel prediction process, the prediction signal is generated by a linear model using collocated reconstructed sample values from different channels. In other words, the sample predictor 308 may be derived from already decoded samples of temporally aligned block from other channels. The parameters (e.g., indicating one or more of: one or more channels to predict from and weights for a weighted combination) of the linear model are derived from already reconstructed samples on a left adjacent template of size tSize = 16 (e.g., a template of any size, e.g., with a length of a power of two).

[0245] The cross channel prediction mode may be invoked if the syntax elements block_match-ing_of_cross_channel_pred_flag and cross_channel_pred_flag are equal to 1. Cross channel prediction with a single hypothesis (cc_pred_mult_hyp_flag equal to zero) and cross channel prediction with two hypothesis (cc_pred_mult_hyp_flag equal to one) are supported.

[0246] For the cross channel mode, which is only applicable if m > 0 (e.g., m indicating a total number of channels or a number of channels of a channel group, e.g. , wherein a first channel is assigned the number m=0), a reference channel index m°efwith 0 < m°ef< m is transmitted (or more reference channel indices in case of more than one reference channel). In the case of multi-hypothesis cross channel prediction, also a second reference channel index mjefwith 0 < mjef< m is transmitted.

[0247] FH250307PEP-2026097809feIf cc_pred_mult_hyp_flag is equal to zero and m = 1, m°efis not signaled and inferred to be 0. Similarly, if cc_pred_mult_hyp_flag is not equal to zero and m = 2, m°efand mjefare not signaled and are inferred to be 0 and 1, respectively.

[0248] The values of mJrefmay be coded predictively by invoking the corresponding values from previous blocks as a prediction. The difference to this prediction may be determined by the syntax elements cc_pred_abs_chd_greaterO_flag[ j ], cc_pred_abs_chd_minus1[j] and cc_pred_chd_sign_flag[j] (or any other one or more syntax elements).

[0249] Description of DC prediction mode

[0250] For the proposed DC prediction mode (e.g., wherein the sample predictor is a block with a sequence of samples having the same value), the predictions signal may be defined as the mean value of the four preceding already reconstructed sample values (e.g., or any other numberof sample values, e.g., one, two, three, five, or more). For example, the last four (or any other number) decoded sample values of a block immediately preceding the currently coded block may be used to determine a common sample value (e.g., dcVal). The mean value may be further processed, e.g., rounded). Thus, one puts

[0251] 3

[0252] dcVal = rec[m] [sk— 4 + p] + 2) » 2

[0253]

[0254] p = 0

[0255] (e.g., with the first sample skof the block to be decoded, e.g., wherein rec[m] indicates already reconstructed samples of channel m) and sets

[0256] pred[j] = dcVal, 0 < j < lk.

[0257] The proposed DC prediction mode is used if the value of the syntax element block_match-ing_or_cross_channel_pred_flag is equal to zero and if the value of the syntax element block_pred_mode is equal to 0. The sample predictor may be defined by a length lk-1 and may have the value dcVal at every sample entry.

[0258] Description of the LMS prediction

[0259] The LMS prediction process in the DCT-ll-residual-domain basically works as follows. Let

[0260] FH250307PEP-2026097809fex[0],...,x[ / c], x[k + 1], ...,x[ / V — 1]

[0261] be the original residual DCT-II coefficients (e.g., available at the encoder 10 without any changes from quantization) in block B (e.g., in the transform domain, wherein the transform block has a length N).

[0262] Assume thatx[ / c] is to be predicted.

[0263] Let

[0264] x [k + 1], ... , x [ / V — 1]

[0265] be the already reconstructed residual DCT-II coefficients (e.g., previously reconstructed coefficients 216, e.g., wherein x [k + 1] and x[k + 1] may be or may almost be identical apart from deviations caused by quantization and dequantization) corresponding to the x[k + 1], ...,x[N - 1] (e.g., in case the scan order for prediction goes from right to left or high frequency to low frequency) and let

[0266] x [AiH / c], ...,x [AM] [ / C],

[0267] denote the reconstructed DCT-residual coefficients of previous channels Alt...,AMthat are to be invoked in the LMS prediction process.

[0268] Assume that weights v , ...,vLand w1;... wMare given (e.g., predetermined and / or updated by training). Then the LMS-prediction signal may be computed as:

[0269] ZL

[0270] Vi ■ x[k + i] + / Wj • x[Ay][k].

[0271]

[0272] 1=1

[0273] Here, L is the LMS order in the same-channel-direction (e.g., L may be a number or maximum number of preceding already reconstructed coefficients; e.g., wherein M is the number of channels to be used for coefficient prediction). As can be seen in the equation above, multiple previously reconstructed coefficients 216 may be used from the same block, whereas only one already reconstructed coefficient may be used per other channel (however, more than one coefficient per other channel is also possible). Pred[k] in the equation above may be the predictor 222. Next, the residual

[0274] res[ / c] = x[ / c] — pred[k]

[0275] is computed (e.g., based on or defines by correction 226) and quantized at the encoder (e.g., wherein x[k] may only be available for the encoder 10, but the encoder 10 may need to quantize and dequantize the resulting res[k] in order to obtain the same value that the decoder 12 may obtain). The reconstructed, i.e. ‘inversely’ quantized value

[0276] res[ / c]

[0277] FH250307PEP-2026097809feis available both to the encoder and to the decoder (e.g., wherein res[ / c] may be encoded into the data stream 16 and transmitted to the decoder 12). This gives the next reconstructed residual value

[0278] x [Zc] = pred[k] + res[ / c].

[0279] In the equation above, x [Zc] may be the currently reconstructed coefficient 214 and pred[k] may be the coefficient predicted by the decoder 12. Finally, an update of the weights v1;..., vLand w1;... wMis computed, e.g., by invoking the old values of the weights v1;..., vLand w1;... wM(or only one of these two set of weights in case the prediction is performed only with one set of weights), the value res [ / c] and the inputs x [k + 1], ... ,x [L] and x [4i][ / c], ...,x [ / lM][ / r] to proceed to predict the next coefficient x [k - 1] until all required coefficients are predicted. This update essentially corresponds to a gradient update of each weight with respect to minimizing the current error res[ / c]. The learning rate of this update can depend on the energy of previously and the current values of res.

[0280] Fig. 17 shows a schematic view of an encoder 10 and decoder 12 for coding a data stream 16. Any feature disclosed for the encoder 10 and decoder 12 with reference to fig. 17 may be used in any other encoder 10 and / or decoder 12 disclosed herein and vice versa.

[0281] The above description is extended in the following by the presentation of implementation examples. Before this, however, the description proceeds with a presentation of a possible framework or codec into which the embodiments described above as well as the examples described further below may be built into. Many details described in this framework are, however, optional when being combined with any of the above or subsequently described embodiments. To be more precise, the framework is described with respect to Fig. 17 which shows an encoder for encoding a multi-channel digital signal 14 into a datastream 16 as well as decoder 12 for decoding the multi-channel digital signal 14 from datastream 16. This description of Fig. 17 shall be seen as a presentation of new embodiments of the present application which result when combining any of the embodiments described above or any of the examples described subsequently or even any of the claimed subject matters is combined with the decoder 12 or encoder 10 of Fig. 17 either by adopting all details / functional-ities described with respect to Fig. 17 or with leaving-out some of the details / functionalities described with respect to Fig. 17. Sometimes such “optional” features of Fig. 17 are explicitly identified as being optional with respect to the combination of the previously and subsequently described embodiments, but the just-mentioned possible combinations of the pre-viously / subsequently explained embodiments with the description of Fig. 17 shall not be

[0282] FH250307PEP-2026097809ferestricted to the these explicitly identified variations of Fig. 17 in terms of leaving-out certain features.

