Decoders, encoders, methods, data streams and computer programs using a parametrizable function

Abstract

Embodiments of the invention describe decoders for decoding a digital time-varying signal from a data stream using a parametrizable function. Further embodiments describe encoders for encoding a digital time-varying signal into a data stream using a parametrizable function. Further embodiments comprise methods, data streams and computer programs.

Description

[0001] Decoders, Encoders, Methods, Data Streams and Computer Programs Using A Parametrizable Function

[0002] Description

[0003] Technical Field

[0004] Embodiments comprise decoders, encoders, methods for decoding, methods for encoding, computer programs and data streams using a parametrizable function for coding a digital timevarying signal.

[0005] Embodiments of the invention may also be called “sample extrapolation encoding of a digital time-varying signal”.

[0006] Background of the Invention

[0007] Predictive block coding is a widely used concept in coding. Generally, a constant prediction or a linear prediction is applied to obtain values for a block that is to be predicted. These concepts, however, while being computationally efficient, may not depict a best prediction of values of the block that is to be predicted, leading to high residuals that need to be coded. More sophisticated predictive algorithms, on the other hand, can improve prediction accuracy but at the cost of increased computational complexity and a higher demand for side information, thereby reducing overall coding efficiency.

[0008] Therefore, a concept is desired which offers an improved compromise between computational costs and coding efficiency.

[0009] This is achieved by the subject-matter of the independent claims of the present application.

[0010] Further embodiments according to the invention are defined by the subject-matter of the dependent claims of the present application.

[0011] Summary of the Invention

[0012] In the following, embodiments and aspects of the invention will be described which can be used individually or in combination with any of the features and functionalities and details described herein.

[0013] PCT_FH250111PEP-2025353041.DOCX An embodiment according to a first aspect of the invention comprises a decoder for decoding a digital time-varying signal 92 (e.g. a one-dimensional digital time-varying signal; e.g. a biomedical waveform signal; e.g. a one-dimensional waveform signal; e.g. a channel out of a multi-channel signal) from a data stream 16, configured to decode the digital time-varying signal 92 from the data stream 16 in temporal blocks 140 (e.g. in blocks 140) by decoding each of temporal blocks of a predetermined type (e.g. those having the corresponding extrapolation mode signalled in the data stream via a corresponding mode syntax element 402) among the temporal blocks 140 of the digital time-varying signal 92 by extrapolation from a previously decoded portion 404 of the digital time-varying signal 92, which precedes a leading end 406 of the respective temporal block of the predetermined type, by predicting the respective temporal block of the predetermined type using a parametrizable function 408 (e.g. a line with a damped slope) which is parametrizable in terms of a leading-end slope 410 and a leading-end offset 412 at the leading end 406 of the respective temporal block of the predetermined type, and comprises a monotonic change in slope from the leading end 406 of the respective temporal block of the predetermined type towards a trailing end 414 of the respective temporal block of the predetermined type so as to start, from the leading end, in a manner changing towards zero slope (i.e. if the leading end slope is positive, the slope is getting smaller, and if the leading end slope is negative, the slope is getting larger, i.e. in both cases changing towards zero, wherein the slope may even pass the zero in order to change its sign), and by setting 416 the leading-end slope 410 and the leading-end offset 412 at the leading end 406 of the respective temporal block of the predetermined type based on the previously decoded portion 404; decoding a prediction residual 418 from the data stream 16 (e.g. from a transform 420 signalled in the data stream), and

[0014] PCT_FH250111PEP-2025353041.DOCX reconstructing 424 the respective temporal block of the predetermined type by correcting 422 (e.g. by addition) the parametrizable function 408 using the prediction residual 418.

[0015] Furthermore, an embodiment according to the first aspect of the invention comprises an encoder for encoding a digital time-varying signal 92 into a data stream 16, configured to encode the digital time-varying signal 92 into the data stream 16 in temporal blocks 140 by encoding each of temporal blocks of a predetermined type (e.g. those having the corresponding extrapolation mode signalled in the data stream via a corresponding mode syntax element 402) among the temporal blocks 140 of the digital time-varying signal 92 by extrapolation from a previously encoded portion 404 of the digital time-varying signal 92, which precedes a leading end 406 of the respective temporal block of the predetermined type, by predicting the respective temporal block of the predetermined type using a parametrizable function 408 which is parametrizable in terms of a leading-end slope 410 and a leading-end offset 412 at the leading end 406 of the respective temporal block of the predetermined type, and comprises a monotonic change in slope from the leading end 406 of the respective temporal block of the predetermined type towards a trailing end 414 of the respective temporal block of the predetermined type so as to start, from the leading end, in a manner changing towards zero slope (i.e. if the leading end slope is positive, the slope is getting smaller, and if the leading end slope is negative, the slope is getting larger, i.e. in both cases changing towards zero, wherein the slope may even pass the zero in order to change its sign), and by setting 416 the leading-end slope 410 and the leading-end offset 412 at the leading end 406 of the respective temporal block of the predetermined type based on the previously encoded portion 404; encoding a prediction residual 418 into the data stream 16 (e.g. from a transform 420 signalled in the data stream), and wherein the respective temporal block of the predetermined type is reconstructable 424 by correcting 422 (e.g. by addition) the parametrizable function 408 using the prediction residual 418.

[0016] The inventors have found that the use of a parametrizable function which comprises a monotonic change in slope, such that the slope of the parametrizable function gradually changes and may even change its sign, provides improved coding efficiency when predicting values of the respective temporal block. By varying the slope, the function can flexibly approximate both increasing and decreasing signal trends, thereby providing a smooth and continuous extrapolation from the previously decoded portion.

[0017] By monotonically changing the slope, the parametrizable function is capable of mimicking the natural temporal evolution typically exhibited by digital time-varying signals. Such signals often

[0018] PCT_FH250111PEP-2025353041.DOCX exhibit quasi-periodic or oscillatory behavior, for instance, as in electroencephalographic (EEG) signals, where variations correspond to neuronal activity. Consequently, embodiments allow to capture signal tendencies more faithfully, which can result in smaller prediction residuals that need to be coded. In other words, by combining the parametrizable function, or more precisely its values, with the decoded or encoded prediction residuals, the respective temporal block can be reconstructed with overall less bits. Since the parametrizable function already constitutes an efficient approximation of the underlying signal progression, the magnitude of the remaining prediction residuals is reduced. Consequently, fewer bits are required to encode or decode these residuals in the data stream, thereby improving overall coding efficiency.

[0019] Furthermore, by applying the above described concept selectively to temporal blocks of a predetermined type, a targeted pre-selection of suitable temporal blocks can be achieved. This enables the encoder and / or decoder to employ the proposed extrapolation technique using the parametrizable function specifically for those signal portions (or temporal block) where it yields a significant coding benefit.

[0020] In the following, references to specific types or configurations of encoders may be omitted for the sake of conciseness. However, it should be understood that the features and functionalities described herein are equally applicable to all encoders according to the embodiments of the present invention.

[0021] According to an embodiment of the invention, the decoder is configured so that the parametrizable function 408 is, except for the leading-end slope 410 and the leading-end offset 412, independent from the previously decoded portion 404 (e.g. 406).

[0022] According to an embodiment of the invention, the parametrizable function 408 is further parametrizable in terms of a rate of the monotonic change, and the decoder is configured to set the rate depending on a length of the respective temporal block of the predetermined type (e.g. in different words, the function 408 is defined by its leading end slope and leading end offset and, from that onwards, is stretched horizontally to fit into the block 140; see the part (... szExt - j) » logSzExt; note, however, that the slopeCurr in the formula above (e.g. and described herein) is NOT the local slope of function 408 (e.g. 408I), but rather defines the slope of a “pointer” from the leading-end offset at the leading end which points for the jthsample position from the leading end 406 to the function 408 at position j).

[0023] PCT_FH250111PEP-2025353041.DOCX By adjusting the rate of the monotonic change as a function of the temporal block length, the overall shape of the parametrizable function can be efficiently controlled. This can allow the extrapolation behavior to be optimally tuned to the length of the signal segment under consideration. For longer temporal blocks, for instance, a more gradual or, depending on the signal characteristics, a more pronounced rate of monotonic change may be advantageous. Consequently, by adapting the rate of the monotonic change, the parametrizable function can provide an improved approximation of the original signal progression, resulting in smaller prediction residuals and, therefore, enhanced coding efficiency.

[0024] According to an embodiment of the invention, the decoder is configured to parametrize the parametrizable function 408 depending on a length 428 of the respective temporal block of the predetermined type such that a trailing-end slope 431 of the parametrizable function 408 at the trailing end of the respective temporal block deviates, in magnitude, by less than 0.1 from zero slope, or, alternatively, deviates, in magnitude, by less than 0.1 from minus the leading-end slope 410. In the latter case, the parametrizable function 408 thereby exhibits an extremal point within the interior of the temporal block.

[0025] According to an embodiment of the invention, the decoder is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types (e.g. those having S=1 and those having S=2; e.g. according to a transform type syntax element 436 decoded from the data stream 16), and parametrize the parametrizable function depending on a length 428 of the respective temporal block of the predetermined type such that a trailing- end slope of the parametrizable function at the trailing end of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from zero slope if the respective temporal block of the predetermined type is classified into a first sub-type, and a trailing-end slope of the parametrizable function at the trailing end of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from minus the leading-end slope if the respective temporal block of the predetermined type is classified into a second sub-type.

[0026] According to an embodiment, the encoder is configured to classify the respective temporal block of the predetermined type into one of the set of sub-types depending on a sub-type syntax element signalled (e.g. and encoded thereinto by the encoder and decoded therefrom by the decoder) in the data stream for the respective temporal block of the predetermined type.

[0027] PCT_FH250111PEP-2025353041.DOCX This classification and adaptive parametrization can allow the decoder to select the more appropriate extrapolation behavior for a given temporal block, depending on the signal characteristics and block length. Consequently, the decoder can be able to apply the most suitable transform thereby improving prediction accuracy and reducing the magnitude of residuals.

[0028] According to an embodiment the decoder is configured to parametrize the parametrizable function depending on a length of the non-available portion such that an opposite-end slope of the parametrizable function (514) at the opposite end (516) of the non-available portion (504) (e.g. alternatively of the respective temporal block of the predetermined type) deviates, in magnitude, by less than 0.1 from the zero slope or a trailing-end slope (431) of the parametrizable function (408) at the opposite end (516) of the non-available portion (504) (e.g. alternatively of the respective temporal block of the predetermined type) deviates, in magnitude, by less than 0.1 from minus the end slope (410).

[0029] According to an embodiment of the invention, the decoder is configured to, in the setting the leading-end slope and leading-end offset at the leading end of the respective temporal block of the predetermined type based on the previously decoded portion, determine the leadingend offset using a sum of sample values of the previously decoded portion (e.g. p[ szArr - 4 ] + p[ szArr - 3 ] + p[ szArr - 2] + p[ szArr - 1 ] +2), and determine the leading-end slope using a sum of addends, wherein each addend depends on a difference between sample values of the previously decoded portion (e.g. weighted sum 15*( p[ szArr - 1 ] - p[ szArr -4 ] ) +5*( p[ szArr - 2 ] - p[ szArr - 3 ] )). Note, szArr may be an index of the leading-end of the respective temporal block and shall be larger than or equal to 4.

[0030] By deriving the leading-end offset and slope from a weighted combination of previously decoded sample values, the decoder can generate a more accurate initial approximation of the temporal block

[0031] According to an embodiment of the invention, the parametrizable function includes a sum of an offset term (e.g. ( offset«( log2SzExt + log2(S) ) ) (e.g. this term may define the offset at the leading-end of a pointer 450 pointing to a function value 452 of the function 408 for same position j within the block 140) and a linear sample position dependent term (e.g. slopeCurr*stepCurr with slopeCurr = slope * ( S * szExt - j ) and stepCurr = 25 + j*10; ) (e.g. this term may define the pointer slope 454 of the pointer 450 which varies according to position j in that the pointer “tilts” for increasing distance from the leading end towards zero slope), wherein the decoder is configured to compute the offset term depending on the leading-end

[0032] PCT_FH250111PEP-2025353041.DOCX offset (e.g. offset), compute the linear sample position dependent term based on a product comprising a first factor term corresponding to a product (e.g. slope * ( S * szExt - j ) ) between the leading-end slope (e.g. slope) on the one hand and a difference between a length term (e.g. S * szExt) which linearly depends on a length (e.g. szExt) of the respective temporal block of the predetermined type and a sample position value (e.g. j ) measuring a sample position between the leading end and the trailing end from the leading end (e.g. as an offset from the lead end) on the other hand, and a second factor (e.g. 25 + j*10) related to the sample position value according to an affine mapping.

[0033] By combining an offset term with a linear sample-position dependent term that adapts according to both block length and sample position, the parametrizable function can flexibly approximate the shape of the temporal block, improving prediction accuracy and reducing residuals for more efficient coding

[0034] According to an embodiment of the invention, the decoder is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types (e.g. those having S=1 and those having S=2), and wherein the length term is a product of the length (e.g. szExt) of the respective temporal block of the predetermined type and a damping rate parameter (e.g. S) of the parametrizable function, and parametrize the parametrizable function depending on a length 428 of the respective temporal block of the predetermined type such that the damping rate parameter is larger (e.g. so that, when j assumes the maximum value corresponding to the trailing end, the difference remains larger; in case of S=1 , the pointer gets zero sloped at the trailing end) if the respective temporal block of the predetermined type is classified into a first sub-type, than if the respective temporal block of the predetermined type is classified into a second sub-type.

[0035] Such embodiments can enable the parametrizable function to more accurately capture the underlying signal dynamics of each temporal block, thereby reducing prediction residuals and enhancing overall coding efficiency.

[0036] According to an embodiment of the invention, the decoder is configured to set the damping rate parameter to 2 if the respective temporal block of the predetermined type is classified into a first sub-type (e.g. configured to set the damping rate parameter to if the respective temporal block of the predetermined type is classified into a first sub-type), and 1 if the respective temporal block of the predetermined type is classified into a second sub-type.

[0037] PCT_FH250111PEP-2025353041.DOCX According to an embodiment of the invention, the parametrizable function has the sum of an offset term and the linear sample position dependent term as a nominator and as a denominator a product comprising a first factor corresponding to the length (e.g. szExt) of the respective temporal block of the predetermined type, and / or a second factor corresponding to the damping rate parameter (e.g. realized by the sum of the corresponding logarithms thereof and the right shift using the sum).

[0038] According to an embodiment of the invention, the decoder is configured to decode each of temporal blocks of a further predetermined type among the temporal blocks 140 of the digital time-varying signal 92 by a sample-wise prediction and a sample wise prediction correction using a prediction residual signal for the respective temporal blocks of the further predetermined type with using the parametrizable function for sample-wise prediction parametrized for a 1-sample block-length.

[0039] Performing sample-wise prediction and correction using the parametrizable function configured for a 1-sample block length enables highly accurate reconstruction of the digital time-varying signal, reduces prediction residuals, and thereby improves coding efficiency.

[0040] According to an embodiment of the invention, the temporal blocks 140 of the digital timevarying signal 92 are non-overlapping and the decoder is configured to decode the prediction residual from the data stream by decoding a transform 420 of the prediction residual 418 from the data stream and applying a re-transformation (e.g. involving basis functions coinciding in length with a length of the temporal blocks of the predetermined type; e.g. a sample rate of the digital time-varying signal 92 is above, or equal to, a Nyquist rate of the transform domain; a number of coefficients of the transform coincides with a number of samples of the respective temporal block) onto the transform.

[0041] In addition to the non-overlapping block-based residual decoding, the use of transform-based re-transformation can enable precise frequency-domain modeling of the residual signal within each temporal block. This approach can allow for an exact reconstruction of the residual signal without aliasing, as the sampling rate of the digital time-varying signal is maintained at or above the Nyquist rate for the transform domain.