[0283] In Fig. 17, the multi-channel digital signal 14 is illustrated by way of an array of samples with the samples being illustrated as small squares 18. Each line / row corresponds to a certain channel of the multi-channel digital signal 14. Each channel of signal 14 may have associated therewith a respective channel ID and Fig. 17 shows these channels as being ordered according to their channel ID along vertical axis 20 which, thus, corresponds to a “source” channel axis 20. The horizontal axis 22 corresponds to time so that samples 18 forming one column, or being horizontally aligned, are samples belonging to one common time instant. Such set / column of temporally co-located samples 18 is illustrated in Fig. 17 at 24.

[0284] Each channel, thus, forms a digital time-varying signal or time / amplitude or time-to-ampli-tude signal. The multi-channel digital signal m might have been obtained by at least one of Electrocardiography, Electroencephalography, Electromyography or seismic measurement. Differently speaking, the multi-channel digital signal might be a bio-physiological waveform data such as an electroencephalography (EEG) signal, an electrocardiogram (ECG), or an electromyography (EMG) signal, or seismic waveform data. However, each channel / signal might alternatively be another sort of waveform signal data such as scalar media data such as an audio signal and the signal 14 might be a multi-channel audio signal.

[0285] Fig. 17 illustrates the option according to which signal 14 is not coded directly, i.e. , in the original domain 26, but in a so-called “coded domain” 28 which might differ from the original domain 26 by one or more of 1) channel transformation, 2) channel permutation and 3) temporal mutual channel alignment. The channel transformation, if applied, transforms, per sample time instant, a set or column 24 of samples from domain 26 to domain 28. Thus, in domain 28, the sample pitch and the time axis is the same as in domain 26, but the meaning of the channels is different, i.e., the “source” channels of domain 26 become transformed channels in domain 28. Accordingly, the vertical axis in Fig. 17 for domain 28 is denoted as 32. Note that the channel transformation might leave the number of channels unchanged so that there is the same number of channels in domain 26 as well as domain 28, but different approaches are also possible. Generally, the channel transformation would aim at reducing redundancy and trying to condense the channels’ energy onto a fewer number of channels in domain 28. As said, the channel transformation is optional. Accordingly, in general terms, the channels in domain 28 are called “coded channels” in order to distinguish

[0286] FH250307PEP-2026097809fethem from the “original” or “source” channels of digital signal 14 in domain 26. The permutation is also optional and may be used in combination with, or without, the channel transformation. If used in combination with the channel transformation, the permutation may be performed prior to and / or or subsequent to the channel transformation in order to per-mute / sort the source channels prior to transformation and the coded channels subsequent to the channel transformation. The channel transformation might be a DCT, DST, FFT or any other transformation. The temporal mutual alignment is also optional and might be seen as a constant temporal alignment between the source channels or the coded channels.

[0287] The module in encoder 10 performing the one or more of channel transformation, channel permutation and temporal mutual alignment is indicated in Fig. 17 as block 34. Side information 36 might be used in order to signal information on one or more of the following: 1) The channel transformation used, 2) information on the permutation(s) among the source channels and / or coded channels and 3) information on the mutual temporal alignment / de-lays between the source channels or coded channels wherein the temporal mutual alignment might be restricted to full sample precision. A corresponding block 38 in decoder 12 performs the reverse step, i.e., performs one or more of: 1) a channel retransformation, 2) a re-permutation of the source channels and / or coded channels and 3) a temporal re-align-ment of the source channels or coded channels. Note, that if no channel transformation takes place, the coded channels are, in fact, equal to the source channels except for being temporally mutually aligned or being differently sorted due to permutation. Block 38 might be controlled by the before-mentioned side information 36.

[0288] Thus, the “actual coding” relates to the coded channels in domain 28. In the coded domain 28, the coded channels are depicted in Fig. 17 as lines or rows of samples 40, each extending along time axis 22, the coded channels being depicted one on top of the other along coded channel axis 32 - potentially ordered according to a coded channel ID they have associated therewith - so as to result into an array of samples 40. Again, although Fig. 17 depicts the case that the number of source channels equals the number of coded channels, the number might be different. Further, if channel transformation is used, while there is no longer a clear association between source channels on the one hand and coded channels on the other hand, the temporal association remains: For each temporally co-located samples 24, there is a corresponding temporally co-located set 42 of samples 40 of the coded channels, wherein the set 42 in domain 28 is a column and might be a set of horizontally

[0289] FH250307PEP-2026097809femutually offset samples in case of, and according to, the mutual temporal alignment, if applied. In case of Fig. 17, it has been assumed that no such temporal alignment took place so that both sets 42 and 24 are pure columns in the time / channel representation.

[0290] The actual coding is done in units of so-called temporal blocks 30. The term “block” or “temporal block” 30 is used so as to denote both a temporal portion of the multi-channel signal in domain 28, i.e. , the set of coded channels, as well as a temporal portion of a certain coded channel. That is, for each temporal block 30, each coded channel has a temporal block such as block 140 depicted for some temporal block 30c and same are mutually colocated. The coding is done sequentially along these blocks 140, by following a coding / de-coding order, which traverses the blocks 140 temporal block 30 by temporal block 30 with traversing temporally co-located blocks of the coded channels along a channel order corresponding to the order of the coded channels along axis 32. This coding / decoding order is illustrated in Fig. 17 at 60. That is, in case of temporal block 140 being the block currently to be coded / decoded, the previously decoded / encoded temporal blocks include all preceding temporal blocks of all coded channels as well as the temporally co-located temporal blocks of coded channels preceding the coded channel 92 of temporal block 140 in channel order. These previously coded / decoded temporal blocks and their samples are illustrated in Fig. 17 by way of shading. In this regard, note that in Fig. 17, merely one temporal block 140 has been illustrated explicitly in order to reduce the complexity of Fig. 17. Thus, in the specification herein, reference sign 140 is sometimes used to indicate the currently en-coded / decoded temporal block or to stand representatively for all temporal blocks. Further, as depicted in Fig. 17, the partitioning of signal 14 into temporal blocks 30 and 140, respectively, might be done in a manner so that these blocks 30 and 140, respectively, are nonoverlapping.

[0291] The actual coding in units of the temporal blocks 140 is performed predictively. That is, the encoder 10 comprises a block predictor 62 which predicts the samples of the currently coded temporal block 140, thereby yielding a prediction signal 64, and the prediction residual 66 formed by a subtraction between the actual sample values of temporal block 140 and the predicted samples of prediction signal 64 formed at a subtractor 68 is coded into the datastream 16 by residual coder 70. The residual coding in residual coder 70 may, or may not, involve a coding error by means of quantization. In any case, block predictor 62 uses the reconstructable version as being available by previously coded temporal blocks in order to obtain the prediction signal 64. This reconstructable version 72 might be derived at encoder 10 by means of a residual decoder 74 which reverses, potentially under coding loss,

[0292] FH250307PEP-2026097809fesuch as quantization, e.g. by means of dequantization, the residual signal 76 as coded into datastream 16, and an adder 78 which sums-up prediction signal 64 and the reconstructable residual signal 80 as obtained by residual decoder 74. To be more precise, let’s call the channel-individual temporal blocks 140 subblocks with temporally collocated subblocks of all channels forming a temporal block 30. Then, the prediction in module 62 or, to be more precise, the prediction at encoder and decoder, is performed in units of the subblocks 140, i.e. subblock wise. The encoder is free to choose different prediction modes for the subblocks within one block 30. As explained in more detail herein, within one block 30, one subblock 140 may be predicted based on one or more subblocks previously - according to the decoding order 60 - en / decoded within this block 30, while another subblock 140 within that block 30 might be coded / decoded based on the previously en / decoded subblock 140 of the same channel (but within the previous block 30). The transform residual en / decoding is then performed subblock wise by use of a one-dimensional transform signaled in the data stream as described hereinbelow.