[0042] According to an embodiment of the invention, the decoder is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types (e.g. those having S=1 and those having S=2), and parametrize the parametrizable function depending on a length of the respective temporal block of the predetermined type such that a trailing-end

[0043] PCT_FH250111PEP-2025353041.DOCX slope of the parametrizable function at the trailing end of the respective temporal block of the predetermined type deviates (e.g. is deviates), in magnitude, by less than 0.1 from zero slope if the respective temporal block of the predetermined type is classified into a first sub-type, and a trailing-end slope of the parametrizable function at the trailing end of the respective temporal block of the predetermined type deviates (e.g. is deviates), in magnitude, by less than 0.1 from minus the leading-end slope if the respective temporal block of the predetermined type is classified into a second sub-type, and wherein the decoder is configured to decode the prediction residual from the data stream by decoding a transform of the prediction residual from the data stream and applying a first re-transformation (e.g. DST-4 or DST-7) involving basis functions having a magnitude increasing towards the trailing end onto the transform if the respective temporal block of the predetermined type is classified into the first sub-type, and a second re-transformation (e.g. DCT such as DCT-2) involving basis functions having constant magnitude if the respective temporal block of the predetermined type is classified into the second sub-type.

[0044] Building upon sub-type classification and adaptive parametrization, embodiments can further improve reconstruction by selecting re-transformation strategies that match the signal characteristics of each block. By integrating sub-type dependent parametrization with tailored transform re-transformation, embodiment can enhance coding efficiency.

[0045] According to an embodiment of the invention, the decoder selects, for each of the temporal blocks, a prediction mode out of a set of prediction modes with the predetermined type representing a predetermined prediction mode out of the set of prediction modes, wherein the set of prediction modes further comprises a DC prediction mode according to which a prediction signal (e.g. prediction residual) of the respective temporal block is determined to be (and / or, for example, determined by) a constant function with a determination of a constant of the constant function based on predetermined already decoded samples preceding the respective temporal block, a block-copy prediction mode according to which the prediction signal (e.g. prediction residual) of the respective temporal block is predicted based on one or more block reference block portions of already decoded samples preceding the respective transform-coded temporal offset relative to the respective transform-coded temporal block at a position signalled for the respective temporal block in the data stream, a cross-channel prediction mode according to which the prediction signal (e.g. prediction residual) of the respective temporal block is predicted based on one or more reference coded channels out of coded channels which represent a multi-channel signal 14 coded into the data stream and to be decoded from the data stream by the decoder, and one of which is represented by the

[0046] PCT_FH250111PEP-2025353041.DOCX digital time-varying signal 92, and a bypass prediction mode according to which the prediction signal (e.g. prediction residual) of the respective temporal block is set to zero.

[0047] According to an embodiment of the invention, the digital time-varying signal 92 is obtained by at least one of Electrocardiography, Electroencephalography, Electromyography or seismic measurement, and / or wherein the digital time-varying signal 92 is a bio-physiological waveform data such as an electroencephalography (EEG) signal, an electrocardiogram (ECG), or an electromyography (EMG) signal, or is a seismic waveform signal.

[0048] The inventors have found, that the above-described digital time-varying signal types are especially suitable to be coded with an embodiment of the invention.

[0049] According to an embodiment of the invention, the decoder is configured to predict each of temporal blocks of an even further predetermined type among the temporal blocks 140 of the digital time-varying signal 92 by block prediction (e.g. block-copy operation) from a reference portion of the digital time-varying signal 92 by padding a non-available portion of the reference portion abutting an end of the reference portion using the parametrizable function.

[0050] The inventors have found that the principle of using a parametrizable function for extrapolation, previously applied for residual coding, can be extended to signal padding. By employing this approach, the decoder can generate a realistic extension of the reference portion without introducing abrupt discontinuities or high-frequency artifacts. This not only facilitates accurate reconstruction of blocks where complete reference data is not available, but also improves the effectiveness of subsequent prediction or coding operations, reduces residual energy, and maintains high-fidelity representation of the original digital time-varying signal. Such a strategy is particularly beneficial for signals with periodic, quasi-periodic, or smoothly varying components, including bio-physiological waveforms and seismic data, where preserving signal continuity is critical for both analysis and compression efficiency.

[0051] The description now proceeds with embodiments according to the second aspect of the invention. It should be noted, however, that the features and functionalities described with respect to the parametrizable function in embodiments related to the first aspect of the invention may also be applied to a parametrizable function in an embodiment according to the second aspect of the invention, and vice versa. Consequently, the technical effects, advantages, and implementation options described for one embodiment are equally relevant and combinable with the other regards to a corresponding embodiment of the other aspect.

[0052] PCT_FH250111PEP-2025353041.DOCX An embodiment according to the second aspect of the invention comprises a decoder for decoding a digital time-varying signal 92 from a data stream 16, configured to decode the digital time-varying signal 92 from the data stream 16 in temporal blocks 140 by predicting each of temporal blocks of a predetermined type among the temporal blocks 140 of the digital time-varying signal 92 by block prediction (e.g. such as a block-copy operation) 500 from a reference portion 502502 of the digital time-varying signal 92 by padding a non-available portion 504 (e.g. due to interpolation, i.e. a sub-sample offset which is, for example, signalled for the block 140 in the data stream and locates the portion 502 relative to the block 140, portion may be larger / longer than the block to be predicted, namely by the filter kernel size of the interpolation filter and non-availability reasons may include restricted access to the past so that the part of portion 502 needed for the interpolation filter reaches out to the left beyond such restriction, or a part of portion 502 reaches out to far towards the future in a reference channel, i.e. to the right) of the reference portion 502 abutting an end 506 (e.g. leading- end / leading end or trailing-end / trailing end; e.g. in Fig. 5, it is the leading end) of an available portion 508 of the reference portion 502 using a parametrizable function 514 which is parametrizable in terms of an end slope 510 and an end offset 512 at an abutting end of the non-available portion abutting the end 506 of the available portion 508, and comprises a monotonic change in slope from the abutting end of the available portion of the reference portion towards an opposite end 516 of the non-available portion 504 so as to start, from the abutting end, in a manner changing towards zero slope, and by setting the end slope and end offset at the abutting end (e.g. end) of the non-available portion based on the available portion of the reference portion.

[0053] Furthermore, an embodiment according to second aspect of the invention comprises an encoder for encoding a digital time-varying signal 92 into a data stream 16, configured to encode the digital time-varying signal 92 into the data stream 16 in temporal blocks 140 by predicting each of temporal blocks of a predetermined type among the temporal blocks 140 of the digital time-varying signal 92 by block prediction (such as a block-copy operation) 500 from a reference portion 502 of the digital time-varying signal 92 by padding a non-available portion 504 (e.g. due to interpolation, i.e. a sub-sample offset which is, for example, signalled for the block 140 in the data stream and locates the portion 502 relative to the block 140, portion may be larger / longer than the block to be predicted, namely by the filter kernel size of the interpolation filter and non-availability reasons may include restricted access to the past so that the part of portion 502 needed for the interpolation filter reaches out to the left beyond such restriction, or a part of portion 502 reaches out to far towards the future in a reference channel, i.e. to the right) of the reference portion 502 abutting an end 506 (e.g. leading end or trailing end; in Fig. 5, it is the leading end) of an available portion 508 of the reference portion

[0054] PCT_FH250111PEP-2025353041.DOCX 502 using a parametrizable function 514 which is parametrizable in terms of an end slope 510 and an end offset 512 at an abutting end of the non-available portion abutting the end 506 of the available portion 508, and comprises a monotonic change in slope from the abutting end of the available portion of the reference portion towards an opposite end 516 of the non- available portion 504 so as to start, from the abutting end, in a manner changing towards zero slope, and by setting the end slope and end offset at the abutting end (e.g. end) of the non- available portion based on the available portion of the reference portion.

[0055] Embodiments according to the second aspect of the invention enable a padding process for digital time-varying signals by employing the parametrizable function for extrapolating a non- available portion of a reference portion of the signal. In such embodiments, the decoder (and analogously the encoder) performs block prediction, such as a block-copy operation, in which a reference portion is used to predict a current temporal block. When part of the reference portion required for prediction is unavailable, e.g., due to interpolation at a sub-sample offset, limited access to past samples, or reference data extending beyond accessible temporal boundaries, the non-available portion abutting the available portion of the reference data is reconstructed using the parametrizable function.

[0056] The parametrizable function allows a smooth and continuous extrapolation of the signal into the non-available region. This results in a reference portion that is both temporally consistent and physically plausible, thereby preventing discontinuities and, for example, enabling an efficient block copy operation.

[0057] Moreover, the parametrizable function according to embodiments of the second aspect of the invention may be identical, analogous or different to the parametrizable function described in connection with embodiments of the first aspect of the invention. Accordingly, features and functionalities described with respect to the parametrizable function of the first aspect may also be applied to the parametrizable function of the second aspect, and vice versa.

[0058] The end of the available portion may be understood synonymously with the abutting end of the abutting end of the non-available portion, or rather the end the abutting end indicate a same location that exhibits a border between the available and non-available portion. Therefore, both may be referenced with the same reference sign.

[0059] In the following description, references to specific types or configurations of some encoders may be omitted for the sake of conciseness. However, it should be understood that the

[0060] PCT_FH250111PEP-2025353041.DOCX features and functionalities described herein are equally applicable to all encoders according to the embodiments of the present invention and vice versa.

[0061] According to an embodiment of the invention, the decoder is configured to parametrize the parametrizable function depending on a length of the non-available portion such that an opposite-end slope of the parametrizable function 514 at the opposite end of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from the zero slope or a trailing-end slope 431 of the parametrizable function 408 at the opposite end of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from minus the end slope 410 (e.g. thereby comprising a extremal point in the inner of the non-available portion).

[0062] According to an embodiment of the invention, the decoder is configured to, in the setting the end slope and end offset at the abutting end of the non-available portion based on the available portion, determine the end offset using a sum of sample values of the available portion (e.g. p[ szArr - 4 ] + p[ szArr - 3 ] + p[ szArr - 2] + p[ szArr - 1 ] +2), and determine the end slope using a sum of addends, wherein each addend depends on a difference between sample values of the available portion (e.g. weighted sum 15*( p[ szArr - 1 ] - p[ szArr -4 ] ) +5*( p[ szArr - 2 ] - p[ szArr - 3 ] )).

[0063] According to an embodiment of the invention, the parametrizable function includes a sum of an offset term (e.g. ( offset«( log2SzExt + log2(S) ) ) and a linear sample position dependent term (e.g. slopeCurr*stepCurr with slopeCurr = slope * ( S * szExt - j ) and stepCurr = 25 + j*10), wherein the decoder is configured to compute the offset term depending on the end offset (e.g. offset), compute the linear sample position dependent term based on a product comprising a first factor term corresponding to a product (e.g. slope * ( S * szExt -j ) ) between the end slope (e.g. slope) on the one hand and a difference between a length term (e.g. S * szExt) which linearly depends on a length (e.g. szExt) of the non-available portion and a sample position value (e.g. j ) measuring a sample position between the abutting end and the opposite end from the abutting end (e.g. as an offset from the lead end) on the other hand, and a second factor (e.g. 25 + j*10) related to the sample position value according to an affine mapping.

[0064] According to an embodiment of the invention, the decoder is configured to decode for the respective temporal block of the predetermined type a sub-sample pointer for locating the reference portion in the digital time-varying signal or in a reference channel of the digital timevarying signal.

[0065] PCT_FH250111PEP-2025353041.DOCX According to an embodiment of the invention, the temporal blocks 140 of the digital timevarying signal 92 are non-overlapping and the decoder is configured to decode for the respective temporal block of the predetermined type a prediction residual from the data stream by decoding a transform (420) of the prediction residual 418 from the data stream and applying a re-transformation (e.g. involving basis functions coinciding in length with a length of the temporal blocks of the predetermined type; e.g. a sample rate of the digital time-varying signal (92) is above, or equal to, a Nyquist rate of the transform domain; a number of coefficients of the transform coincides with a number of samples of the respective temporal block) onto the transform, and correct a predictor obtained by the block prediction (such as a block-copy operation) (500) using the prediction residual.

[0066] According to an embodiment of the invention, the digital time-varying signal 92 is obtained by at least one of Electrocardiography, Electroencephalography, Electromyography or seismic measurement, and / or wherein the digital time-varying signal 92 is a bio-physiological waveform data such as an electroencephalography (EEG) signal, an electrocardiogram (ECG), or an electromyography (EMG) signal, or is a seismic waveform signal.

[0067] Features and functionalities may only be described with respect to a decoder or an encoder however it is noted that those are also applicable to a corresponding encoder / decoder. Furthermore, it is also noted that features and functionalities described with respect to a decoder or encoder are also applicable to a corresponding method, data stream or computer program, and vice versa.

[0068] Brief Description of the Drawings

[0069] 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:

[0070] Fig. 1 shows three parametrizable functions according to embodiments;

[0071] Fig. 2 shows a comparison between existing extrapolation methods and a parametrizable functions according to embodiments;

[0072] Fig. 3 shows three further parametrizable functions according to embodiments;

[0073] PCT_FH250111PEP-2025353041.DOCX Fig. 4a+b show schematic illustrations of decoders according to embodiments related to the first aspect of the invention;

[0074] Fig. 5 shows a schematic illustration of a decoder according to embodiments related to the second aspect of the invention;

[0075] Fig. 6 shows a schematic illustration of an encoder according to embodiments related to the first aspect of the invention;

[0076] Fig. 7 shows a schematic illustration of an encoder according to embodiments related to the second aspect of the invention;

[0077] Fig. 8 shows a schematic illustration of a possible framework embodiments may be implemented in;

[0078] Fig. 9 shows a schematic illustration of a method for decoding according to embodiments related to the first aspect of the invention;

[0079] Fig. 10 shows a schematic illustration of a method for decoding according to embodiments related to the second aspect of the invention;

[0080] Fig. 11 shows a schematic illustration of a method for encoding according to embodiments related to the first aspect of the invention; and

[0081] Fig. 12 shows a schematic illustration of a method for encoding according to embodiments related to the second aspect of the invention.

[0082] Detailed Description of the Embodiments

[0083] Equal or equivalent elements or elements with equal or equivalent functionality may be denoted in the following description by equal or equivalent reference numerals even if occurring in different figures.

[0084] In 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

[0085] PCT_FH250111PEP-2025353041.DOCX 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.

[0086] Embodiments of the invention make use of a parametrizable function. Figs. 1 to 3 exemplarily illustrate different embodiments of such parametrizable function, each of which may be implemented in any of the embodiments described herein, whether for encoding or for decoding purposes.

[0087] The description now proceeds with a description of decoder and encoder according to the first aspect of the invention (e.g. with regards to Fig. 4a, 4b and Fig. 6) before continuing with a description of a decoder and encoder according to the second aspect of the invention (e.g. with regards to Fig. 5 and Fig. 7). Afterwards the description resumes to a general description of embodiments (e.g. with regards to Fig. 1 to 3) that can be implemented into any of the previously described decoder and encoder.

[0088] Fig. 4a shows a schematic illustration of a decoder 12 according to embodiments related to the first aspect of the invention.

[0089] The decoder 12 is configured to decode, from a data stream 16, a digital time-varying signal 92 using block-wise predictive coding. The decoder 12 is configured to decode the digital timevarying signal 92 block-wise and for decoding a respective block of a predetermined type, the decoder 12 may comprise a parametrizable function applier 463 for applying a parametrizable function 408 having a leading-end slope 410 and a leading-end offset 412. The parametrizable function 408 depicts a basis for the residual coding, wherein the decoder 12 is configured to, using a prediction residual decoder 464, decode a prediction residual 418 from the data stream 16 and reconstruct 424 the digital time-varying signal 92 by correcting 422 the parametrizable function 408 using the prediction residual 418. For example, a result of the correction 422 may depict a reconstructed signal for the respective block 140.

[0090] The decoder 12 obtains the leading-end slope 410 using a leading-end slope obtainer 461 that is configured to obtain the leading-end slope 410 based on a previously decoded portion 404 of the digital time-varying signal 92. Similar, the decoder 12 obtains the leading-end offset 412 using a leading-end offset obtainer 462 that is configured to obtain the leading-end offset 412 based on the previously decoded portion 404.

[0091] PCT_FH250111PEP-2025353041.DOCX The leading-end offset 412 may depict an offset of the parametrizable function 408 at a leading-end 406. The leading-end slope 410 may depict a slope of the parametrizable function 408 at the leading-end 406.

[0092] As shown, the parametrizable function 408 changes monotonically, initially moving toward a zero slope. This monotonic change may continue beyond zero slope, but it always follows the initial direction.

[0093] The previously decoded portion 404 may, for example, be a portion of the digital time-varying signal 92 that immediately precedes the leading-end or may, for example, be a portion of the digital time-varying signal 92 that is placed elsewhere, for example, at a position based on meta information codable from the data stream 16.

[0094] Fig. 4b shows a schematic illustration of a decoder 12 according to embodiments related to the first aspect of the invention. The decoder 12 can be understood as an extension of the decoder 12 of Fig. 4a. Features and functionalities of the decoder 12 may be taken individually or in combination and be implemented in the decoder 12 of Fig. 4a. In the following description of the decoder 12, redundant explanations of the units of the decoder 12 that are also available or have already been described with respect to the decoder 12 of Fig. 4a are omitted.

[0095] Certain embodiments may be described the following. The decoder 12 is reconstructing 424 the digital time-varying signal by use of block-wise predictive transform coding. By doing so the decoder 12 decodes a prediction residual 418 based on transform (e.g. coefficients) 420 derived from the data stream 16. To obtain the prediction residual 418 the decoder 12 may apply a transformation from a sequential domain into a sample domain (e.g. spatial domain). The obtained prediction residuals 418 is used by the decoder to correct 422 a parametrizable function 408 to obtain a reconstruction of the digital time-varying signal 92 (e.g. for a respective block). The parametrizable function 408 may depend on a slope of a previously decoded portion, an offset of a previously decoded portion, a dampening rate parameter (e.g. S) derived from the data stream 16, and / or a length 428 of the respective temporal block. Described herein are different advantageous configurations according to embodiments of the parametrizable function 408.

[0096] The decoder 12 may decode, from the data stream 16, additional meta information helping in the predictive coding of the digital time-varying signal 92. For determining if the respective block is of the predetermined type, the decoder 12 may be decoder a mode syntax element 402 that indicates, if the respective block is of the predetermined type. In other words, the

[0097] PCT_FH250111PEP-2025353041.DOCX mode syntax element 402 may indicate if the predictive coding scheme shall be applied or optionally, which predictive coding scheme shall be applied.

[0098] As further metadata, the decoder 12 may decode a sub-type syntax element 430 from the data stream 16, which allows the decoder to classify the respective temporal block of the predetermined type into one of a set of sub-types, for example those being associated with S=1 or S=2, with S being a damping rate parameter of the parametrizable function 408.

[0099] The decoder 12 may parametrize the parametrizable function 408 depending on a length 428 of the respective temporal block of the predetermined type. In particular, for a temporal block classified as a first sub-type, the decoder 12 may set the parametrizable function such that a trailing-end slope 431 at the trailing end 414 of the block deviates, in magnitude, by less than 0.1 from zero slope. For a temporal block classified as a second sub-type, the trailing-end slope 431 at the trailing end 414 of the block may deviate, in magnitude, by less than 0.1 from the negative of the leading-end slope 410.

[0100] Furthermore, in setting the leading-end slope 410 and leading-end offset 412 at the leading end 406 of the respective temporal block based on the previously decoded portion 404, the decoder 12 may determine the leading-end offset 412 using a sum of sample values of the previously decoded portion, for example p[szArr-4] + p[szArr-3] + p[szArr-2] + p[szArr-1] + 2, wherein p is an array of samples with szArr denoting an index of the leading-end 406. Other embodiments may determine an average sample value of the previously decoded portion and use the average sample value as the leading-end offset 412. The decoder 12 may similarly determine the leading-end slope 410 using a sum of addends, wherein each addend depends on a difference between sample values of the previously decoded portion, for example a weighted sum 15*(p[szArr-1] - p[szArr-4]) + 5*(p[szArr-2] - p[szArr-3]). Other embodiments may determine an average slope of the previously decoded portion and use the average slope as the leading-end slope 410.

[0101] In some embodiments, the parametrizable function 408 used by the decoder 12 comprises a sum of an offset term and a linear sample position dependent term. The offset term may, for example, be implemented as (offset « (log2SzExt + log2(S))), with log2SzExt being Iog2 of an extrapolation size (e.g. log 2 of length 428), and defines the offset at the leading-end of a pointer 450 pointing to a function value 452 of the function 408 for a given sample position j within the temporal block 140. The linear sample position dependent term may, for example, be expressed as slopeCurr * stepCurr, where slopeCurr = slope * (S * szExt - j) and stepCurr = 25 + j*10. This term defines a pointer slope 454 of the pointer 450, which varies according

[0102] PCT_FH250111PEP-2025353041.DOCX to the sample position j, such that the pointer “tilts” progressively from the leading end towards zero slope at the trailing end.

[0103] The decoder 12 can be configured to compute the offset term based on the leading-end offset, for example using the offset determined from the previously decoded portion 404. The linear sample position dependent term can be computed as a product of two factors: a first factor corresponding to the product of the leading-end slope and the difference between a length term (e.g., S * szExt), which may depend linearly on the block length szExt, and / or a sample position value j, measuring the offset from the leading end; and a second factor, for example 25 + j*10, which defines an affine mapping of the sample position.

[0104] In some embodiments, the decoder 12 can be configured to classify the respective temporal block of the predetermined type into one of a set of sub-types, for example those having S=1 or S=2, with S being a dampening rate parameter. The parametrizable function 408 includes a length term defined as the product of the length szExt of the respective temporal block and the damping rate parameter S. The decoder 12 parametrizes the parametrizable function depending on the length 428 of the temporal block such that the damping rate parameter is larger, e.g., so that, when j assumes the maximum value corresponding to the trailing end, the difference remains larger, and in the case of S=1 , the pointer reaches zero slope at the trailing end, if the temporal block is classified into the first sub-type than if it is classified into the second sub-type.

[0105] In particular embodiments, the decoder 12 may explicitly set the damping rate parameter S to 2 if the temporal block is classified into the first sub-type, and to 1 if the temporal block is classified into the second sub-type. By adjusting the damping rate parameter in this manner, the slope progression of the parametrizable function 408 can be adapted to the characteristics of the block.

[0106] In certain embodiments, the temporal blocks 140 of the digital time-varying signal 92 may be non-overlapping. The decoder 12 can be configured to decode the prediction residual 418 from the data stream 16 by first decoding a transform 420 of the prediction residual 418 and subsequently applying a re-transformation to the transform coefficients. The re-transformation may, for example, involve basis functions whose length coincides with the length (e.g. 428) of the temporal blocks 140 of the predetermined type. In particular embodiments, the sample rate of the digital time-varying signal 92 can be above, or equal to, the Nyquist rate of the transform domain, and the number of coefficients of the transform corresponds to the number of samples in the respective temporal block.

[0107] PCT_FH250111PEP-2025353041.DOCX The decoder 12 is further configured to decode a prediction residual 418 from the data stream 16 by decoding a transform of the residual and applying a re-transformation depending on the block sub-type. For blocks classified as the first sub-type, a first re-transformation may be applied, for example a DST-4 or DST-7 transform, which employs basis functions having a magnitude increasing towards the trailing end of the block. For blocks classified as the second sub-type, a second re-transformation may be applied, for example a DCT such as DCT-2, employing basis functions of approximately constant magnitude over the block.

[0108] In certain embodiments, the decoder 12 decodes a transform type syntax element 436 from the data stream 16 for the respective temporal block. The value of this syntax element 436 determines which re-transformation is applied: a first re-transformation using basis functions with increasing magnitude towards the trailing end if the element assumes a first state, and a second re-transformation with constant-magnitude basis functions if the element assumes a second state. This allows adaptive selection of the transform type according to the properties of the block and signal.

[0109] Transforms may be understood as operations that convert a sequence of samples into a different domain, for example from a frequency in a spatial domain. Variations of transforms according to embodiments may include different types of discrete cosine transforms (DCTs), discrete sine transforms (DSTs), and other orthogonal or biorthogonal transforms, each selected to match the statistical properties of the residual signal and the prediction mode used during coding of the digital time-varying signal 92.

[0110] Furthermore, for each temporal block, the decoder 12 may select a prediction mode from a set of prediction modes. The predetermined type may represent one of these modes, while other modes may include one or more of the following:

[0111] • A DC prediction mode, where the prediction signal for the block is set to a constant function, with the constant value determined from previously decoded samples preceding the block.

[0112] A block-copy prediction mode, where the block is predicted by copying one or more reference portions of previously decoded blocks, optionally using a temporal offset signaled in the data stream.

[0113] PCT_FH250111PEP-2025353041.DOCX • A cross-channel prediction mode, where the prediction is based on one or more reference coded channels from a multi-channel signal, such as in multi-lead ECG or multi-microphone audio recordings, with the respective channel corresponding to the current digital time-varying signal 92.

[0114] • A bypass prediction mode, where the prediction residual 418 is set to zero, effectively signaling that no prediction is used for the block.

[0115] The description now proceeds with a description of an encoder 10, depicted in Fig. 6 according to embodiments related to the first aspect of the invention. It is to be understood that any functions or functionalities described with respect to the decoder 10 of Fig. 4a or Fig. 4b are, mutatis mutandis, likewise applicable to an encoder 10 of the Fig. 6.

[0116] Fig. 6 shows a schematic illustration of an encoder 10 according to embodiments related to the first aspect of the invention.

[0117] The encoder 10 is configured to encode a digital time-varying signal 92 into a data stream 16 using block-wise predictive coding. For encoding a respective temporal block 140, the encoder 10 comprises a prediction residual encoder 668 for encoding a prediction residual 418 into the data stream 16 such that the respective temporal block is reconstructable with the prediction residual 418 and a parametrizable function 408. To obtain the prediction residual 418 the encoder 10 may break up 669 the respective temporal block into the parametrizable function 408 and the prediction residual 418 by calculating a difference between the parametrizable function 408 and the respective temporal block.

[0118] The encoder 10 may be configured to determine the parametrizable function based on a leading-end slop 410 and a leading end offset 412 that determine a slope and offset of the parametrizable function at a leading-end 406 of the respective temporal block. For such purpose, the encoder 10 may comprise a leading-end slope setter 665 for setting the leadingend slope 410 based on a previously coded region 404 and a leading-end offset setter for setting the leading-end offset 412 based on a previously coded region 404. The parametrizable function may be obtained by a parametrizable function applier 463 that obtains the parametrizable function 408 to be used in the residual coding of the respective temporal block.

[0119] The decoder 12 may be configured to support temporal blocks 140 of different lengths 428. In particular, the decoder can determine the length of a transform-coded temporal block based

[0120] PCT_FH250111PEP-2025353041.DOCX on a length parameter signaled in the data stream 16. This allows the decoder 12 to flexibly adapt the block length depending on the characteristics of the digital time-varying signal 92, such as its temporal complexity or rate of variation.

[0121] In some embodiments, the decoder 12 may further be configured to switch between different supported block lengths at predefined borders between consecutive temporal blocks. The switching may likewise be controlled according to the length parameter signaled in the data stream 16. By dynamically adjusting the block length at such predetermined boundaries, the decoder can improve coding efficiency and reconstruction accuracy, for example by using shorter blocks in regions of rapid signal change and longer blocks in more stationary regions.

[0122] The decoder 12 may further be configured to predict specific temporal blocks 140 of an additional predetermined type using block prediction. In such embodiments, a reference portion (e.g. 502) of the digital time-varying signal 92 is utilized, wherein a non-available portion (e.g. 504) abutting an end (e.g. 506) of the reference portion is padded. The padding may be performed using the parametrizable function 408 to achieve a smooth continuation of the reference portion (e.g. 502) into the non-available region. This concept is exemplarily illustrated in embodiments related to the second aspect of the invention; however, it is noted that both concepts may also be combined within a single embodiment.

[0123] The description now proceeds with a description of decoder and encoder according to the second aspect of the invention with regards to the Fig. 5 and Fig. 7.

[0124] Fig. 5 shows a schematic illustration of a decoder 12 according to embodiments related to the second aspect of the invention.

[0125] The decoder 12 is configured to decode a digital time-varying signal 92 in temporal blocks 140 from a data stream 16. In particular, the decoder 12 is configured to predict each temporal block of a predetermined type by performing a block prediction 500, such as a block-copy operation, from a reference portion 502 of the digital time-varying signal 92. For this purpose, the decoder 12 leverages a parametrizable function 514, which may correspond to or be the same as the parametrizable function 408 described in other embodiments. The parametrizable function 514 is applied to obtain a non-available portion 504 of the reference portion 502 that abuts an end 506, for example a leading-end or a trailing-end, of an available portion 508 of the reference portion 502.

[0126] PCT_FH250111PEP-2025353041.DOCX The non-available portion 504 may arise, for example, due to interpolation operations, such as when a sub-sample offset is signaled for the temporal block in the data stream and locates the reference portion 502 relative to the block, or when the reference portion extends beyond accessible past or future samples, e.g., due to restrictions in the decoder’s memory or in multichannel reference signals. The parametrizable function 514 is parametrizable in terms of an end slope 510 and an end offset 512 at the abutting end of the non-available portion 504, and comprises a monotonic change in slope from the abutting end of the available portion 508 towards the opposite end 516 of the non-available portion 504, starting from the abutting end in a manner that gradually changes towards zero slope. The decoder 12 sets the end slope 510 and end offset 512 at the end of the non-available portion 504 based on the characteristics of the available portion 508 of the reference portion 502.

[0127] The parametrizable function 514 is dependent on an end offset 512 and on an end slope 510. The end offset 512 may define the value of the function at the abutting end 506 of the available portion 508 of the reference portion 502. In other words, the end offset 512 specifies the starting value of the function for the non-available portion 504 that is to be predicted or padded. The end slope 510 defines the initial rate of change of the parametrizable function 514 at the abutting end 506. This slope determines how quickly the function 514 progresses away from the abutting end as it moves towards the opposite end 516 of the non-available portion 504.

[0128] In certain embodiments, the decoder 12 may be configured to adjust a rate of a change of the slope based on a length 520 of the non-available portion 504. In this way, the progression of the parametrizable function can be tailored to the specific duration or size of the non-available portion.

[0129] In some embodiments, the decoder 12 may be configured to decode, for each temporal block of the predetermined type, a sub-sample pointer from the data stream16. This pointer may be used to locate the reference portion within the digital time-varying signal 92 or within a corresponding reference channel. By using such a pointer, the decoder can accurately determine the position of the reference portion 502 data even when the alignment is not on exact sample boundaries.

[0130] Features and functionalities described with respect to the parametrizable function 408 are also applicable to the parametrizable function 514, or rather the decoder 12 according to embodiments related to the second aspect of the invention. The end offset 512 may relate to the leading-end offset 412 and the end slop 510 may relate to the leading-end slope 510.

[0131] PCT_FH250111PEP-2025353041.DOCX The description now proceeds with a description of an encoder 10, depicted in Fig. 7 according to embodiments related to the first aspect of the invention. It is to be understood that any functions or functionalities described with respect to the decoder 10 of Fig. 5 are, mutatis mutandis, likewise applicable to an encoder 10 of the Fig. 7.