[0293] The decoder 12 decodes the coded channels from data stream 16 in a corresponding manner, i.e., in units of the temporal blocks 30 or in temporal blocks 140, respectively, and using predictive decoding. To this end, the decoder 12 comprises a residual decoder 82, an adder 84 and a block predictor 86 which correspond to, and are mutually connected in the same manner as, elements 74, 78 and 62 of encoder 10. That is, the residual decoder 82 derives from the residual signal 76 in data stream 16 the reconstructable residual signal 80 for a currently decoded temporal block 140 which is then subject to addition with prediction signal 64 derived by block predictor 86 for temporal block 140 on the basis of the reconstructed version 72 of previously decoded temporal blocks at adder 84. The output of adder 84, thus, yields the reconstructed version 72 of the currently decoded temporal block 140 and becomes part of the pool of already decoded samples of previously decoded temporal blocks when the temporal blocks of the coded channels are, in this manner, traversed along cod-ing / decoding order 60 so as to reconstruct the coded channels in the coded domain 28.

[0294] Note that the above description concentrated on the so-called sample prediction where samples of a current block 140 are predicted based on reconstructed samples of one or more previously decoded blocks, but coding interdependencies, namely intra-channel and inter-channel coding dependencies may be exploited not only in terms of sample prediction, but also in terms of other coding tools involving, for instance, parameter prediction and / or context derivation.

[0295] FH250307PEP-2026097809feIn order to enable a high degree of random access capability, some of the temporal blocks 30 may be coded in a random access manner meaning that the coded channels therein are coded independent from previous temporal blocks 30. Imagine, for instance, that temporal blocks 30b and 30e are random access temporal blocks. Then, none of the temporal channel blocks 140 in temporal block 30b as well as 30e would depend on any preceding temporal block 140 and no coding dependency would cross these temporal blocks 30b and 30e, that is no temporal block 140 within any of temporal block 30b-30d would be coded depending on any block 140 temporally preceding temporal block 30b, and no temporal block 140 within any of temporal block 30e and following would be coded depending on any block 140 temporally preceding temporal block 30e.

[0296] Thus, in other words, coding dependencies are restricted so as to not reach-out beyond the border of a random access temporal block 30b and 30e towards any preceding temporal block 30. Such restriction might also hold for intermediate temporal blocks 30c to 30d between random access temporal blocks 30b and 30e in that same may not depend on any temporal block preceding the leading one among the random access temporal blocks 30b and 30e, here block 30b. Accordingly, leading temporal borders of the random access temporal blocks 30b and 30e are indicated by bold lines in Fig. 17. In a variant, the restriction is not valid for all en / decoding stages. For instance, while the grouping might hold true for prediction, but the residual en / decoding dependencies might cross borders between channel groups. It might be the case, for instance, that for the entropy coding and decoding, all channels are coded jointly, i.e. using a single arithmetic coding engine, butthat for the sake of prediction and reconstruction, the channels are grouped as described into independent groups such that, after entropy decoding, each such group can be reconstructed completely independently from each other group. This means that no prediction of sample values or any other information is supported between different channel groups.

[0297] Further, it might be that the coding of the coded channels also interrupts or restricts interchannel dependencies. For example, one or more of the coded channels might be coded as random access coded channels so that same do not use inter-channel dependencies, but merely intra-channel dependencies. The restriction of inter-channel coding dependencies might follow the channel order 32: that is, coding of these random access coded channels and the intermediate coded channels therebetween would be restricted so as to not reach-out beyond such a random access coded channel toward any coded channel preceding that random access coded channel in channel order along axis 32. Two such random access coded channels 88a and 88b and the resulting inter-channel dependency borders

[0298] FH250307PEP-2026097809feare illustrated in Fig. 17. Note that the restriction of inter-channel dependencies might be differently and is illustrated here merely as an example where the definition of, along channel order 32, interspersed random access channels 88a and 88b defines channel groups covering contiguous channels along the channel order 32. Other groups of channels might be defined, which do not necessarily follow the channel order 32, and inter-channel dependencies might be restricted not to render any channel of one group dependent on a channel of any other group, and within each group the inter-channel dependencies may also by restricted or each channel might by coded inter-channel dependent on any previously coded channel within its channel group.

[0299] The block predictor 62 and 86 of encoder 10 and decoder 12, respectively, operate synchronously, i.e. , they generate the same prediction signal 64 based on the previously en-coded / decoded samples of previously encoded / decoded temporal blocks 140. On encoder side 10, the prediction for a certain temporal block 140 may be accompanied or determined by one or more prediction parameters. Same might be determined on encoder side based on a rate / distortion optimization. These prediction parameters 90 are coded into data stream 16 and they are decoded from data stream 16 and used by block predictor 86 so as to perform the same prediction.

[0300] It might be that encoder 10 and decoder 12 support more than one prediction mode. For instance, encoder 10 and decoder 12 may support an intra prediction mode (which mode may also be called block-copy mode) according to which the currently encoded / decoded temporal block 140 is predicted based on the reconstructable sample values of previously encoded / decoded temporal blocks of the same coded channel to which the currently encoded / decoded temporal block 140 belongs, which is coded channel 92 in the example of Fig. 17. Additionally or alternatively, encoder 10 and decoder 12 may support an inter-prediction mode (which mode may also be called cross-channel prediction mode) according to which the currently encoded / decoded temporal block 140 is predicted based on the reconstructable sample values of previously encoded / decoded temporal blocks of one or more coded channels preceding - in coding order 32 - the coded channel 92 to which the currently encoded / decoded temporal block 140 belongs. Additionally or alternatively, there may be a mixed prediction mode according to which the prediction signal 64 is obtained by both, re-constructed / reconstructable sample values of previously encoded / decoded temporal blocks of coded channel 92 itself as well as reconstructed / reconstructable sample values of one or more coded channels preceding coded channel 92 in channel order along axis 32. Beyond this, there may be temporal blocks 140 which are coded without any prediction at

[0301] FH250307PEP-2026097809feencoder 10 and decoded without any prediction at decoder 12 such as the first temporal blocks 140 in the tiles 94 resulting from mutually separating the temporal blocks by means of the random access borders 96 on the one hand and the random access channel borders 98 on the other hand. This corresponds to the prediction signal 64 being set to zero and this may form an additional mode which could be called bypass mode. Additionally, or alternatively, there may be other modes such as ones deriving a DC predictor or linear function predictor for block 64 based on immediately preceding samples which immediately precede block 140. The prediction parameters 90 may, thus, contain for a currently encoded / de-coded temporal block 140 a prediction mode flag or prediction mode indicator indicating the prediction mode to be used for this currently encoded / decoded temporal block 140 and, optionally, one or more parameters parameterizing the prediction mode to be used for this currently encoded / decoded temporal block 140. It might also be that the prediction parameters are themselves coded predictively from already reconstructed blocks 140. In this prediction process, the laid out random-access capabilities in channel- and temporal-direction are, as an example, always maintained, i.e. the mentioned prediction of prediction parameters may never be supported across such a random access segment.