[0132] Fig. 7 shows a schematic illustration of the encoder 10 according to embodiments related to the second aspect of the invention.

[0133] The encoder 10 is configured to encode a digital time-varying signal 92 into a data stream 16 in temporal blocks 140. In particular, the encoder 10 is configured to predict each temporal block of a predetermined type by performing a block prediction 500, such as a block-copy operation, from a reference portion 502 of the digital time-varying signal 92. For this purpose, the encoder 10 may employ a parametrizable function 514, which may correspond to or be the same as the parametrizable function 408 described in other embodiments. The parametrizable function 514 is applied to generate or pad a non-available portion 504 of the reference portion 502 that abuts an end 506, such as a leading end or trailing end, of an available portion 508 of the reference portion 502.

[0134] The parametrizable function 514 is parametrizable in terms of an end slope 510 and an end offset 512 at the abutting end of the non-available portion 504, and comprises a monotonic change in slope from the abutting end of the available portion 508 towards the opposite end 516 of the non-available portion 504, starting from the abutting end in a manner that gradually changes towards zero slope. The encoder 10 determines and sets the end slope 510 and end offset 512 at the end of the non-available portion 504 based on the characteristics of the available portion 508 of the reference portion 502.

[0135] Features and functionalities described with respect to the parametrizable function 408 are also applicable to the parametrizable function 514, or rather the decoder 12 according to embodiments related to the second aspect of the invention. The end offset 512 may relate to the leading-end offset 412 and the end slop 510 may relate to the leading-end slope 510.

[0136] The description now proceeds with a general description of embodiments that can be implemented into any of the previously described decoder 12 and encoder 10 (e.g. into a decoder and / or encoder related to the first and / or second aspect of the invention). References made solely to embodiments of the first or second aspect of the invention are provided for conciseness only, and no limitation to either aspect of the invention is intended or should be inferred.

[0137] PCT_FH250111PEP-2025353041.DOCX Embodiments of the invention may also be titled “Sample Extrapolation in Coding of a Digital Time-Varying Signal”.

[0138] The description now continues with a discussion of the motivation for the embodiments, while simultaneously presenting features and functionalities of the invention. The reference signs used herein refer to the figures described later on, and the features and functionalities described in this section may be incorporated into the embodiments described later, and vice versa.

[0139] Embodiments of the invention deal with the coding of biomedical or general waveform data. Examples of such data are ECG-signals, PPG-signals, EMG signals, seismic data or audio signals. The data may in general be organized into multiple channels. Thus, the sample values of the signal X (e.g. the digital time-varying signal 92) to be coded can be represented as X[c][n], where c denotes a channel index and n denotes a time-index. The number of channels shall be denoted by C. In the general setup of the present application, the signal X (e.g. the digital time-varying signal 92) is coded in a block-wise way with predictive coding and transform coding of prediction residuals (e.g. 418). This may mean that a sequence of consecutive blocks bk(e.g. temporal blocks 140) with starting position skand length lk(e.g. length 428) is formed where the sample values on bkare the values

[0140] X[c] [n], sk< n < sk+ lk, 0 < c < C.

[0141] For a channel m with 0 < m < C, the sample values on the channel-wise sub-block bk mof the block bk(e.g. temporal block 140) are the sample values

[0142] X[m] [n], sk< n < sk+ lk.

[0143] In a certain embodiment, the blocks bkand bk+1may always be adjacent, which means thatsfc+i =sk + lk- The block-lengths lk(e.g. length 428) can be coded in the bitstream (e.g. data stream 16). They might vary or might all be equal. Typically, they (e.g. the lengths 428) should be taken from a fixed set of integral powers of two.

[0144] For coding according to embodiments, one may start with the block b0and then code the blocks sequentially, i.e. , one codes bk+1after having coded bk. Moreover, each block bk(e.g.

[0145] PCT_FH250111PEP-2025353041.DOCX 140) may be coded by sequentially coding the blocks bk m(e.g. 140), i.e., one starts with coding bk 0(e.g. 140) and codes bk m+1(e.g. 140) after having coded bk m(e.g. 140).

[0146] On each block bk m(e.g. 140), a prediction and a transform coding may be performed. This means that first, the sample values on bk m(e.g. of a respective temporal block 140) are predicted out of already coded and reconstructed sample values [Z] [n] (e.g. sample values of previously decoded portion 404, where either I < m and 0 < n < sk+ lk(e.g. sample values of previous channels at all temporal positions including the current positions) or I = m and 0 < n < sk(e.g. sample values of the current channel at previous temporal positions). Next, according to certain embodiments, the residual samples are transformed either by a trigonometric transform or by the identity transform, where the latter is possibly coupled with a sample-wise residual prediction, the resulting transform coefficients are quantized and the quantization indices are entropy coded (e.g. a transform 420 may be decodable from the data stream 16).

[0147] An example for the aforementioned prediction, that may also optionally be implemented in embodiments of the invention, is a block-copy prediction, where the prediction sample values are generated by the sample values of a previous consecutive portion of length lk, the location of the portion being coded in the bit-stream, where the location might also be a fractional location. Thus, the prediction is generated as

[0148] X[sk— offset + 1], 0 < 1 < lk, where X are the reconstructed sample values and where offset denotes a temporal offset that is transmitted in the bit-stream. The latter block-copy prediction is a very effective method if the underlying waveform signal has a periodic behavior.

[0149] Similarly, invoking sample values from previous channels might yield a very accurate prediction if strong correlations between channels exist in the underlying waveform signal (e.g. in the multi-channel digital time-varying signal 14; e.g. in the digital time-varying signal 92). However, it is asserted herein that there are important use-cases where neither the blockcopy prediction nor any kind of cross-channel prediction is feasible.

[0150] First, the cross channel prediction might not be feasible either because one wants to generate a prediction for the very first channel, i.e. for a block bk 0, or because one wants to enable an independent processing of the channels in order to increase the encoder (e.g. 10) throughput or to enable rapid accessing at a decoder (e.g. 12).

[0151] PCT_FH250111PEP-2025353041.DOCX 1

[0152] Second, the block-matching prediction might not be feasible because the encoder (e.g. 10) might have extremely limited computational resources or because a very small encoding delay is required. In such a case, the determination of the actual periodicity of the signal, i.e. the determination of the offset offset from the above equation, might be impossible since it typically requires to perform some sort of an encoder-side search algorithm. Another case where a block-matching prediction might be infeasible occurs if the underlying periodicity of the signal has a very large value but if it is undesirable to access reconstructed samples that are very far away in temporal distance. For example, it might be the case that the blockmatching prediction becomes very effective if the value offset is set to 10000. This would mean that at an encoder and at a decoder, at least 10000 previously coded and reconstructed sample values would need to be stored. This might be too much of a complexity burden for example in the case of an encoder that is implemented on a wearable device in order to compress and transmit the wearer’s continuously measured ECG-data in real time and on the fly. To give another example where the block-matching is not effective, it is pointed out that, in order to guarantee random access, the block-matching offset may not range over a randomaccess border. For this reason, at every random-access point, the first blocks (except the very first block) have previous reconstructed samples at their disposal which are, however, not very far away in a temporal direction and for which, thus, no periodicity might be observed.

[0153] As a consequence, embodiments of the invention support block-based prediction that exploits short-term temporal correlations. Alternatively, or additionally, embodiments of the invention can be configured to use sample-based prediction making use of the same short-term correlation. The existence of such correlations can be assumed since typical waveform data are sampled from a continuous time signal at a reasonably high sampling rate. To predict sample values using short-term correlations according to certain embodiments means that sample values

[0154] X[c][sk+ 1], 0 < 1 < lkare predicted out of the immediately adjacent reconstructed sample values (e.g. samples values of the previously decoded portion 404; e.g. other embodiments may utilize a previously decoded portion lying further back) [c][sk— i], 0 < i < V, where V is the prediction order. Thus, the prediction according to embodiments can be regarded as an extrapolation (e.g. through the use of the parametrizable function 408, 514) of the sample values [c][sk- i], 0 < i < V.

[0155] PCT_FH250111PEP-2025353041.DOCX From now on, the prediction value for the sample value X[c][sk+ 1], 0 < 1 < lk, shall be denoted by pred[l].

[0156] The problem that is dealt with herein is thus to define a suitable prediction mode or suitable prediction modes which are defined as extrapolations.

[0157] The most obvious and simple extrapolation method would be obtained by using the mean value as a constant prediction value, i.e. to set pred[l] = p for all 0 < 1 < lk. However, while such a prediction might be quite stable, it does not seem to represent the typical behavior of waveforms (e.g. as in digital time-varying signals) well since it is rather unlikely that waveforms suddenly become constant. For this reason, embodiments of the invention may be described as generating a non-constant prediction by extrapolation. Thus, taking into account again that the digital signal (e.g. digital time-varying signal 92) is sampled from a continuous signal which can even be expected to be smooth, a more appropriate extrapolation seems to be to invoke some kind of Taylor series extrapolation. In order to limit the computational complexity, it is proposed to restrict to a first order Taylor approximation. This essentially may mean that the prediction is generated by a straight line whose slope and offset in y-direction are determined from the values [c] [sk- i], 0 < i < V. However, it is pointed out that, while a straight line might yield a rather accurate prediction for sample values [c][sk+ 1] for small values of / , the prediction might become more and more inaccurate the larger I becomes.

[0158] While this phenomenon may generally not be completely be avoided for an extrapolation, embodiments of the present application are based on that the phenomenon of a degrading prediction quality far away from the left side of the block becomes extremely severe in particular for the case of extrapolating by a straight line. The reason for this is that typical waveforms (e.g. as the digital time-varying signal 92) have extremal points at which the sign of the derivative changes and thus, a straight line (e.g. as in a linear extrapolation) generates an inappropriate prediction after an extremal point.

[0159] One possibility to mitigate this problem would be to make the block size (e.g. a length 428) so small that the extrapolation by a straight line generates a good prediction for the whole block. However, if the block size is too small, the energy compaction property of the transform cannot be exploited sufficiently in the transform coding of the prediction residual, which leads to a significant bitrate increase. Another possibility would be to model the extrapolation by a more complex shape than a straight line, for example by even transmitting an extremal point of the

[0160] PCT_FH250111PEP-2025353041.DOCX to-be-predicted waveform on the block bk mand to transmit two slopes. However, it is observed that in general, the signaling overhead to transmit such side information is too large and cannot be compensated by the possibly increased prediction quality.

[0161] As a consequence, embodiments of the invention use an extrapolation with a damped slope. More precisely, the following design criteria are proposed for the extrapolation prediction according to embodiments:

[0162] 1. Certain embodiments of the invention are be configured such that the shape of the extrapolation prediction (e.g. the parametrizable function 408, 514) should be close to a straight line at the left boundary (e.g. at a leading-end 406) where the shape of the straight line (i.e. slope and offset in y-direction) should be determined from the adjacent reconstructed sample values [c] [sk- I], 0 < i < V

[0163] 2. Certain embodiments of the invention are be configured such that when approaching the right boundary of the block (e.g. respective temporal block 140), the shape of the extrapolation (e.g. the shape of the parametrizable function) should look more and more like a straight line with zero slope, i.e. like a constant function. Alternatively, the shape of the extrapolation (e.g. the shape of the parametrizable function) should be reversed at the right block boundary (i.e., the slope at the right boundary should be the negative value of the slope at the left boundary, e.g. -1*slope).

[0164] In other words, certain embodiments can be configured such that the shape of the extrapolation prediction (e.g., the parametrizable function 408, 514) is close to a straight line at the left boundary (e.g., at a leading-end 406). The slope and offset in the y-direction of this straight line are determined from the previously decoded portion.

[0165] In other words, certain embodiments can be configured such that, when approaching the right boundary of the block (e.g., the respective temporal block 140), the parametrizable function gradually approximates a straight line with zero slope.

[0166] In other words, certain embodiments can be configured such that, at the right boundary (e.g. at a leading-end 406) of a respective block, the shape the parametrizable function is reversed relative to the left boundary. In this case, the slope at the right boundary may be equal to or may be close to the negative of the slope at the left boundary.

[0167] PCT_FH250111PEP-2025353041.DOCX As a consequence of these design criteria of certain embodiments of the invention, it is noted in particular that the sample distance to the left boundary (e.g. the leading-end 406) of the position at which the parametrizable function becomes nearly constant may depend on the block-size (e.g. the length 428): In certain embodiments, for smaller blocks, the parametrizable function becomes nearly constant after fewer samples than for larger blocks. Since it is envisioned that the parametrizable function according to embodiments is to be applied within a hybrid framework invoking variable block-sizes, it is envisioned that an encoder (e.g. 10) adapts to the most appropriate shape of the parametrizable function implicitly by selecting an appropriate block size.

[0168] Fig. 1 exemplarily illustrates parametrizable functions according to embodiments of the invention for three different block lengths 428', 428", and 428"'. Specifically, Fig. 1 depicts a parametrizable function 408' corresponding to a block length 428' comprising four samples to be extrapolated, a parametrizable function 408" corresponding to a block length 428" comprising sixteen samples to be extrapolated, and a parametrizable function 408'" corresponding to a block length 428'" comprising sixty-four samples to be extrapolated. As can be seen from Fig. 1 , the parametrizable functions 408', 408", and 408'" may each be defined by a continuation of a slope corresponding to a preceding portion of sample values, or more specifically, corresponding to a previously decoded portion 404 of the digital time-varying signal 92 and the samples contained therein. Furthermore, the depicted parametrizable functions 408', 408", and 408'" clearly illustrate the monotonic change in slope from the leading end 406 of the respective temporal block towards the trailing end 414.

[0169] In certain embodiments, the monotonic change in slope may be implemented as a strictly monotonic change, that ensures a continuously varying slope without reversals and stagnations.

[0170] The parametrizable functions 408', 408", and 408'" (denoted 408) illustrated in Fig. 1 may each correspond to a respective temporal block of the predetermined type, wherein a decoder or encoder has classified the respective temporal block into a first sub-type. As shown, the parametrizable functions 408 exhibit a gradual change in slope such that the slope at the last sample of the respective temporal block approaches zero. In other words, the slope at the trailing end of the respective block deviates, in magnitude, by less than 0.1 from zero, thereby providing a smooth transition towards a flat or steady state at the end of the block.

[0171] The parametrizable functions 408', 408", and 408'" may also be introduced as parametrizable function 514 into embodiments related to the second aspect of the invention.

[0172] PCT_FH250111PEP-2025353041.DOCX In other words, Fig. 1 illustrates exemplary extrapolation function (e.g. the parametrizable functions 408) according to embodiments in dependency of the block shape (e.g. length 428) for the case of block-sizes 4, 16 and 64 (e.g. lengths 428’, 428” and 428”’).

[0173] Fig. 2 illustrates a comparison of the parametrizable function 408"' of Fig. 1 , here denoted 418, with two alternative extrapolation methods with respect to a digital time-varying signal 92 (also referred to as the original data). In particular, a linear extrapolation 291 and a DC prediction 292 are shown. In the linear extrapolation 291 , the slope of the previously decoded portion 404 (i.e., preceding sample values) is continued linearly. As can be observed, this approach provides a poor approximation of the digital time-varying signal 92 in the later stages of the block, particularly for samples beyond index 40. The DC prediction 292 similarly exhibits suboptimal performance, yielding an inaccurate extrapolation for the first half of the temporal block 140. By contrast, the parametrizable function 408 achieves a more faithful representation of the signal across the entire block, owing to its monotonic change in slope and parametrization based on the preceding portion of the signal.

[0174] In comparison, the parametrizable function 408 according to embodiments of the invention provides a good extrapolation over the entire extrapolation period, from the sample at index 0 to the sample at index 64, corresponding to a temporal block of the digital time-varying signal 92.

[0175] In other words, Fig. 2 illustrates a benefit of the proposed extrapolation according to embodiments with damped slope.

[0176] In Fig. 2, it can be clearly observed that a DC-prediction (e.g. 292; e.g. constant prediction) yields a “too flat” extrapolation while a linear extrapolation (e.g. 291) “overshoots” too much, whereas the proposed prediction (e.g. 408, 514) according to embodiments with damped slope (here, converging to zero slope at the right block boundary (e.g. at the trailing end)) is a decent compromise between both predictions.

[0177] It is further noted that for the case that the current portion (e.g. the respective temporal block) of the waveform signal (e.g. of the digital time-varying signal 92) is extremely noisy or consists of noise only, embodiments of the invention can be configured to use the mean value as an extrapolation (e.g. as the parametrizable function) only. Thus, it is proposed to use a switching between multiple extrapolation methods, one being the mean value extrapolation and one being one realization of said extrapolation method with damped slope.

[0178] PCT_FH250111PEP-2025353041.DOCX In other words, certain embodiments are configured such that when the digital time-varying signal 92 of the current respective block is extremely noisy (e.g. wherein a noise measure is above a certain threshold) or consists predominantly of noise a simplified extrapolation approach is used. In particular, the mean value of the block can be used as the parametrizable function 408 (e.g. 514). This adaptive approach may allow the system to maintain reasonable prediction quality even under challenging noisy conditions.

[0179] In the following, an explicit embodiment of the invention is described that may also be titled “Example implementation of the extrapolation prediction with damped slope”

[0180] As a possible embodiment, the following fixed-point implementation of the aforementioned prediction with damped slope is proposed:

[0181] Such embodiments may also be titled “Linear extrapolation process of an array to the right”

[0182] Input to this process (e.g. the parametrizable function 408, 514) can be:

[0183] - an input array size szArr >= 4 (e.g. a length of the total previously decoded portion 404; e.g. an index of the leading-end 406),

[0184] - an input array of sample values p[ i ] with 0 <= i < szArr (e.g. sample values of the previously decoded portion 404),

[0185] - an extrapolation size log2SzExt >= 0 (e.g. a length 428 of the respective temporal block or a base 2 logarithm of the length 428)

[0186] Output to this process (e.g. the parametrizable function 408, 514) can be the array values extrapolated to the right p[ szArr + j ] with 0 <= j < (1 « log2SzExt) and, optionally, an extrapolated mean value meanValExtr (e.g. a mean value of the extrapolated sample values).

[0187] The variable szExt (e.g. the length 428 of the respective temporal block 140) can be set to 1 « log2SzExt.

[0188] The variable slope (e.g. the leading-end slope 410) can be set to 15*( p[ szArr - 1 ] - p[ szArr -4 ] ) +5*( p[ szArr - 2 ] - p[ szArr - 3 ] ).

[0189] The variable offset (e.g. the leading-end offset 412) can be set to ( ( p[ szArr - 4 ] + p[ szArr - 3 ] + p[ szArr - 2] + p[ szArr - 1 ] +2 ) « 7 ).

[0190] For 0 <= j < szExt, the extrapolated array values p[ szArr + j ] can be defined as follows:

[0191] PCT_FH250111PEP-2025353041.DOCX - One sets stepCurr = 25 + j*10.

[0192] - One sets slopeCurr = slope * ( S * szExt - j ).

[0193] - One sets p[ szArr + j ] = ( ( offset«( log2SzExt + log2(S) ) ) + slopeCurr*stepCurr ) » ( log2SzExt + log2(S) + 9 ).

[0194] (( offset«( log2SzExt + log2(S) ) may, for example be the offset term; slopeCurr*stepCurr may, for example, be the linear sample position dependent term) One sets meanValExtr = (offset + 20 * slope ) » 9.

[0195] In other words, the parametrizable function 408, 514 for the values of the respective temporal block with a length 428 of SzExt may be depicted as:

[0196] (log2SzExt + log2(S) + 9 ) with n being the index of the leading-end 406 (>= 4; e.g. previously denoted szArr) and S being the damping parameter.

[0197] Note that S (e.g. damping rate parameter; e.g. damping rate parameter derived from sub-type syntax element 430) is a predefined strength constant, whose value can be, preferably, 2. However, to achieve the abovementioned alternative predictive-shape behavior (e.g. the shape of the extrapolation being reversed at the right block boundary, for a negative slope), S may also be set to 1 instead of 2, other values are also plausible. This causes the prediction vector to reach its maximum magnitude midway inside the given block of size szExt (instead of at the right block boundary, as previously), while the prediction values at the left and right block boundaries are very similar (they are not identical due to rounding errors in the above fixed-point implementation). The following figure depicts how the shapes of the extrapolation vectors for block sizes of 4, 16, or 64 samples change in comparison to those in the previous figures when S = 1 .

[0198] Fig. 3 illustrates exemplary parametrizable functions according to embodiments of the invention, specifically the parametrizable functions 408', 408", and 408"' (collectively denoted 408). These parametrizable functions may be understood as relating to predetermined temporal blocks classified as belonging to a second sub-type. The functions demonstrate how the parametrizable function can be adapted to different block lengths while maintaining a

[0199] PCT_FH250111PEP-2025353041.DOCX smooth progression of slope throughout the block. In particular, the slope of the parametrizable functions 408', 408", and 408"' at the respective trailing ends 414', 414", and 414'" may approach, or be equal to, the negative of the slope at the corresponding leading ends 406.

[0200] The parametrizable functions 408', 408", and 408'" may also be introduced as parametrizable function 514 into embodiments related to the second aspect of the invention.

[0201] Fig. 3 depicts a view of the for S = 1 instead of 2.

[0202] Note, further, that when applicable in a given realization, the choice of S may be signaled to the decoder, as part of the bitstream data and on a block-by-block basis, wherein one bit may suffice. For example, a value of 0 for this one bit may indicate S = 1 , while a bit value of 1 may indicate S = 2. The choice of S = 1 or S = 2 may be made, in the encoder, using a ratedistortion optimization search.

[0203] Note that, for the preferred case of S = 2, the abovementioned calculation of p[ szArr + j ] reduces to p[ szArr + j ] = ( ( offset«( log2SzExt + 1) ) + slopeCurr*stepCurr ) » ( log2SzExt + 10 ).

[0204] For the special case of log2SzExt = 0, implying a block size of szExt = 1 in this exemplary embodiment, a sample-wise prediction according to the above proposal is achieved (performing identically regardless of whether S is 1 or 2). More specifically, in this case, the “block” contains only a single sample (i.e. , at index j = 0), which is predictively coded and decoded using the above-described extrapolation process (i.e., the extrapolation result stored in p[ szArr ]) as a prediction value. Once this single sample has been reconstructed, by adding said prediction value to the decoded residual sample value, the next “block”, again containing only one sample in this case, can be predictively coded using the above proposal, making use of the last four reconstructed samples (the last of which now equals the sample reconstructed in the previous step). Note that, for this log2SzExt = 0 case, the calculation of p[ szArr + j ] reduces to p[ szArr + j ] = ( ( offset«log2(S) ) + slope* S*stepCurr ) » ( log2(S) + 9 ) = ( offset + 25 * slope ) » 9.

[0205] Embodiments of the invention can be combined with specific dedicated transforms. The disclosed embodiments may generally imply that the prediction residual 418 is of larger absolute value at the right boundary (e.g. trailing end 414) than at the left boundary (e.g.

[0206] PCT_FH250111PEP-2025353041.DOCX leading-end 406) when S = 2. This may in general not be the case for other types of prediction like block-matching prediction. Thus, when using S = 2, in certain embodiments it is proposed to invoke a trigonometric transform whose first basis function has a higher magnitude at the right part of the block than at the left part. An example for such a transform is the DST-4 or the DST-7. In specific “fast” applications, an encoder (e.g. 10) might only test such DST transforms when testing extrapolation based predictions while restricting to the DCT (preferably, DCT-2) among the trigonometric transforms when using e.g. block-matching prediction. In further specific applications, the signaling of the transform might be coupled with the signaling of the prediction mode. This means that the signaling is designed in such a way that a DST-transform with said shape of the first basis function is cheaper to code when an extrapolation-based prediction is used than when a block-matching prediction is used.

[0207] The description now proceeds with additional embodiments according to the second aspect of the invention, specifically an application of the extrapolation method to padding processes.

[0208] Some applications may require sample vectors to be extrapolated, e.g. backward or forward in time, without implementing predictive coding / decoding of sample data by subtraction or addition of the extrapolation values. In this case, embodiments of the invention according to the second aspect may be applied. Specifically, the example extrapolation with damped slope presented previously(with either S = 1 or S = 2) may be applied - more specifically, the extrapolation results (e.g. of the parametrizable function 418) p[ szArr + j ], 0 <j < szExt - may serve to generate additional “meaningful” sample data at block or signal boundaries beyond which no sample data are available (e.g. for the non-available portion 504) . This so-called padding process (e.g. according to the second aspect of the invention) can improve some prediction corner cases, especially in comparison with simple zero-padding (e.g. or in comparison with a linear padding) or cases when the available sample closest to said block or signal boundary is very different in value from the remaining samples close to said boundary (thereby also often outperforming, e.g., repetition of the closest sample value).

[0209] In a preferred embodiment, the following fixed-point implementation of said extrapolation (e.g. of the parametrizable function 408, 514) with damped slope is proposed, with the extrapolation (e.g. the parametrizable function 408, 514) acting backward in time, i.e., to the left of the given waveform (e.g. left of an available portion 508 of the digital time-varying signal 92):

[0210] In the following, a linear extrapolation process (e.g. with the parametrizable function 408, 514) of an array to the left (e.g. of a non-available portion 504) according to embodiments of the second aspect of the invention is discussed.

[0211] Input to this process (e.g. to the parametrizable function 408, 514)can be:

[0212] PCT_FH250111PEP-2025353041.DOCX - an input array starting position startPos (e.g. a value indicating the end 506 of the available portion 508)

[0213] - an input array size szArr (e.g. length of the available portion 508 (e.g. in terms of number of samples)) with szArr - startPos >= 4,

[0214] - an input array of sample values p[ i ] (e.g. sample values of the available portion 508) with startPos <= i < szArr,

[0215] - an extrapolation size (e.g. length 520 of the non-available portion 504 (e.g. in terms of number of samples)) szExt > 0

[0216] Output to this process (e.g. the parametrizable function (408, 514) are the array values (e.g. extrapolated sample values of the non-available portion 504) extrapolated to the left p[ startPos - 1- j ] with 0 < = j < szExt.

[0217] The mirrored input array sample values pMirror[ i ] with 0 <= i < szArr - startPos are defined as pMirror[ i ] = p[ szArr - 1- i ].

[0218] The extrapolation process to the right described above, with respect to embodiments, can be invoked with the input array size szArr - startPos, the input array pMirror and the extensions size set to Ceil( Log2 ( szExt ) ) as input to obtain the array values pMirror[ szArr + j ] with 0 <= j < ( 1 « Ceil( Log2 ( szExt ) ) ).

[0219] In other words, the parametrizable function 514 according to embodiments related to the second aspect of the invention can be the same parametrizable function 408 as in embodiments according to the first aspect of the invention.

[0220] For 0 <= j < szExt the value pMirror[ szArr + j ] can be assigned to p[ startPos - 1 - j ].

[0221] 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. 8 which shows an encoder for encoding a multi-channel digital signal 14 into a data stream 16 as well as decoder 12 for decoding the multi-channel digital signal 14 from data stream 16. The multichannel digital signal may comprise a plurality of digital time-varying signals 92 as coded in embodiments of the invention. This description of Fig. 8 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. 8 either by

[0222] PCT_FH250111PEP-2025353041.DOCX adopting all details / functionalities described with respect to Fig. 8 or with leaving-out some of the details / functionalities described with respect to Fig. 8. Sometimes such “optional” features of Fig. 8 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 previously / subsequently explained embodiments with the description of Fig. 8 shall not be restricted to the these explicitly identified variations of Fig. 8 in terms of leaving-out certain features.

[0223] In Fig. 8, 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. 8 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. 8 at 24.

[0224] Each channel, thus, forms a digital time-varying signal or time / amplitude or time-to-amplitude 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.

[0225] Fig. 8 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 are 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. 8 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

[0226] PCT_FH250111PEP-2025353041.DOCX 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 them 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 permute / sortthe 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.

[0227] The module in encoder 10 performing the one or more of channel transformation, channel permutation and temporal mutual alignment is indicated in Fig. 8 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 / delays 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-alignment 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.

[0228] Thus, the “actual coding” relates to the coded channels in domain 28. In the coded domain 28, the coded channels are depicted in Fig. 8 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. 8 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 mutually offset samples in

[0229] PCT_FH250111PEP-2025353041.DOCX case of, and according to, the mutual temporal alignment, if applied. In case of Fig. 8, 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.

[0230] 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 co-located. The coding is done sequentially along these blocks 140, by following a coding / decoding 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. 8 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 (e.g. the digital time-varying signal) of temporal block 140 in channel order. These previously coded / decoded temporal blocks and their samples are illustrated in Fig. 8 by way of shading. In this regard, note that in Fig. 8, merely one temporal block 140 has been illustrated explicitly in order to reduce the complexity of Fig. 8. Thus, in the specification herein, reference sign 140 is sometimes used to indicate the currently encoded / decoded temporal block or to stand representatively for all temporal blocks. Further, as depicted in Fig. 8, 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 non-overlapping.

[0231] 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 data stream 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, such as quantization, e.g. by means of dequantization, the residual signal 76 as coded into data stream 16, and an adder 78 which sums-up prediction signal 64 and the reconstructable residual

[0232] PCT_FH250111PEP-2025353041.DOCX signal 80 as obtained by residual decoder 74. To be more precise, let’s call the channelindividual 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.

[0233] 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 coding / decoding order 60 so as to reconstruct the coded channels in the coded domain 28.

[0234] 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 inter dependencies, namely intra-channel and interchannel 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.

[0235] In 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

[0236] PCT_FH250111PEP-2025353041.DOCX 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.

[0237] 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. 8. 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, but that 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.

[0238] 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 are illustrated in Fig. 8. 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

[0239] PCT_FH250111PEP-2025353041.DOCX 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.

[0240] 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 encoded / 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.

[0241] 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. 8. 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, reconstructed / 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 encoder 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

[0242] PCT_FH250111PEP-2025353041.DOCX immediately preceding samples which immediately precede block 140. The prediction parameters 90 may, thus, contain for a currently encoded / decoded 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.

[0243] 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.

[0244] 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 manner with 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 alignment 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.

[0245] It should be noted that the temporal blocks 30 might, other than illustrated in Fig.5, vary in block length rather than being of a constant length as depicted in Fig. 8. For instance, encoder

[0246] PCT_FH250111PEP-2025353041.DOCX 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.

[0247] As to the residual coder and residual decoder 70 and 82, they may use transform coding / decoding 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 nonwindowed, 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 retransformation 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.

[0248] The transform coefficients might be encoded by quantization, i.e. they may be quantized with the quantized coefficients then being coded in the data stream 16. Dequantization may occur 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 yields 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.

[0249] PCT_FH250111PEP-2025353041.DOCX 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 (e.g. the inverse transform being applied to reconstruct the time-domain signal). 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 coded. 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.

[0250] 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 be performed frame wise. That is, in each channel, the sequence of blocks 140 may be partitioned into

[0251] PCT_FH250111PEP-2025353041.DOCX 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.

[0252] 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.

[0253] 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 be coded into the data stream en block in a transform domain or samplewise in time domain.

[0254] Note 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 sample-wise 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

[0255] PCT_FH250111PEP-2025353041.DOCX 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.