[0302] As mentioned, the aforementioned coding dependencies ought not to cross any of the borders 96 and 98 not only result from the just-described sample prediction capabilities of block predictor 62 and 86, respectively, but may optionally also result from other mechanisms such as parameter prediction according to which parameters such as the aforementioned prediction parameters 90 for a certain temporal block 140 are predicted based on coding parameters conveyed in the data stream 16 for any previous temporal block, or context derivation for context-adaptive entropy coding / decoding any coding parameter such as the prediction parameters 90 or any other side information such as side information 76 and 36 for temporal block 140 based on any coding parameter conveyed in the data stream 16 for any preceding temporal block.

[0303] That is, summarizing, the encoder 10 encodes the multi-channel signal 14 by transferring it into the coded domain 28 and then coding the coded channels into data stream 16 in the just-described block-wise and predictive manner, wherein decoder 12 decodes the coded channels of coded domain 28 from data stream 16 and the corresponding block-wise and predictive mannerwith then gaining the multi-channel signal 14 in its original form 26 based on the coded channels in coded domain 28 by means of segment 38. As said, the channel transformation is optional and if not used, each sample 40 in the coded domain 28 really corresponds to one sample 18 in the original domain 26. If, further, the temporal mutual

[0304] FH250307PEP-2026097809fealignment is not used, each sample 40 exactly corresponds to a sample 18 in the original domain 26 at exactly the same time instant or, differently speaking, all temporally co-located samples 40 in coded domain 28 remain mutually temporally co-located in the original domain 26.

[0305] It should be noted that the temporal blocks 30 might, other than illustrated in Fig. 17, vary in block length rather than being of a constant length as depicted in Fig. 17. For instance, encoder 10 may decide on the length of blocks 30 and signal the block length of blocks 30 (and the corresponding temporal blocks 140 of the coded channels) within data stream 16. Such signaling might be done on block level, such as for each temporal block 30 or, differently speaking for each temporally aligned bundle of blocks 140, so that the encoder may decide on the block size on the fly, or the block length might be signaled in the stream 16 on a larger scope such as for a sequence of blocks or even the whole stream 16.

[0306] As to the residual coder and residual decoder 70 and 82, they may use transform coding / de-coding in order to convey the residual signal 76 in data stream 16. That is, the residual signal 80 may be conveyed in data stream 16 in transform or spectral domain by way of transform coefficients in residual signal 76. The transform domain might be a DCT, DST or an FFT. The transform may be non-overlapping, i.e. it may only transform residual signal 80 and its re-transform may only cover residual signal 76 within block 140, and / or may be non-windowed, i.e. the residual signal might be transformed without any transform window used to temporally shape the residual signal 80 before the transform. The transform domain, i.e. the transformation leading from time domain to transform domain which is used by the encoder to transform the prediction residual signal 80 to be coded und the corresponding re-transformation leading from transform domain to time domain which is used by the decoder to derive the prediction residual signal 80, or the transformation, might be selected from a set of available transforms including, for instance, one or more of 1) one or more DCTs, 2) one or more DSTs and 3) an identity transform according to which the prediction residual signal 80 is coded into the data stream 14 in time domain directly. The transform may be critically sampled in that the number of transform coefficients resulting from the samples of one block 140 may equal the number of samples of block 140. Again, the samples might be the residual samples or may be, in case of the bypass mode, the channel samples directly.

[0307] The transform coefficients might be encoded by quantization, i.e. they may be quantized with the quantized coefficients then being coded in the datastream 16. Dequantization may

[0308] FH250307PEP-2026097809feoccur at decoding. For quantization, either a scalar uniform reconstruction quantizer or a low complexity vector quantizer might be used. In order to determine the quantization indices, the encoder may perform some optimization algorithm such as a rate-distortion optimized scalar quantization, or a trellis quantization with the goal to approximately minimize an approximated Lagrangian rate-distortion cost. At the decoder, the reconstruction process that yiels the transform coefficients may be conducted by multiplying the coded quantization indices with a certain step-size and, in case of the use of a low-complexity vector quantizer, by additionally invoking a state-machine based on the parity of previously decoded quantization indices in order to reconstruct the current quantization index.

[0309] In order to control the quantization noise, the transform coefficients might be subject to noise shaping. Spectral noise shaping may be used to shape the quantization noise spectrally. This may be done by signaling in the data stream spectral-band scale factors, i.e. a scale factor per spectral band, which represent a transfer function of a spectral filter which approximates the spectral envelope of the signal within the current block 140 (or its prediction residual, respectively), or signaling filter coefficients defining a temporal filter having a filter transfer function which approximates the spectral envelope of the signal within the current block 140 (or its prediction residual, respectively). On encoder side, spectral noise shaping may be applied in spectral domain by multiplying an inverse of scale factors, either directly signaled in the data stream or derivable from the filter coefficients by filter-to-factor conversion, with the transform coefficients before quantization. That is, at encoder, the coefficients are shaped by the inverse of the spectral envelope. At decoder side, spectral shaping may be applied in spectral domain by multiplying scale factors, either directly signaled in the data stream or derived from the filter coefficients by filter-to-factor conversion, with the transform coefficients, with then . That is, at decoder, the coefficients are shaped by the spectral envelope before applying retransformation. Additionally or alternatively, temporal noise shaping might be applied. To this end, TNS filter coefficients might be determined and signaled by the encoder. The TNS filter coefficients may represent a transfer function which approximates the temporal envelope of the current block 140 (or its residual signal). The encoder may apply TNS filtering using the filter coefficients by spectrally filtering the possibly spectrally shaped transform coefficients so as to filter them with a transfer function corresponding to an inverse of the temporal envelope. The TNS filter coefficients might be derived by linear prediction analysis of the possibly spectrally shaped transform coefficients so as to derive a linear prediction filter, then used as TNS filter, which minimizes a prediction residual when spectrally applied on the possibly spectrally shaped transform coefficients. At the encoder, the TNS filtered coefficients are then quantized and entropy

[0310] FH250307PEP-2026097809fecoded. At decoder side, the inverse takes place: the possibly spectrally shaped transform coefficients are inversely TNS filtered before applying retransformation. Additionally or alternatively, noise filling might be used. The filling may be applied to zero-quantized portions of the spectrum and controlled by the encoder via corresponding noise filling parameters.

[0311] As to the encoding / decoding the block or sequence of quantized transform coefficients of a current block into / from the data stream 16, arithmetic coding, such as context-adaptive binary arithmetic coding, CABAC, may be used. The CABAC encoding / decoding may by performed frame wise. That is, in each channel, the sequence of blocks 140 may be partitioned into immediately consecutive blocks 140, which form frames. This partitioning may be equal among the channels so that, again, a frame denotes both a temporal portion within each channel individually, as well as a temporal portion of the multi-channel signal, i.e. a collection of temporally aligned frames. Within each frame, the sequence of blocks 140 are CABAC en / decoded with once initializing the contexts and resetting the internal CABAC state at the beginning and then updating the contexts’ probabilities during en / decoding the respective frame. That is, blocks 140 are CABAC decodable merely in units of frames. The context initialization might be done independent from previous frames, or depending on the contexts as manifesting itself at the end of, of during, the en / decoding a previous frame.

[0312] Some deblocking processing might be used to avoid blocking artifacts. If, alternatively, an overlapped transform is used, an overlap-add processing with re-transforms of immediately preceding / succeeding temporal blocks of the same coded channel might be used in order to completely reconstruct the current temporal block’s 140 residual signal 76.

[0313] Besides such transform-(residual)-coded blocks there might be temporal blocks 140 which, additionally or alternatively, are coded using, besides the block prediction by block predictor 62 / 86 - which could be called a primary prediction - a secondary sample-wise prediction of the residual samples in residual block 66 such as by predicting a current sample’s residual sample by means of already decoded values of preceding - in sample coding order - residual samples in block 66 or 80, with then correcting same by means of a secondary-prediction-residual sample decoded from the data stream 16. The secondary-prediction-residual samples for such a block may coded into the data stream en block in a transform domain or sample-wise in time domain.

[0314] FH250307PEP-2026097809feNote that the afore-mentioned spectral shaping of the residual signal of a block 140 might be seen as a sample wise residual prediction, i.e. the case where filter coefficients are signaled for a block which define a temporal filter having a filter transfer function which approximates the spectral envelope of the residual signal within a current block 140. In sample wise residual prediction, the residual predictor on a current block 140 might either be chosen out of a fixed set of prediction modes, where an index to such a residual prediction mode is signaled in the bit-stream, or the residual prediction mode might be ‘signal adaptive’. In the latter case, prediction filter coefficients for the residual predictor are determined at the encoder by solving for example a linear equation, and are then quantized and transmitted to the decoder. At the decoder, the coefficients are inverse quantized and then the samplewise prediction is conducted with these coefficients. The number of used coefficients may vary per block and might also be signaled in the bit-stream. Additionally, it might optionally (i.e. indicated by some information in the bit-stream) be supported to invoke collocated samples from a previous block for the sample wise residual prediction. Finally, the coefficients of the sample wise residual prediction might be coded predictively, i.e. be predicted from used coefficients of a previous block, where only the differences to the current coefficients are transmitted.

[0315] A final note shall be made with respect to the juxtaposition of frames, blocks 140, channels and channel groups and regarding decoding order. The description above already described the fact that the channels might be grouped into channel group with each channel group being coded independently from each other, meaning that the blocks 140 in a certain channel group are coded without dependencies from channels outside their channel group. The decoding order 60, thus, would traverse the channels channel-group individually, channel group by channel group. Within each channel group, the blocks 140 are traversed as described: all temporally aligned blocks 140 of all channels fist, then proceeding with the next blocks 140 and so forth. A frame may have a sequence of blocks of a channel group encoded thereinto along the mentioned decoding order, such as n temporally consecutive blocks 140 for all channels of a channel group. IF the channel group had m channels, m*n block104 would, thus, be coded into the frame. As mentioned, there might be dependent frames, for which the CABAC contexts are adopted from the preceding frame of the same channel group, i.e. the one having encoded the immediately preceding block 140. For such dependent frames, not only CABAC contexts may be adopted from the preceding frame, but it may also be allowed to allow for prediction from the preceding frame to the dependent frame. Prediction, and possibly also any coding dependencies, towards channels outside the channel group and, within the channel group, towards frames temporally preceding the

[0316] FH250307PEP-2026097809femostly recently previously en / decoded independent frame would be disallowed. Thus, each tile shown in Fig. 17 by bold lines may represent a sequence of an independent frame flowed by zero, one or more dependent frames.

[0317] As mentioned before, Fig. 17 only represents a possible “framework” into which the previously described embodiments and the embodiments described subsequently may be built into. Many modifications may be performed with respect to Fig. 17, and some of these modifications might be mentioned in the subsequent description with respect to certain ones of the subsequently described embodiments, but these modifications shall then be treated as being also applicable with respect to other ones of the subsequently described embodiments.

[0318] The description is now resumed with respect to the announced subsequently described implementation examples and further embodiments where the digital time-varying signal is not restricted to be a channel of a multi-channel signal or to be a multi-channel signal, but where same may only be a single digital scalar signal.

[0319] Further Remarks:

[0320] Also, it should be noted that individual aspects described herein can be used individually or in combination. Thus, details can be added to each of said individual aspects without adding details to another one of said aspects.

[0321] Moreover, features and functionalities disclosed herein relating to a method can also be used in an apparatus (configured to perform such functionality). Furthermore, any features and functionalities disclosed herein with respect to an apparatus can also be used in a corresponding method. In other words, the methods disclosed herein can be supplemented by any of the features and functionalities described with respect to the apparatuses.

[0322] Also, any of the features and functionalities described herein can be implemented in hardware or in software, or using a combination of hardware and software, as will be described in the section “implementation alternatives”.

[0323] Implementation alternatives:

[0324] FH250307PEP-2026097809feDepending on certain implementation requirements, embodiments of the invention can be implemented in hardware or in software. The implementation can be performed using a digital storage medium, for example a floppy disk, a DVD, a Blu-Ray, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, having electronically readable control signals stored thereon, which cooperate (or are capable of cooperating) with a programmable computer system such that the respective method is performed. Therefore, the digital storage medium may be computer readable.

[0325] Some embodiments according to the invention comprise a data carrier having electronically readable control signals, which are capable of cooperating with a programmable computer system, such that one of the methods described herein is performed.

[0326] Generally, embodiments of the present invention can be implemented as a computer program product with a program code, the program code being operative for performing one of the methods when the computer program product runs on a computer. The program code may for example be stored on a machine readable carrier.

[0327] Other embodiments comprise the computer program for performing one of the methods described herein, stored on a machine readable carrier.

[0328] In other words, an embodiment of the inventive method is, therefore, a computer program having a program code for performing one of the methods described herein, when the computer program runs on a computer.

[0329] A further embodiment of the inventive methods is, therefore, a data carrier (or a digital storage medium, or a computer-readable medium) comprising, recorded thereon, the computer program for performing one of the methods described herein. The data carrier, the digital storage medium or the recorded medium are typically tangible and / or non-transitionary.

[0330] 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 the sequence of signals may for example be configured to be transferred via a data communication connection, for example via the Internet.

[0331] FH250307PEP-2026097809feA further embodiment comprises a processing means, for example a computer, or a programmable logic device, configured to or adapted to perform one of the methods described herein.

[0332] A further embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.

[0333] A further embodiment according to the invention comprises an apparatus or a system configured to transfer (for example, electronically or optically) a computer program for performing one of the methods described herein to a receiver. The receiver may, for example, be a computer, a mobile device, a memory device or the like. The apparatus or system may, for example, comprise a file server for transferring the computer program to the receiver.

[0334] In some embodiments, a programmable logic device (for example a field programmable gate array) may be used to perform some or all of the functionalities of the methods described herein. In some embodiments, a field programmable gate array may cooperate with a microprocessor in order to perform one of the methods described herein. Generally, the methods are preferably performed by any hardware apparatus.

[0335] The apparatus described herein may be implemented using a hardware apparatus, or using a computer, or using a combination of a hardware apparatus and a computer.

[0336] The apparatus described herein, or any components of the apparatus described herein, may be implemented at least partially in hardware and / or in software.

[0337] The methods described herein may be performed using a hardware apparatus, or using a computer, or using a combination of a hardware apparatus and a computer.

[0338] The methods described herein, or any components of the apparatus described herein, may be performed at least partially by hardware and / or by software.