[0256] 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. I F the channel group had m channels, m*n blockl 04 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 mostly recently previously en / decoded independent frame would be disallowed. Thus, each tile shown in Fig. 8 by bold lines may represent a sequence of an independent frame flowed by zero, one or more dependent frames.

[0257] As mentioned before, Fig. 8 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. 8, 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.

[0258] 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

[0259] PCT_FH250111PEP-2025353041.DOCX 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.

[0260] The description now continues with further features and functionalities that may be implemented into any of the above described embodiments of the inveniton. The description starts with a general overview of of prediction, coninunes with a more detailed description of embodiments perfoming a line fitting prediction and a more detailed description of embodiments performing a block matching prediction (e.g. block-copy operation). In the following, embodiments are described with respect to a multichannel digital signal, wherein the digital time-varying signal 92 is one of a plurality of channels of the multichannel digital signal.

[0261] Prediction of embodiments of the invention may also be described the following.

[0262] The prediction signal sample values (e.g. the extrapolated signal; e.g. sample values of the parametrizable function (e.g. 408; e.g. 514); e.g. sample values of the padding of the non- available portion 504) pred[j],0 < j < Ik

[0263] (e.g. wherein Ik denotes a length 428 of the respective temporal block) may, according to embodiments, be generated on each block bk,mout of already reconstructed sample values (e.g. out of a previously decoded portion 404, out of sample values of a previously decoded portion 404)

[0264] {rec[m][p], 0<p<sk}

[0265] (e.g. wherein m denotes a channel index of the digital time-varying signal 92; e.g. wherein Sk denotes a last index up until the signal has been decoded) that belong to the same channel or out of already reconstructed sample values {rec[r][p], 0<p<Sk+lk, 0<r<m} that belong to a previously coded channel in channel order.

[0266] Embodiments of the inveniton may support four prediction modes : A DC Prediction mode, the Line fitting prediction mode according to embodiments, a Cross-Channel Prediction mode and the Block-Matching Prediction mode according to embodiments. Certain embodiments may also support to completely bypass the prediction. The prediction mode may be selected per block (e.g. 140) bk,mand may be signaled in the bit stream (e.g. data stream 16).

[0267] Embodiments may also be describes as performing a line fitting prediction. In embodiments, the prediction signal (e.g. the extrapolated signal; the parametrizable function (e.g. 408; e.g.

[0268] PCT_FH250111PEP-2025353041.DOCX 514); e.g. from the padding of the non-available portion 504) can, for example be defined by a line with a damped slope where the slope is determined by the four reconstructed sample values (e.g. of the previously decoded portion 404) preceding the current block (e.g. the respective temporal block; e.g. the available portion 508). For example, with leading-end slope (e.g.410) pit if=15*(rec[m][Sk-1]-rec[m][Sk-4])+5*(rec[m][Sk-2]-rec[m][Sk-3]) leading-end offset (e.g. 412): offset=(rec[m][Sk-4]+rec[m][Sk-3]+rec[m][Sk-2]+rec[m][Sk-1]+2)«7 in certain embodiments, it is proposed to set pred[j]=((offset«3)+ if*(8-j)*(25+10*j))»12, 0<j<lk.

[0269] The proposed line fitting prediction mode according to embodiments may be used if the value of the syntax element block_matching_or_cross_channel_pred_flag is equal to zero and if the value of the syntax element block_pred_mode (e.g. the transform type syntax element 436) is equal to 1.

[0270] Embodiments of the invention may be configured to perform an Extension of prediction signals the left adjacent samples

[0271] To obtain an extension of the prediction residual to the left (e.g. for a non-available portion 504), which can, for example, be invoked in a sample wise prediction of reconstructed residuals, for each prediction signal, an extension pred[- tSize+j],O<j < tSize to tSize=16 many left adjacent prediction values (e.g. length 520 of unavailable portion 504 of reference portion 502) can be done according to embodiments.. Here, to determine the extended prediction (e.g. for the non-available portion 504), essentially the same process that was used in the respective prediction generation (e.g. according to embodiments related to the first aspect of the invention) on the given block can be invoked. Given the extended prediction signal, an extended residual signal (e.g. of the reference portion) may be defined as resExt[j]=rec[m][sk-tSize+j]- pred[-tSize+j],O<j<tSize . Here, a specified padding process for sample values at unavailable sample locations (e.g. for unavailable portion 504) is invoked.

[0272] PCT_FH250111PEP-2025353041.DOCX Embodiments of the invention related to the first aspect of the invention may also be defined as below with regard to first to fourth specific embodiment. Features and functionalities of embodiments of the invention described above can individually or in combination be inserted into any of the first to fourth specific embodiments described below. This is also indicated by the descriptive use of reference signs.

[0273] Please note: text in brackets of the specific embodiments is not intended to be part of the specific embodiments, but provides explanations, examples, or optional features, which may optionally be integrated into the specific embodiments.

[0274] Further, the reference signs are not to be understood as limiting the specific embodiments, but are to ease the understanding of the specific embodiments.

[0275] A first specific embodiment is related to a decoder 12 for decoding a multi-channel (e.g. digital time-varying) signal 14 (e.g. a biomedical waveform data) from a data stream (16), configured to decode each channel (e.g. representing a digital time-varying signal 92) of the multi-channel digital time-varying signal 14 from the data stream 16 in blocks 140 by decoding a current block coded in a line fitting prediction mode by generating a prediction signal (e.g. prediction signal sample) based on already reconstructed sample values (e.g. that belong to the same channel; e.g. that belong to a previously coded channel in a channel order), wherein the decoder 12 is configured to generate the prediction signal, for the current block, using a line 408 with a damped slope (e.g., wherein the decoder is configured to determine the damped slope by use of the already reconstructed sample values) (e.g. four reconstructed sample values] (e.g. immediately; e.g. out of the already reconstructed sample values) preceding the current block (e.g. wherein individual sample values of the prediction signal are generated by a value of the line with the damped slope)

[0276] (e.g. decoding, from the data stream (16), a residual signal (e.g. reconstructed residual sample values) using transform coding; and forming, sample-wise, a sum over the prediction signal and the residual signal to obtain final reconstructed sample values of the current block (e.g. rec[ j ]=pred[ j ]+ res[ j ]). ) (e.g. the decoder may also comprise a DC Prediction mode, a Cross-Channel Prediction mode and a Block-Matching Prediction mode; e.g. it may also be supported by the decoder to completely bypass the prediction).

[0277] PCT_FH250111PEP-2025353041.DOCX A first specific embodiment is related to a decoder 12 according to the first specific embodiment, wherein the decoder 12 is configured to determine the prediction signal by computing pred[j] = ((offset«3)+ if*(8-j)*(25+10*j))»12, 0<j<lk(e.g. wherein

[0278] ((offset«3)+ if*(8-j)*(25+10*j))»12, 0<j<lkis the line with the damped slope; e.g., representing an extrapolation using the line with the damped slope), wherein m denotes the current channel, if = 15*(rec[m][sk-1]-rec[m][sk-4])+5*(rec[m][sk-2]-rec[m][sk-3]), offset = (rec[m][sk-4]+rec[m][sk-3]+rec[m][sk-2]+rec[m][sk-1]+2)«7, pred[j] is the prediction signal being lksamples long, skis the position of the first sample of the current block, rec[m][p] 0<p<skdenotes already reconstructed sample values for the current channel m.

[0279] A third specific embodiment is related to an encoder 10 for encoding a multi-channel (e.g. a digital time-varying) signal 14 (e.g. a biomedical waveform data) into a data stream 16, configured to encode each channel (e.g. representing a digital time-varying signal (92)) of the multi-channel digital time-varying signal (14) into the data stream (16) in blocks (140) by encoding a current block, in a line fitting prediction mode, by generating a prediction signal (e.g. prediction signal sample values) based on already reconstructed sample values (e.g. already encoded sample values)(e.g. that belong to the same channel; e.g. that belong to a previously coded channel in a channel order), wherein the encoder 10 is configured to generate the prediction signal, for the current block, using a line 408 with a damped slope (e.g., wherein the encoder is configured to determine the damped slope by use of the already reconstructed sample values)(e.g. four reconstructed sample values)(e.g. immediately; e.g. out of the already reconstructed sample values) preceding the current block (e.g. wherein individual sample values of the prediction signal are generated by a value of the line with the damped slope)

[0280] (encoding, into the data stream 16, a residual signal (e.g. reconstructed residual sample values) using transform coding, such that forming, sample-wise, a sum over the prediction signal and the residual signal yields sample values of the current block (e.g. rec[ j ]=pred[ j ]+ res[ j ]])

[0281] PCT_FH250111PEP-2025353041.DOCX (e.g. the encoder may also comprise a DC Prediction mode, a Cross-Channel Prediction mode and a Block-Matching Prediction mode; e.g. it may also be supported by the encoder to completely bypass the prediction).

[0282] A fourth specific embodiment is related to an encoder 10 according to the third specific embodiment, wherein the encoder 10 is configured to determine the prediction signal by computing pred[j] = ((offset«3)+ if*(8-j)*(25+10*j))»12, 0<j<lk wherein m denotes the current channel, if = 15*(rec[m][Sk-1]-rec[m][Sk-4])+5*(rec[m][Sk-2]-rec[m][Sk-3]), offset = (rec[m][Sk-4]+rec[m][Sk-3]+rec[m][Sk-2]+rec[m][Sk-1]+2)«7, pred[j] is the prediction signal being Ik samples long,

[0283] Sk is the position of the first sample of the current block, rec[m][p] 0<p<Sk denotes already reconstructed sample values for the current channel m.

[0284] Even further embodiments describe computer programs performing any of the further specific embodiments described above.

[0285] Even more further embodiments describe data stream generated by any of the especific embodiments or the further specific embodiments described above.

[0286] The description now proceeds with a description of methods according to embodiments of the inveniton.

[0287] The description now proceeds with a description of methods according to embodiments of the inveniton.

[0288] Fig. 9 shows a schematic illustration of a method 9001 for decoding according to embodiments related to the first aspect of the invention. The method 9001 for decoding a digital time-varying signal from a data stream comprises decoding 9002 the signal in temporal blocks. Each temporal block of a predetermined type is decoded by extrapolating from a previously decoded portion of the signal that precedes the leading end of the block. The method 9001 further comprises predicting 9003 the temporal block using 9003 a parametrizable function. This function is defined in terms of a leading-end slope and a leading-end offset at the leading-end of the block, and exhibits a monotonic change in slope from the leading end towards the trailing

[0289] PCT_FH250111PEP-2025353041.DOCX end, starting from the leading end in a manner that gradually approaches zero slope. The leading-end slope and offset are set based on the previously decoded portion. The method 9001 further includes decoding 9005 a prediction residual from the data stream, and reconstructing 9006 the temporal block by correcting the parametrizable function using the prediction residual.

[0290] Fig. 10 shows a schematic illustration of a method 10007 for decoding according to embodiments related to the second aspect of the invention. The method 10007 for decoding a digital time-varying signal from a data stream comprises decoding 10008 the signal in temporal blocks. Each temporal block of a predetermined type is predicted by block prediction from a reference portion of the signal. In this process, a non-available portion of the reference portion that abuts an end of an available portion is reconstructed by padding using 8009 a parametrizable function. The parametrizable function is defined in terms of an end slope and an end offset at the abutting end of the non-available portion, and exhibits a monotonic change in slope from the abutting end of the available portion towards the opposite end of the non- available portion. The slope begins at the abutting end and gradually changes toward zero. The end slope and end offset are set based on the characteristics of the available portion 508 of the reference portion.

[0291] Fig. 11 shows a schematic illustration of a method 11010 for encoding according to embodiments related to the first aspect of the invention. The method 11010 for encoding a digital time-varying signal into a data stream comprises encoding 11011 the signal in temporal blocks. Each temporal block of a predetermined type is encoded by extrapolating from a previously encoded portion of the signal that precedes the leading end of the block. The method further comprises predicting 11012 the respective temporal block using 11012 a parametrizable function. This function is defined in terms of a leading-end slope and a leadingend offset at the leading end of the block, and exhibits a monotonic change in slope from the leading end towards the trailing end, starting from the leading end in a manner that gradually approaches zero slope. The leading-end slope and offset are set based on the previously encoded portion. The method 11010 also includes encoding 11014 a prediction residual into the data stream. The respective temporal block of the predetermined type is reconstructable by correcting the parametrizable function using the prediction residual, thereby ensuring accurate reconstruction of the original signal.

[0292] Fig. 12 shows a schematic illustration of a method 12015 for encoding according to embodiments related to the second aspect of the invention. The method 12015 for encoding a digital time-varying signal into a data stream comprises encoding 12016 the signal in

[0293] PCT_FH250111PEP-2025353041.DOCX temporal blocks. Each temporal block of a predetermined type is predicted by block prediction from a reference portion of the signal. In this process, a non-available portion of the reference portion that abuts an end of an available portion is reconstructed by padding using a parametrizable function. The parametrizable function is defined in terms of an end slope and an end offset at the abutting end of the non-available portion, and exhibits a monotonic change in slope from the abutting end of the available portion towards the opposite end of the non- available portion. The slope begins at the abutting end and gradually changes toward zero. The end slope and end offset are set based on the characteristics of the available portion of the reference portion, thereby enabling accurate prediction and efficient encoding of the temporal block.

[0294] Implementation alternatives

[0295] Although some aspects have been described in the context of an apparatus, it is clear that these aspects also represent a description of the corresponding method, where a block or device corresponds to a method step or a feature of a method step. Analogously, aspects described in the context of a method step also represent a description of a corresponding block or item or feature of a corresponding apparatus. Some or all of the method steps may be executed by (or using) a hardware apparatus, like for example, a microprocessor, a programmable computer or an electronic circuit. In some embodiments, one or more of the most important method steps may be executed by such an apparatus.

[0296] Depending 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.

[0297] 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.

[0298] 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

[0299] PCT_FH250111PEP-2025353041.DOCX methods when the computer program product runs on a computer. The program code may for example be stored on a machine-readable carrier.

[0300] Other embodiments comprise the computer program for performing one of the methods described herein, stored on a machine-readable carrier.

[0301] 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.

[0302] 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.

[0303] 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.

[0304] A 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.

[0305] A further embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.

[0306] 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.

[0307] 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

[0308] PCT_FH250111PEP-2025353041.DOCX microprocessor in order to perform one of the methods described herein. Generally, the methods are preferably performed by any hardware apparatus.

[0309] 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.

[0310] The apparatus described herein, or any components of the apparatus described herein, may be implemented at least partially in hardware and / or in software.

[0311] 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.

[0312] The methods described herein, or any components of the apparatus described herein, may be performed at least partially by hardware and / or by software.

[0313] 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, to 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.

[0314] PCT_FH250111PEP-2025353041.DOCX

Claims

Claims1. Decoder (12) for decoding a digital time-varying signal (92) from a data stream (16), configured to decode the digital time-varying signal (92) from the data stream (16) in temporal blocks (140) by decoding each of temporal blocks of a predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by extrapolation from a previously decoded portion (404) of the digital time-varying signal (92), which precedes a leading end (406) of the respective temporal block of the predetermined type, by predicting the respective temporal block of the predetermined type using a parametrizable function (408) which is parametrizable in terms of a leading-end slope (410) and a leadingend offset (412) at the leading end (406) of the respective temporal block of the predetermined type, and comprises a monotonic change in slope from the leading end (406) of the respective temporal block of the predetermined type towards a trailing end (414) of the respective temporal block of the predetermined type so as to start, from the leading end, in a manner changing towards zero slope, and by setting (416) the leading-end slope (410) and the leading-end offset (412) at the leading end (406) of the respective temporal block of the predetermined type based on the previously decoded portion (404); decoding a prediction residual (418) from the data stream (16), and reconstructing (424) the respective temporal block of the predetermined type by correcting (422) the parametrizable function (408) using the prediction residual (418).