[0339] The above described embodiments are merely illustrative for the principles of the present invention. It is understood that modifications and variations of the arrangements and the details described herein will be apparent to others skilled in the art. It is the intent, therefore,

[0340] FH250307PEP-2026097809feto be limited only by the scope of the impending patent claims and not by the specific details presented by way of description and explanation of the embodiments herein.

[0341] The description is now resumed with respect to the announced subsequently described embodiments / claims.

[0342] FH250307PEP-2026097809fe

Claims

67Claims1. Decoder (12) for block-wise decoding a digital time-varying signal (28; 92) from a data stream (16), configured todecode (204) a signaled residual transform (202) for a currently decoded block (140a) from the data stream (16);sequentially reconstruct (206) coefficients (208) of a residual transform (210) of the currently decoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) based on filter coefficients (218) to obtain a predictor (222) for the currently reconstructed coefficient (214), andcorrecting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstruct (228) the currently decoded block (140a) by correcting (306) a sample predictor (308) for the currently decoded block (140a) based on the residual transform (210),wherein the decoder is configured to, in sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a), adapt (230) the filter coefficients (218) based on1) a) the predictor (222) for the currently reconstructed coefficient (214), and the signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214), or b) the currently reconstructed coefficient (214), and2) the one or more previously reconstructed coefficients (216) of the residual transform (210)so as to be used for a subsequently reconstructed coefficient (232) of the residual transform (210).

2. Decoder (12) of claim 1,FH250307PEP-2026097809fe68wherein the digital time-varying signal (28; 92) comprises multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, from the data stream (16),wherein the decoder is configured to reconstruct each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream,wherein the currently decoded block (140a) is of a first channel (302) and the decoder is configured toreconstruct (228) the currently decoded block (140a) based on the residual transform (210) by correcting (306) a sample predictor (308) of the currently decoded block (140a) based on the residual transform (210), andin sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a),depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently decoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently decoded block (140a),predict (220) the currently reconstructed coefficient (214) of the residual transform (210) additionally from a corresponding coefficient (308) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306), andadapt (230) the filter coefficients (218) additionally based on the corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306).

3. Decoder (12) of claim 2, configured so that the predetermined relationship is fulfilled if one or all of one or more out of the following conditions are true:a) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are a block matching mode according to which the sample predictor forFH250307PEP-2026097809fe69a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs, displaced from the block to be predicted by a temporal offset, and the temporal offset is equal among the one or more blocks (304) and the currently decoded block (140a), b) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels, c) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are all of a prediction mode comprised by a set of prediction modes comprising, or consisting of, one or more of:a zero prediction mode according to which a zero signal is used as the sample predictor for a block to be predicted,a DC prediction mode according to which the sample predictor for a block to be predicted a constant signal of a constant value with predicting the constant value based on a previously reconstructed portion of the digital time-varying signal, a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels.

4. Decoder (12) according to claim 2 or 3,wherein criteria for predetermined relationship comprises one or more of the use of same prediction mode,the use of one or more same prediction parameters.

5. Decoder (12) according to any of claims 2 to 4,wherein the decoder is configured to perform the steps of predicting and adapting depending on whether the blocks (304) of all reference channels (306), which are temporally aligned to the currently decoded block (140a), fulfill a predetermined relationship.

6. Decoder (12) according to any of claims 2 to 4,wherein the decoder is configured to perform the steps of predicting and adapting onlyFH250307PEP-2026097809fe70using corresponding coefficient (309) of the residual transform (310) of blocks (304) of the reference channels (306) that fulfill the predetermined relationship.

7. Decoder (12) according to any of claims 1 to 6,wherein in predicting (220) the currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) based on filter coefficients (218) to obtain a predictor (222) for the currently reconstructed coefficient (214), the decoder is configured tosubject the one or more previously reconstructed coefficients (216) of the residual transform (210) to a weighted sum, wherein at least one filter coefficients form weights for one or more previously reconstructed coefficients (216).8 Decoder (12) according to any of claims 1 to 7,wherein in adapting (230) the filter coefficients (218), the decoder is configured to determine adapted filter coefficients so that a difference between the corrected predictor (222) and an updated predictor obtained from the one or more previously reconstructed coefficients (216) of the residual transform (210) based on the adapted filter coefficients (218) is smaller than a difference between the corrected predictor (222) and the predictor (222) obtained from the un-adapted filter coefficients (218).9 Decoder (12) according to any of claims 1 to 8,wherein the decoder is configured to sequentially reconstruct the coefficients (208) of a residual transform (210) in a coding order (210) from a transform coefficient of the residual transform (210) of the currently decoded block (140a) associated with a highest frequency to a transform coefficient of the residual transform (210) of the currently decoded block (140a) associated with a lowest frequency.

10. Method (400) for block-wise decoding a digital time-varying signal (28; 92) from a data stream (16), the method comprisingdecoding (204) a signaled residual transform (202) for a currently decoded block (140a) from the data stream (16);sequentially reconstructing (206) coefficients (208) of a residual transform (210) of the currently decoded block (140a) byFH250307PEP-2026097809fe71predicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) based on filter coefficients (218) to obtain a predictor (222) for the currently reconstructed coefficient (214), andcorrecting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstructing (228) the currently decoded block (140a) by correcting (306) a sample predictor (308) for the currently decoded block (140a) based on the residual transform (210),wherein the method comprises, in sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a), adapting (230) the filter coefficients (218) based on1) a) the predictor (222) for the currently reconstructed coefficient (214), and the signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214), or b) the currently reconstructed coefficient (214), and2) the one or more previously reconstructed coefficients (216) of the residual transform (210)so as to be used for a subsequently reconstructed coefficient (232) of the residual transform (210).

11. Encoder (10) for block-wise encoding a digital time-varying signal (28; 92) into a data stream (16), configured toencode (204) a signaled residual transform (202) for a currently encoded block (140a) into the data stream (16) so that coefficients (208) of a residual transform (210) of the currently encoded block (140a) is sequentially reconstructable (206) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) based on filter coefficients (218) to obtain a predictor (222) for the currently reconstructed coefficient (214), andFH250307PEP-2026097809fe72correcting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),wherein the currently encoded block (140a) is reconstructable by correcting (306) a sample predictor (308) for the currently encoded block (140a) based on the residual transform (210),wherein the sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a) comprises adapting (230) the filter coefficients (218) based on1) a) the predictor (222) for the currently reconstructed coefficient (214), and the signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214), or b) the currently reconstructed coefficient (214), and2) the one or more previously reconstructed coefficients (216) of the residual transform (210)so as to be used for a subsequently reconstructed coefficient (232) of the residual transform (210).

12. Encoder (10) of claim 11,wherein the digital time-varying signal (28; 92) comprises multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, into the data stream (16),wherein the encoder is configured to encode each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and encoding a residual transform for the respective block encoded into the data stream for correcting the sample predictor so as to reconstruct the respective block,wherein the currently encoded block (140a) is of a first channel (302) andthe currently encoded block (140a) is reconstructable (228) based on the residual transform (210) by correcting (306) a sample predictor (308) of the currently encoded block (140a) based on the residual transform (210), andFH250307PEP-2026097809fein sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a),depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently encoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently encoded block (140a),predicting (220) the currently reconstructed coefficient (214) of the residual transform (210) additionally from a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306), andadapting (230) the filter coefficients (218) additionally based on the corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306).