2. Decoder (12) of claim 1 , wherein the decoder (12) is configured so that the parametrizable function (408) is, except for the leading-end slope (410) and the leading-end offset (412), independent from the previously decoded portion (404).PCT_FH250111PEP-2025353041.DOCX3. Decoder (12) of any of claims 1 or 2, wherein the decoder (12) is configured to parametrize the parametrizable function (408) based on the previously decoded portion (404) exclusively in terms of the leading-end slope (410) and leading-end offset (412).

4. Decoder (12) of any of claims 1 to 3, wherein the parametrizable function (408) is further parametrizable in terms of a rate of the monotonic change, and the decoder (12) is configured to set the rate depending on a length (428) of the respective temporal block of the predetermined type.

5. Decoder (12) of any of claims 1 to 4, wherein the decoder (12) is configured to parametrize the parametrizable function (408) depending on a length (428) of the respective temporal block of the predetermined type such that a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from the zero slope, or a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from minus the leading-end slope (410).

6. Decoder (12) of any of claims 1 to 5, wherein the parametrizable function (408) is differentiable on a domain of sample positions between the leading end (406) and the trailing end (414).

7. Decoder (12) of any of claims 1 to 6, wherein the decoder (12) is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types, and parametrize the parametrizable function (408) depending on a length (428) of the respective temporal block of the predetermined type such thatPCT_FH250111PEP-2025353041.DOCXa trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from zero slope if the respective temporal block of the predetermined type is classified into a first sub-type, and a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from minus the leading-end slope if the respective temporal block of the predetermined type is classified into a second sub-type.

8. Decoder (12) of claim 7, wherein the decoder (12) is configured to classify the respective temporal block of the predetermined type into one of the set of sub-types depending on a sub-type syntax element (430) signalled in the data stream (16) for the respective temporal block of the predetermined type.

9. Decoder (12) of claim 7 and 8, wherein the set of sub-types exclusively consists of the first and second sub-type.

10. Decoder (12) of any of claims 1 to 9, wherein the decoder (12) is configured to, in the setting the leading-end slope (410) and leading-end offset (412) at the leading end (406) of the respective temporal block of the predetermined type based on the previously decoded portion (404), determine the leading-end offset (412) using a sum of sample values of the previously decoded portion (404), and determine the leading-end slope (410) using a sum of addends, wherein each addend depends on a difference between sample values of the previously decoded portion (404).

11. Decoder (12) of any of claims 1 to 10, wherein the parametrizable function (408) includes a sum of an offset term and a linear sample position dependent term and, wherein the decoder (12) is configured to compute the offset term depending on the leading-end offset (412), compute the linear sample position dependent term based on a product comprisingPCT_FH250111PEP-2025353041.DOCXa first factor term corresponding to a product between the leading-end slope (410) on the one hand and a difference between a length term which linearly depends on a length (428) of the respective temporal block of the predetermined type and a sample position value measuring a sample position between the leading end (406) and the trailing end (414) from the leading end (414) on the other hand, and a second factor related to the sample position value according to an affine mapping.

12. Decoder (12) of claim 11 , wherein the decoder (12) is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types, and wherein the length term is a product of the length (428) of the respective temporal block of the predetermined type and a damping rate parameter of the parametrizable function (408), and parametrize the parametrizable function (408) depending on a length (428) of the respective temporal block of the predetermined type such that the damping rate parameter is larger if the respective temporal block of the predetermined type is classified into a first sub-type, than if the respective temporal block of the predetermined type is classified into a second sub-type.

13. Decoder (12) of claim 12, wherein the decoder (12) is configured to set the damping rate parameter to 2 if the respective temporal block of the predetermined type is classified into a first sub-type, and 1 if the respective temporal block of the predetermined type is classified into a second sub-type.

14. Decoder (12) of claim 12, wherein the parametrizable function (408) has the sum of an offset term and the linear sample position dependent term as a nominator and as a denominator a product comprising a first factor corresponding to the length (428) of the respective temporal block of the predetermined type, and / or a second factor corresponding to the damping rate parameter.

15. Decoder (12) of any of claims 1 to 14, configured toPCT_FH250111PEP-2025353041.DOCXdecode each of temporal blocks of a further predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by a sample-wise prediction and a sample wise prediction correction using a prediction residual signal (418) for the respective temporal blocks of the further predetermined type with using the parametrizable function (408) for sample-wise prediction parametrized for a 1-sample block-length.

16. Decoder (12) of any of claims 1 to 15, wherein the temporal blocks (140) of the digital time-varying signal (92) are non-overlapping and the decoder (12) is configured to decode the prediction residual (418) from the data stream by decoding a transform (420) of the prediction residual (418) from the data stream (16) and applying a re-transformation onto the transform (420).

17. Decoder (12) of any of claims 1 to 16, wherein the decoder (12) is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types, and parametrize the parametrizable function (408) depending on a length (428) of the respective temporal block of the predetermined type such that a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from zero slope if the respective temporal block of the predetermined type is classified into a first sub-type, and a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from minus the leading-end slope if the respective temporal block of the predetermined type is classified into a second sub-type, and wherein the decoder (12) is configured to decode the prediction residual (418) from the data stream (16) by decoding a transform (420) of the prediction residual (418) from the data stream and applying a first re-transformation involving basis functions having a magnitude increasing towards the trailing end (414) onto the transform (420) if the respective temporal block of the predetermined type is classified into the first sub-type, andPCT_FH250111PEP-2025353041.DOCXa second re-transformation involving basis functions having constant magnitude if the respective temporal block of the predetermined type is classified into the second subtype.

18. Decoder (12) of any of claims 1 to 17, configured to decode a transform type syntax element (436) from the data stream (16) for the respective temporal block of the predetermined type, and decode the prediction residual (418) from the data stream (16) by decoding a transform (420) of the prediction residual (418) from the data stream (16) and applying a first re-transformation involving basis functions having a magnitude increasing towards the trailing end (414) onto the transform (420) if the transform type syntax element (436) assumes a first state, and a second re-transformation involving basis functions having constant magnitude if the transform type syntax element (436) assumes a second state.

19. Decoder (12) of any of claims 1 to 18, wherein the decoder (12) selects, for each of the temporal blocks (140), a prediction mode out of a set of prediction modes with the predetermined type representing a predetermined prediction mode out of the set of prediction modes, wherein the set of prediction modes further comprises a DC prediction mode according to which a prediction signal of the respective temporal block is determined to be a constant function with a determination of a constant of the constant function based on predetermined already decoded samples preceding the respective temporal block, a block-copy prediction mode according to which the prediction signal of the respective temporal block is predicted based on one or more block reference block portions of already decoded samples preceding the respective transform-coded temporal offset relative to the respective transform-coded temporal block at a position signalled for the respective temporal block in the data stream,PCT_FH250111PEP-2025353041.DOCXa cross-channel prediction mode according to which the prediction signal of the respective temporal block is predicted based on one or more reference coded channels out of coded channels which represent a multi-channel signal (14) coded into the data stream and to be decoded from the data stream by the decoder, and one of which is represented by the digital time-varying signal (92), and a bypass prediction mode according to which the prediction signal of the respective temporal block is set to zero.

20. Decoder (12) of any of claims 1 to 19, wherein the digital time-varying signal (92) is obtained by at least one of Electrocardiography, Electroencephalography, Electromyography or seismic measurement, and / or wherein the digital time-varying signal (92) is a bio-physiological waveform data such as an electroencephalography (EEG) signal, an electrocardiogram (ECG), or an electromyography (EMG) signal, or is a seismic waveform signal.21 . Decoder (12) of any of claims 1 to 20, configured to support different lengths (428) of temporal blocks and set a length of the transform-coded temporal blocks according to a length parameter in the data stream.

22. Decoder (12) of any of claims 1 to 21 , configured to support different lengths of the temporal blocks and switch between the different lengths of the transform-coded temporal blocks at predetermined borders between consecutive temporal blocks according to a length parameter in the data stream.

23. Decoder (12) of any of claims 1 to 22, configured to predict each of temporal blocks of an even further predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by block prediction (500) from a reference portion (502) of the digital time-varying signal (92) by padding a non-available portion (504) of the reference portion abutting an end (506) of the reference portion (502) using the parametrizable function (408).PCT_FH250111PEP-2025353041.DOCX24. Decoder (12) for decoding a digital time-varying signal (92) from a data stream (16), configured to decode the digital time-varying signal (92) from the data stream (16) in temporal blocks (140) by predicting each of temporal blocks of a predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by block prediction (500) from a reference portion (502) of the digital time-varying signal (92) by padding a non-available portion (504) of the reference portion (502) abutting an end (506) of an available portion (508) of the reference portion 502 using a parametrizable function (514) which is parametrizable in terms of an end slope (510) and an end offset (512) at an abutting end of the non-available portion abutting the end (506) of the available portion (508), and comprises a monotonic change in slope from the abutting end of the available portion of the reference portion towards an opposite end (516) of the non- available portion (504) so as to start, from the abutting end, in a manner changing towards zero slope, and by setting the end slope (510) and end offset (512) at the abutting end of the non- available portion (504) based on the available portion (508) of the reference portion.

25. Decoder (12) of claim 24, wherein the decoder (12) is configured so that the parametrizable function (514) is, except for the leading-end slope (510) and the leading-end offset (512), independent from the available portion (508).

26. Decoder (12) of any of claims 24 to 25, wherein the decoder (12) is configured to parametrize the parametrizable function (514) based on the available portion (508) exclusively in terms of the leading-end slope (510) and leading-end offset (512).

27. Decoder (12) of any of claims 24 to 26, wherein the parametrizable function (514) is further parametrizable in terms of a rate of the monotonic change, and the decoder (12) is configured to set the rate depending on a length (520) of the non-available portion (504).PCT_FH250111PEP-2025353041.DOCX28. Decoder (12) of any of claims 24 to 27, wherein the decoder (12) is configured to parametrize the parametrizable function (514) depending on a length (520) of the non- available portion (504) such that an opposite-end slope of the parametrizable function (514) at the opposite end (516) of the non-available portion (504) deviates, in magnitude, by less than 0.1 from the zero slope or a trailing-end slope (431) of the parametrizable function (408) at the opposite end (516) of the non-available portion (504) deviates, in magnitude, by less than 0.1 from minus the end slope (410).

29. Decoder (12) of any of claims 24 to 28, wherein the parametrizable function (514) is differentiable on a domain of sample positions between the end (506) and the opposite end (516).

30. Decoder (12) of any of claims 24 to 29, wherein the decoder (12) is configured to, in the setting the end slope (510) and end offset (512) at the abutting end (506) of the non- available portion (504) based on the available portion (508), determine the end offset (512) using a sum of sample values of the available portion (508), and determine the end slope (510) using a sum of addends, wherein each addend depends on a difference between sample values of the available portion (508).

31. Decoder (12) of any of claims 24 to 30, wherein the parametrizable function (514) includes a sum of an offset term and a linear sample position dependent term, wherein the decoder (12) is configured to compute the offset term depending on the end offset (512), compute the linear sample position dependent term based on a product comprising a first factor term corresponding to a product between the end slope (510) on the one hand and a difference between a length term which linearly dependsPCT_FH250111PEP-2025353041.DOCXon a length (520) of the non-available portion (504) and a sample position value measuring a sample position between the abutting end (506) and the opposite end from the abutting end (506) on the other hand, and a second factor related to the sample position value according to an affine mapping.

32. Decoder (12) of any of claims 24 to 31 , configured to decode for the respective temporal block of the predetermined type a sub-sample pointer for locating the reference portion in the digital time-varying signal (92) or in a reference channel of the digital time-varying signal (92).

33. Decoder (12) of any of claims 24 to 32, wherein the temporal blocks (140) of the digital time-varying signal (92) are non-overlapping and the decoder (12) is configured to decode for the respective temporal block of the predetermined type a prediction residual from the data stream (16) by decoding a transform (420) of the prediction residual (418) from the data stream (16) and applying a re-transformation onto the transform (420), and correct a predictor obtained by the block prediction (500) using the prediction residual.

34. Decoder (12) of any of claims 24 to 33, wherein the digital time-varying signal (92) is obtained by at least one of Electrocardiography, Electroencephalography, Electromyography or seismic measurement, and / or wherein the digital time-varying signal (92) is a bio-physiological waveform data such as an electroencephalography (EEG) signal, an electrocardiogram (ECG), or an electromyography (EMG) signal, or is a seismic waveform signal.

35. Decoder (12) of any of claims 24 to 34, configured to support different lengths of temporal blocks and set a length of the transformcoded temporal blocks according to a length parameter in the data stream (16).

36. Decoder (12) of any of claims 24 to 35,PCT_FH250111PEP-2025353041.DOCXconfigured to support different lengths of the temporal blocks and switch between the different lengths of the transform-coded temporal blocks at predetermined borders between consecutive temporal blocks according to a length parameter in the data stream (16).

37. Encoder (10) for encoding a digital time-varying signal (92) into a data stream (16), configured to encode the digital time-varying signal (92) into the data stream (16) in temporal blocks (140) by encoding each of temporal blocks of a predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by extrapolation from a previously encoded portion (404) of the digital time-varying signal (92), which precedes a leading end (406) of the respective temporal block of the predetermined type, by predicting the respective temporal block of the predetermined type using a parametrizable function (408) which is parametrizable in terms of a leading-end slope (410) and a leadingend offset (412) at the leading end (406) of the respective temporal block of the predetermined type, and comprises a monotonic change in slope from the leading end (406) of the respective temporal block of the predetermined type towards a trailing end (414) of the respective temporal block of the predetermined type so as to start, from the leading end, in a manner changing towards zero slope, and by setting (416) the leading-end slope (410) and the leading-end offset (412) at the leading end (406) of the respective temporal block of the predetermined type based on the previously encoded portion (404); encoding a prediction residual (418) into the data stream (16), and wherein the respective temporal block of the predetermined type is reconstructable (424) by correcting (422) the parametrizable function (408) using the prediction residual (418).PCT_FH250111PEP-2025353041.DOCX38. Encoder (10) of claim 37, wherein the encoder (10) is configured so that the parametrizable function (408) is, except for the leading-end slope (410) and the leading-end offset (412), independent from the previously encoded portion (404).

39. Encoder (10) of any of claims 37 or 38, wherein the encoder (10) is configured to parametrize the parametrizable function (408) based on the previously encoded portion (404) exclusively in terms of the leading-end slope (410) and leading-end offset (412).

40. Encoder (10) of any of claims 37 to 39, wherein the parametrizable function (408) is further parametrizable in terms of a rate of the monotonic change, and the encoder (10) is configured to set the rate depending on a length (428) of the respective temporal block of the predetermined type.

41. Encoder (10) of any of claims 37 to 40, wherein the encoder (10) is configured to parametrize the parametrizable function (408) depending on a length (428) of the respective temporal block of the predetermined type such that a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from the zero slope, or a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from minus the leading-end slope (410).

42. Encoder (10) of any of claims 37 to 41 , wherein the parametrizable function (408) is differentiable on a domain of sample positions between the leading end (406) and the trailing end (414).

43. Encoder (10) of any of claims 37 to 42, wherein the encoder (10) is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types, andPCT_FH250111PEP-2025353041.DOCXparametrize the parametrizable function (408) depending on a length (428) of the respective temporal block of the predetermined type such that a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from zero slope if the respective temporal block of the predetermined type is classified into a first sub-type, and a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from minus the leading-end slope (410) if the respective temporal block of the predetermined type is classified into a second sub-type.

44. Encoder (10) of any of claims 37 to 43, wherein the encoder (10) is configured to classify the respective temporal block of the predetermined type into one of the set of subtypes depending on a sub-type syntax element (430) signalled in the data stream (10) for the respective temporal block of the predetermined type.

45. Encoder (10) of any of claims 37 to 44, wherein the set of sub-types exclusively consists of the first and second sub-type.