13. Encoder (10) of claim 12, configured so that the predetermined relationship is fulfilled if one or all of one or more out of the following conditions are true:a) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently encoded block (140a) are a block matching mode according to which the sample predictor for a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs, displaced from the block to be predicted by a temporal offset, and the temporal offset is equal among the one or more blocks (304) and the currently encoded block (140a), b) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently encoded block (140a) are a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels, c) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently encoded block (140a) are all of a prediction mode comprised by a set of prediction modes comprising, or consisting of, one or more of:a zero prediction mode according to which a zero signal is used as the sample predictor for a block to be predicted,FH250307PEP-2026097809fea DC prediction mode according to which the sample predictor for a block to be predicted a constant signal of a constant value with predicting the constant value based on a previously reconstructed portion of the digital time-varying signal, a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels.

14. Encoder (10) according to claim 12 or 13,wherein criteria for predetermined relationship comprises one or more of the use of same prediction mode,the use of one or more same prediction parameters.

15. Encoder (10) according to any of claims 12 to 14,wherein the encoder is configured to perform the steps of predicting and adapting depending on whether the blocks (304) of all reference channels (306), which are temporally aligned to the currently encoded block (140a), fulfill a predetermined relationship.

16. Encoder (10) according to any of claims 12 to 14,wherein the encoder is configured to perform the steps of predicting and adapting onlyusing corresponding coefficient (309) of the residual transform (310) of blocks (304) of the reference channels (306) that fulfill the predetermined relationship.

17. Encoder (10) according to any of claims 11 to 16,wherein in predicting (220) the currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) based on filter coefficients (218) to obtain a predictor (222) for the currently reconstructed coefficient (214), the encoder is configured tosubject the one or more previously reconstructed coefficients (216) of the residual transform (210) to a weighted sum, wherein at least one filter coefficients form weights for one or more previously reconstructed coefficients (216).

18. Encoder (10) according to any of claims 11 to 17,wherein in adapting (230) the filter coefficients (218), the encoder is configured to determine adapted filter coefficients so that a difference between the corrected predictor (222) and an updated predictor obtained from the one or more previously reconstructed coefficients (216)FH250307PEP-2026097809fe75of the residual transform (210) based on the adapted filter coefficients (218) is smaller than a difference between the corrected predictor (222) and the predictor (222) obtained from the un-adapted filter coefficients (218).

19. Encoder (10) according to any of claims 11 to 18,wherein the encoder is configured to sequentially reconstruct the coefficients (208) of a residual transform (210) in a coding order (210) from a transform coefficient of the residual transform (210) of the currently encoded block (140a) associated with a highest frequency to a transform coefficient of the residual transform (210) of the currently encoded block (140a) associated with a lowest frequency.

20. Method (410) for block-wise encoding a digital time-varying signal (28; 92) into a data stream (16), the method comprisingencoding (204) a signaled residual transform (202) for a currently encoded block (140a) into the data stream (16) so that coefficients (208) of a residual transform (210) of the currently encoded block (140a) is sequentially reconstructable (206) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) based on filter coefficients (218) to obtain a predictor (222) for the currently reconstructed coefficient (214), andcorrecting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),wherein the currently encoded block (140a) is reconstructable by correcting (306) a sample predictor (308) for the currently encoded block (140a) based on the residual transform (210),wherein the sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a) comprises adapting (230) the filter coefficients (218) based on1) a) the predictor (222) for the currently reconstructed coefficient (214), and the signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214), or b) the currently reconstructed coefficient (214), andFH250307PEP-2026097809fe762) the one or more previously reconstructed coefficients (216) of the residual transform (210)so as to be used for a subsequently reconstructed coefficient (232) of the residual transform (210).

21. Data stream having stored thereon a data stream (16) encoded according to claim 20.

22. Computer program product for implementing the method of claim 10 or 20 when being executed on a computer or signal processor.

23. Decoder (12) for block-wise decoding a digital time-varying signal (28; 92) comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, from a data stream (16), configured toreconstruct each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block (140a) of a first channel (302) bydecoding (204) the signaled residual transform (202) for the currently decoded block (140a) from the data stream (16);sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) to obtain a predictor (222) for the currently reconstructed coefficient (214), and correcting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstructing (228) the currently decoded block (140a) by correcting (306) a sample predictor (308) of the currently decoded block (140a) based on the residual transform (210),FH250307PEP-2026097809fe77wherein the decoder is configured to, in sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a),depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently decoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently decoded block (140a), predict (220) the currently reconstructed coefficient (214) of the residual transform (210) additionally from a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306).

24. Decoder (12) of claim 23, configured so that the predetermined relationship is fulfilled if one or all of one or more out of the following conditions are true:a) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are a block matching mode according to which the sample predictor for a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs, displaced from the block to be predicted by a temporal offset, and the temporal offset is equal among the one or more blocks (304) and the currently decoded block (140a), b) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels , c) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are all of a prediction mode comprised by a set of prediction modes comprising, or consisting of, one or more of:a zero prediction mode according to which a zero signal is used as the sample predictor for a block to be predicted,a DC prediction mode according to which the sample predictor for a block to be predicted a constant signal of a constant value with predicting the constant value based on a previously reconstructed portion of the digital time-varying signal,FH250307PEP-2026097809fe78a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels.

25. Decoder (12) according to 23 or 24,wherein criteria for predetermined relationship comprises one or more of the use of same prediction mode,the use of one or more same prediction parameters.

26. Decoder (12) according to any of 23 to 25,wherein the decoder is configured predict (220) the currently reconstructed coefficient (214) of the residual transform (210) additionally from a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306) depending on whether the blocks (304) of all reference channels (306), which are temporally aligned to the currently decoded block (140a), fulfill a predetermined relationship.

27. Decoder (12) according to any of 23 to 26,wherein the decoder is configured predict (220) the currently reconstructed coefficient (214) of the residual transform (210) additionally from a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels only using corresponding coefficient (309) of the residual transform (310) of blocks (304) of the reference channels (306) that fulfill the predetermined relationship.

28. Decoder (12) according to any of 23 to 27,wherein the decoder is configured to sequentially reconstruct (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a) in a coding order (210) from a transform coefficient of the residual transform (210) of the currently decoded block (140a) associated with a highest frequency to a transform coefficient of the residual transform (210) of the currently decoded block (140a) associated with a lowest frequency.

29. Method (420) for block-wise decoding a digital time-varying signal (28; 92) comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, from a data stream (16), the method comprisingFH250307PEP-2026097809fe79reconstructing each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block (140a) of a first channel (302) bydecoding (204) the signaled residual transform (202) for the currently decoded block (140a) from the data stream (16);sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) to obtain a predictor (222) for the currently reconstructed coefficient (214), and correcting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstructing (228) the currently decoded block (140a) by correcting (306) a sample predictor (308) of the currently decoded block (140a) based on the residual transform (210),wherein the decoder is configured to, in sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a),depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently decoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently decoded block (140a), predict (220) the currently reconstructed coefficient (214) of the residual transform (210) additionally from a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306).

30. Encoder (10) for block-wise encoding a digital time-varying signal (28; 92) comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, into a data stream (16), configured toFH250307PEP-2026097809fe80encode each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and encode a residual transform encoded for the respective block into the data stream for correcting the sample predictor to reconstruct the respective block, with encoding a currently encoded block (140a) of a first channel (302) byencoding (204) the signaled residual transform (202) for the currently encoded block (140a) into the data stream (16) from which the currently encoded block (140a) is reconstructable bysequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) to obtain a predictor (222) for the currently reconstructed coefficient (214), and correcting (226) the predictor (222) using an encoded coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstructing (228) the currently encoded block (140a) by correcting (306) a sample predictor (308) of the currently encoded block (140a) based on the residual transform (210),wherein, in sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a),depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently encoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently encoded block (140a), the currently reconstructed coefficient (214) of the residual transform (210) is to be predicted additionally from a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306).FH250307PEP-2026097809fe8131. Method (430) for block-wise encoding a digital time-varying signal (28; 92) comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, into a data stream (16), the method comprisingencoding each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and encode a residual transform encoded for the respective block into the data stream for correcting the sample predictor to reconstruct the respective block, with encoding a currently encoded block (140a) of a first channel (302) byencoding (204) the signaled residual transform (202) for the currently encoded block (140a) into the data stream (16) from which the currently encoded block (140a) is reconstructable bysequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) to obtain a predictor (222) for the currently reconstructed coefficient (214), and correcting (226) the predictor (222) using an encoded coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstructing (228) the currently encoded block (140a) by correcting (306) a sample predictor (308) of the currently encoded block (140a) based on the residual transform (210),wherein, in sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a),depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently encoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently encoded block (140a), the currently reconstructed coefficient (214) of the residual transform (210) is to be predicted additionally from a corresponding coefficient (309)FH250307PEP-2026097809fe82of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306).

32. Data stream having stored thereon a data stream encoded according to the method of claim 31.

33. Computer program product for implementing the method of claim 29 or 31 when being executed on a computer or signal processor.

34. Decoder (12) for block-wise decoding a digital time-varying signal (28; 92) comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, from a data stream (16), configured toreconstruct each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block (140a) of a first channel (302) depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently decoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently decoded block (140a), bydecoding (204) the signaled residual transform (202) for the currently decoded block (140a) from the data stream (16);sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) and a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306) to obtain a predictor (222) for the currently reconstructed coefficient (214), andcorrecting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),FH250307PEP-2026097809fe83reconstructing (228) the currently decoded block (140a) by correcting (306) a sample predictor (308) of the currently decoded block (140a) based on the residual transform (210).

35. Decoder (12) of claim 34, configured so that the predetermined relationship is fulfilled if one or all of one or more out of the following conditions are true:a) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are a block matching mode according to which the sample predictor for a block to be predicted is obtained by intra-channel copy from an already reconstructed portion of a channel to which the block to be predicted belongs, displaced from the block to be predicted by a temporal offset, and the temporal offset is equal among the one or more blocks (304) and the currently decoded block (140a), b) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels, c) the selected prediction modes of the one or more blocks (304) of one or more reference channels (306) and the selected prediction mode of the currently decoded block (140a) are all of a prediction mode comprised by a set of prediction modes comprising, or consisting of, one or more of:a zero prediction mode according to which a zero signal is used as the sample predictor for a block to be predicted,a DC prediction mode according to which the sample predictor for a block to be predicted a constant signal of a constant value with predicting the constant value based on a previously reconstructed portion of the digital time-varying signal, a cross-channel prediction mode according to which the sample predictor for a block to be predicted is obtained by cross-channel prediction from one or more temporally aligned blocks of one or more sample prediction channels.

36. Decoder (12) according to claim 34 or 35,wherein criteria for predetermined relationship comprises one or more of the use of same prediction mode,the use of one or more same prediction parameters.FH250307PEP-2026097809fe8437. Decoder (12) according to any of claims 34 to 36,wherein the decoder is configured to perform the steps of reconstructing and correctingdepending on whether the blocks (304) of all reference channels (306), which are temporally aligned to the currently decoded block (140a), fulfill a predetermined relationship.

38. Decoder (12) according to any of claims 34 to 37,wherein the decoder is configured predict (220) the currently reconstructed coefficient (214) of the residual transform (210) from the one or more corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels only using corresponding coefficient (309) of the residual transform (310) of blocks (304) of the reference channels (306) that fulfill the predetermined relationship.

39. Decoder (12) according to any of 34 to 38,wherein the decoder is configured to sequentially reconstruct (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a) in a coding order (210) from a transform coefficient of the residual transform (210) of the currently decoded block (140a) associated with a highest frequency to a transform coefficient of the residual transform (210) of the currently decoded block (140a) associated with a lowest frequency.

40. Method (440) for block-wise decoding a digital time-varying signal (28; 92) comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, from a data stream (16), the method comprisingreconstructing each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and correcting the sample predictor using a residual transform derived for the respective block from the data stream with reconstructing a currently decoded block (140a) of a first channel (302) depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently decoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently decoded block (140a), byFH250307PEP-2026097809fe85decoding (204) the signaled residual transform (202) for the currently decoded block (140a) from the data stream (16);sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently decoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) and a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306) to obtain a predictor (222) for the currently reconstructed coefficient (214), andcorrecting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstructing (228) the currently decoded block (140a) by correcting (306) a sample predictor (308) of the currently decoded block (140a) based on the residual transform (210).

41. Encoder (10) for block-wise encoding a digital time-varying signal (28; 92) comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, into a data stream (16), configured toencode each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and encode a residual transform encoded for the respective block into the data stream for correcting the sample predictor to reconstruct the respective block, with encoding a currently encoded block (140a) of a first channel (302) depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently encoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently encoded block (140a), byencoding (204) the signaled residual transform (202) for the currently encoded block (140a) into the data stream (16) from which the currently encoded block (140a) is reconstructable by;FH250307PEP-2026097809fe86sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transform (210) and a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306) to obtain a predictor (222) for the currently reconstructed coefficient (214), andcorrecting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstructing (228) the currently encoded block (140a) by correcting (306) a sample predictor (308) of the currently encoded block (140a) based on the residual transform (210).

42. Method (450) for block-wise encoding a digital time-varying signal (28; 92) comprising multiple channels, each partitioned into consecutive blocks which are mutually temporally aligned, into a data stream (16), the method comprisingencoding each block of each channel by predicting the respective block using a selected prediction mode selected for the respective block out of supported prediction modes to obtain a sample predictor for the respective block, and encode a residual transform encoded for the respective block into the data stream for correcting the sample predictor to reconstruct the respective block, with encoding a currently encoded block (140a) of a first channel (302) depending on whether the selected prediction modes of one or more blocks (304) of one or more reference channels (306), which are temporally aligned to the currently encoded block (140a), fulfill a predetermined relationship with the selected prediction mode of the currently encoded block (140a), byencoding (204) the signaled residual transform (202) for the currently encoded block (140a) into the data stream (16) from which the currently encoded block (140a) is reconstructable by;sequentially reconstructing (206) coefficients (208) of the residual transform (210) of the currently encoded block (140a) bypredicting (220) a currently reconstructed coefficient (214) of the residual transform (210) from one or more previously reconstructed coefficients (216) of the residual transformFH250307PEP-2026097809fe(210) and a corresponding coefficient (309) of the residual transform (310) of the one or more blocks (304) of the one or more reference channels (306) to obtain a predictor (222) for the currently reconstructed coefficient (214), andcorrecting (226) the predictor (222) using a signaled coefficient value (224) of the signaled residual transform (202), corresponding to the currently reconstructed coefficient (214),reconstructing (228) the currently encoded block (140a) by correcting (306) a sample predictor (308) of the currently encoded block (140a) based on the residual transform (210).

43. Data stream having stored thereon a data stream encoded according to the method of claim 42.

44. Computer program product for implementing the method of claim 40 or 42 when being executed on a computer or signal processor.FH250307PEP-2026097809fe

Images