46. Encoder (10) of any of claims 37 to 45, wherein the encoder (10) is configured to, in the setting the leading-end slope (410) and leading-end offset (412) at the leading end (406) of the respective temporal block of the predetermined type based on the previously encoded portion (404), determine the leading-end offset (412) using a sum of sample values of the previously encoded portion (404), and determine the leading-end slope (410) using a sum of addends, wherein each addend depends on a difference between sample values of the previously encoded portion (404).

47. Encoder (10) of any of claims 37 to 46, wherein the parametrizable function (408) includes a sum of an offset term and a linear sample position dependent term, wherein the encoder (10) is configured toPCT_FH250111PEP-2025353041.DOCXcompute the offset term depending on the leading-end offset (412), compute the linear sample position dependent term based on a product comprising a first factor term corresponding to a product between the leading-end slope (410) on the one hand and a difference between a length term which linearly depends on a length (428) of the respective temporal block of the predetermined type and a sample position value measuring a sample position between the leading end (406) and the trailing end (414) from the leading end (406) on the other hand, and a second factor related to the sample position value according to an affine mapping.

48. Encoder (10) of claim 47, wherein the encoder (10) is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types, and wherein the length term is a product of the length (428) of the respective temporal block of the predetermined type and a damping rate parameter of the parametrizable function (408), and parametrize the parametrizable function (408) depending on a length (428) of the respective temporal block of the predetermined type such that the damping rate parameter is larger if the respective temporal block of the predetermined type is classified into a first sub-type, than if the respective temporal block of the predetermined type is classified into a second sub-type.

49. Encoder (10) of claim 48, wherein the encoder (10) is configured to set the damping rate parameter to 2 if the respective temporal block of the predetermined type is classified into a first sub-type, and 1 if the respective temporal block of the predetermined type is classified into a second sub-type.

50. Encoder (10) of claim 48, wherein the parametrizable function (408) has the sum of an offset term and the linear sample position dependent term as a nominator and as a denominator a product comprising a first factor corresponding to the length (408) of thePCT_FH250111PEP-2025353041.DOCXrespective temporal block of the predetermined type, and / or a second factor corresponding to the damping rate parameter.51 . Encoder (10) of any of claims 37 to 50, configured to encode each of temporal blocks of a further predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by a sample-wise prediction and a sample wise prediction correction using a prediction residual signal (418) for the respective temporal blocks of the further predetermined type with using the parametrizable function (408) for sample-wise prediction parametrized for a 1-sample block-length.

52. Encoder (10) of any of claims 37 to 51 , wherein the temporal blocks (140) of the digital time-varying signal (92) are non-overlapping and the encoder (10) is configured to encode the prediction residual (418) into the data stream (16) by encoding a transform (420) of the prediction residual (418) into the data stream (16) and applying a re-transformation onto the transform (420).

53. Encoder (10) of any of claims 37 to 52, wherein the encoder (10) is configured to classify the respective temporal block of the predetermined type into one of a set of sub-types, and parametrize the parametrizable function (408) depending on a length (428) of the respective temporal block of the predetermined type such that a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from zero slope if the respective temporal block of the predetermined type is classified into a first sub-type, and a trailing-end slope (431) of the parametrizable function (408) at the trailing end (414) of the respective temporal block of the predetermined type deviates, in magnitude, by less than 0.1 from minus the leading-end slope (410) if the respective temporal block of the predetermined type is classified into a second sub-type, andPCT_FH250111PEP-2025353041.DOCXwherein the encoder (10) is configured to encode the prediction residual (418) into the data stream (16) by encoding a transform (420) of the prediction residual (418) into the data stream (16) and applying a first re-transformation involving basis functions having a magnitude increasing towards the trailing end (414) onto the transform (420) if the respective temporal block of the predetermined type is classified into the first sub-type, and a second re-transformation involving basis functions having constant magnitude if the respective temporal block of the predetermined type is classified into the second subtype.

54. Encoder (10) of any of claims 37 to 53, configured to encode a transform type syntax element (436) into the data stream (16) for the respective temporal block of the predetermined type, and encode the prediction residual (418) into the data stream (16) by encoding a transform of the prediction residual into the data stream (16) and applying a first re-transformation involving basis functions having a magnitude increasing towards the trailing end (414) onto the transform (420) if the transform type syntax element (436) assumes a first state, and a second re-transformation involving basis functions having constant magnitude if the transform type syntax element (436) assumes a second state.

55. Encoder (10) of any of claims 37 to 54, wherein the encoder (10) selects, for each of the temporal blocks, a prediction mode out of a set of prediction modes with the predetermined type representing a predetermined prediction mode out of the set of prediction modes, wherein the set of prediction modes further comprises a DC prediction mode according to which a prediction signal of the respective temporal block is determined to be a constant function with a determination of a constant of the constant function based on predetermined already encoded samples preceding the respective temporal block,PCT_FH250111PEP-2025353041.DOCXa block-copy prediction mode according to which the prediction signal of the respective temporal block is predicted based on one or more block reference block portions of already encoded samples preceding the respective transform-coded temporal offset relative to the respective transform-coded temporal block at a position signalled for the respective temporal block in the data stream, a cross-channel prediction mode according to which the prediction signal of the respective temporal block is predicted based on one or more reference coded channels out of coded channels which represent a multi-channel signal (14) coded into the data stream and to be encoded into the data stream by the encoder, and one of which is represented by the digital time-varying signal (92), and a bypass prediction mode according to which the prediction signal of the respective temporal block is set to zero.

56. Encoder (10) of any of claims 37 to 55, wherein the digital time-varying signal (92) is obtained by at least one of Electrocardiography, Electroencephalography, Electromyography or seismic measurement, and / or wherein the digital time-varying signal (92) is a bio-physiological waveform data such as an electroencephalography (EEG) signal, an electrocardiogram (ECG), or an electromyography (EMG) signal, or is a seismic waveform signal.

57. Encoder (10) of any of claims 37 to 56, configured to support different lengths of temporal blocks and set a length of the transformcoded temporal blocks according to a length parameter in the data stream (16).

58. Encoder (10) of any of claims 37 to 57, configured to support different lengths of the temporal blocks and switch between the different lengths of the transform-coded temporal blocks at predetermined borders between consecutive temporal blocks according to a length parameter in the data stream (16).

59. Encoder (10) of any of claims 37 to 58, configured toPCT_FH250111PEP-2025353041.DOCXpredict each of temporal blocks of an even further predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by block prediction from a reference portion (502) of the digital time-varying signal (92) by padding a non-available portion (504) of the reference portion (502) abutting an end (506) of the reference portion (502) using the parametrizable function (408).

60. Encoder (10) for encoding a digital time-varying signal (92) into a data stream (16), configured to encode the digital time-varying signal (92) into the data stream (16) in temporal blocks (140) by predicting each of temporal blocks of a predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by block prediction (500) from a reference portion (502) of the digital time-varying signal (92) by padding a non-available portion (504) of the reference portion (502) abutting an end (506) of an available portion (508) of the reference portion 502 using a parametrizable function (514) which is parametrizable in terms of an end slope (510) and an end offset (512) at an abutting end (506) of the non-available portion (504) abutting the end (506) of the available portion (508), and comprises a monotonic change in slope from the abutting end (506) of the available portion of the reference portion towards an opposite end (516) of the non-available portion (504) so as to start, from the abutting end (506), in a manner changing towards zero slope, and by setting the end slope (510) and end offset (512) at the abutting end (506) of the non-available portion (504) based on the available portion (506) of the reference portion (502).

61. Encoder (10) of claim 60, wherein the encoder (10) is configured so that the parametrizable function (514) is, except for the leading-end slope (510) and the leading-end offset (512), independent from the available portion (508).PCT_FH250111PEP-2025353041.DOCX62. Encoder (10) of any of claims 60 or 61 , wherein the encoder (10) is configured to parametrize the parametrizable function (514) based on the available portion (508) exclusively in terms of the leading-end slope (510) and leading-end offset (512).

63. Encoder (10) of any of claims 60 to 62, wherein the parametrizable function (514) is further parametrizable in terms of a rate of the monotonic change, and the encoder (10) is configured to set the rate depending on a length (520) of the non-available portion.

64. Encoder (10) of any of claims 60 to 63, wherein the encoder (10) is configured to parametrize the parametrizable function (514) depending on a length (520) of the non- available portion such that an opposite-end slope of the parametrizable function (514) at the opposite end (516) of the non-available portion (504) deviates, in magnitude, by less than 0.1 from the zero slope or a trailing-end slope (431) of the parametrizable function (408) at the opposite end (516) of the non-available portion (504) deviates, in magnitude, by less than 0.1 from minus the end slope (410).

65. Encoder (10) of any of claims 60 to 64, wherein the parametrizable function (514) is differentiable on a domain of sample positions between the end (506) and the opposite end (514).

66. Encoder (10) of any of claims 60 to 65, wherein the encoder (10) is configured to, in the setting the end slope (510) and end offset (512) at the abutting end (506) of the non- available portion (504) based on the available portion (508), determine the end offset (512) using a sum of sample values of the available portion (508), and determine the end slope (510) using a sum of addends, wherein each addend depends on a difference between sample values of the available portion (508).PCT_FH250111PEP-2025353041.DOCX67. Encoder (10) of any of claims 60 to 66, wherein the parametrizable function (514) includes a sum of an offset term and a linear sample position dependent term, wherein the encoder (10) is configured to compute the offset term depending on the end offset (512), compute the linear sample position dependent term based on a product comprising a first factor term corresponding to a product between the end slope (510) on the one hand and a difference between a length term which linearly depends on a length (520) of the non-available portion (504) and a sample position value measuring a sample position between the abutting end and the opposite end (516) from the abutting end (506) on the other hand, and a second factor related to the sample position value according to an affine mapping.

68. Encoder (10) of any of claims 60 to 67, configured to encode for the respective temporal block of the predetermined type a sub-sample pointer for locating the reference portion (502) in the digital time-varying signal (92) or in a reference channel of the digital time-varying signal (92).

69. Encoder (10) of any of claims 60 to 68, wherein the temporal blocks (140) of the digital time-varying signal (92) are non-overlapping and the encoder (10) is configured to encode for the respective temporal block of the predetermined type a prediction residual (418) into the data stream (16) by encoding a transform (420) of the prediction residual (418) into the data stream (16) and applying a re-transformation onto the transform (420), and correct a predictor obtained by the block prediction (500) using the prediction residual (418).

70. Encoder (10) of any of claims 60 to 69, wherein the digital time-varying signal (92) is obtained by at least one of Electrocardiography, Electroencephalography, Electromyography or seismic measurement, and / orPCT_FH250111PEP-2025353041.DOCXwherein the digital time-varying signal (92) is a bio-physiological waveform data such as an electroencephalography (EEG) signal, an electrocardiogram (ECG), or an electromyography (EMG) signal, or is a seismic waveform signal.

71. Encoder (10) of any of claims 60 to 70, configured to support different lengths of temporal blocks and set a length of the transformcoded temporal blocks according to a length parameter in the data stream (16).

72. Encoder (10) of any of claims 60 to 71 , configured to support different lengths of the temporal blocks and switch between the different lengths of the transform-coded temporal blocks at predetermined borders between consecutive temporal blocks according to a length parameter in the data stream (16).

73. Method (9001) for decoding a digital time-varying signal (92) from a data stream (16), the method (9001) comprising: decoding (9002) the digital time-varying signal (92) from the data stream (16) in temporal blocks (140) by decoding each of temporal blocks of a predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by extrapolation from a previously decoded portion (404) of the digital time-varying signal (92), which precedes a leading end (406) of the respective temporal block of the predetermined type, by predicting (9003) the respective temporal block of the predetermined type using (9004) a parametrizable function (408) which is parametrizable in terms of a leading-end slope (410) and a leadingend offset (412) at the leading end (406) of the respective temporal block of the predetermined type, and comprises a monotonic change in slope from the leading end (406) of the respective temporal block of the predetermined type towards a trailing end (414) of the respective temporal block of the predetermined type so as to start, from the leading end, in a manner changing towards zero slope, andPCT_FH250111PEP-2025353041.DOCXby setting (416) the leading-end slope (410) and the leading-end offset (412) at the leading end (406) of the respective temporal block of the predetermined type based on the previously decoded portion (404); decoding (9005) a prediction residual (418) from the data stream (16), and reconstructing (9006) the respective temporal block of the predetermined type by correcting (422) the parametrizable function (408) using the prediction residual (418).

74. Method (10007) for decoding a digital time-varying signal (92) from a data stream (16), the method (10007) comprising: decoding (10008) the digital time-varying signal (92) from the data stream (16) in temporal blocks (140) by predicting each of temporal blocks of a predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by block prediction (500) from a reference portion (502) of the digital time-varying signal (92) by padding a non-available portion (504) of the reference portion (502) abutting an end (506) of an available portion (508) of the reference portion (502) using (10009) a parametrizable function (514) which is parametrizable in terms of an end slope (510) and an end offset (512) at an abutting end of the non-available portion abutting the end (506) of the available portion (508), and comprises a monotonic change in slope from the abutting end of the available portion of the reference portion towards an opposite end (516) of the non-available portion (504) so as to start, from the abutting end, in a manner changing towards zero slope, and by setting the end slope (510) and end offset (512) at the abutting end (506) of the non-available portion based on the available portion of the reference portion.

75. Method (11010) for encoding a digital time-varying signal (92) into a data stream (16), the method (11010) comprising:PCT_FH250111PEP-2025353041.DOCXencoding (11011) the digital time-varying signal (92) into the data stream (16) in temporal blocks (140) by encoding each of temporal blocks of a predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by extrapolation from a previously encoded portion (404) of the digital time-varying signal (92), which precedes a leading end (406) of the respective temporal block of the predetermined type, by predicting (11012) the respective temporal block of the predetermined type using (11013) a parametrizable function (408) which is parametrizable in terms of a leading-end slope (410) and a leadingend offset (412) at the leading end (406) of the respective temporal block of the predetermined type, and comprises a monotonic change in slope from the leading end (406) of the respective temporal block of the predetermined type towards a trailing end (414) of the respective temporal block of the predetermined type so as to start, from the leading end, in a manner changing towards zero slope, and by setting (416) the leading-end slope (410) and the leading-end offset (412) at the leading end (406) of the respective temporal block of the predetermined type based on the previously encoded portion (404); encoding (11014) a prediction residual (418) into the data stream (16), and wherein the respective temporal block of the predetermined type is reconstructable (424) by correcting (422) the parametrizable function (408) using the prediction residual (418).

76. Method (12015) for encoding a digital time-varying signal (92) into a data stream (16), the method (12015) comprising: encoding (12016) the digital time-varying signal (92) into the data stream (16) in temporal blocks (140) by predicting each of temporal blocks of a predetermined type among the temporal blocks (140) of the digital time-varying signal (92) by block prediction (500) from aPCT_FH250111PEP-2025353041.DOCXreference portion (502) of the digital time-varying signal (92) by padding a non-available portion (504) of the reference portion (502) abutting an end (506) of an available portion (508) of the reference portion (502) using (12017) a parametrizable function (514) which is parametrizable in terms of an end slope (510) and an end offset (512) at an abutting end of the non-available portion abutting the end (506) of the available portion (508), and comprises a monotonic change in slope from the abutting end of the available portion of the reference portion towards an opposite end (516) of the non-available portion (504) so as to start, from the abutting end, in a manner changing towards zero slope, and by setting the end slope and end offset at the end of the non-available portion based on the available portion of the reference portion.

77. Data stream (16) encoded using the method according to any of the claims 75 or 76.

78. A computer program for implementing the method of any one of claims 73 to 76 when being executed on a computer or signal processor.PCT_FH250111PEP-2025353041.DOCX

Images