The present application is concerned with an intra-coding concept for use in a block-based codec such as, for example, a hybrid video codec.

Given a certain block, intra prediction is carried out in HEVC by extrapolating the decoded boundary samples of the neighboring blocks following certain patterns, namely 33 angular modes and a DC and a planar modes [1] The one intra prediction mode that minimizes the rate-distortion cost is then signaled to the decoder. Despite the known codecs supporting many Intra Prediction Modes (IPMs), the intra prediction achieved thereby is still subject of development to find better intra predictors leading to higher coding efficiency. This does not only pertain to HEVC but also to other block-based codecs using intra-prediction. Finding a set of intra-prediction modes which are suitable for efficiently coding the inner of blocks requires taking into account the overhead for signaling the intra-prediction mode in terms of signaling overhead and the resulting quality of the predictor obtained by these intraprediction modes due to the fact that a more accurate predictor reduces the prediction residual, thereby reducing the signaling overhead associated with coding the prediction residual. In order to keep the signaling overhead associated with the intra-prediction modes low, intra-predicted blocks should be large, i.e. the granularity at which the intra prediction mode is signaled should be kept coarse, but on the other hand, spatial prediction of larger blocks tends to be less accurate owing to a higher mean sample distance of the samples in the inner of the intra-predicted block, i.e. the ones to be predicted, to the already decoded/encoded samples neighboring this block, i.e. the reference samples. HEVC alleviates this catch-22 a little bit by allowing the transform residual blocks to inherit the intra-prediction mode of their corresponding coding unit relative to which the transform residual blocks form leaf blocks into which the coding unit is sub-divided by multi-tree subdivisioning. However, this still requires signaling overhead for signaling from encoder to decoder the sub-partitioning of respective intra-coded coding units into the transform blocks.

A newly developing intra coding concept is presented by the Intra Sub-Partitions (ISP) coding mode in the newly developing Versatile Video Coding (WC) standard, but here, implementation efficiency improvements would be required.

Thus, it would be favorable to have a concept at hand which further increases the implementation efficiency of intra-coding at comparable coding efficiency.

Accordingly, it is an object of the present invention to provide a concept for intra-coding which is more efficient.

This object is achieved by the subject-matter of the independent claims of the present application.

Advantageous aspects of the present invention are the subject of dependent claims.

In accordance with a first aspect of the present invention, the inventors of the present application realized that one problem encountered when using sub-partitioning in connection with intra-coding is, that the number of sub-partition per block, for which prediction is done individually, should be limited considering the resulting sub-partition size, a wanted minimum throughput of, for example, 16 samples per cycle, and/or a minimum width of coding advance per coding cycle such as a minimum of 4 sample wide advance per prediction. These thoughts led to the idea of construing a flag controlled intra-prediction mode/decision for an intra-coded (predetermined) block which leads to a partitioning of this predetermined block in terms of prediction residual transformation, while the subpartitioning in terms of intra-prediction, i.e. whether the predetermined block is intra-predicted in an integral manner (all at once), or whether the transform partitions are used for sequential and partition-wise intra-prediction with intermediate usage of the prediction residual and the correction of the just intra-predicted sub-partition using the same for intra-predicting the next sub-partition as well, or whether groups of transform partitions for such prediction sub-partitions, may be freely implemented as needed, such as rendering the latter choice dependent on the block size with the aim, for instance, to avoid intra-predictions resulting into too less samples per intra-prediction performed or too less intra-prediction width advance, for instance. Note that the coding and decoding of the transform partitions may be done independent among the transform partitions, i.e. that same may be coded/decoded in parallel, thereby not raising any minimum sample per cycle or width advance per cycle issues. This enables a partitioning of an intra-predicted block into partitions with, for instance, less samples than 16, since more than one partition can be intra-predicted and reconstructed in the same cycle. It is advantageous, if all sub-partitions encoded or decoded in the same cycle comprise together at least 16 samples. Again, according to a variant described herein, the coding coded supports many block sizes, and

depending on the size of the predetermined intra-predicted block and/or depending on its width and/or height, decoder and encoder set the partitioning in terms of prediction to result into one of the following options:

1 ) intra-prediction block-globally, i.e. in toto (at once) or as a whole (in other words, e.g., predicting the whole predetermined block at once or, in even other words, predicting all samples within the predetermined block based on neighboring samples which are located outside that predetermined block exclusively, and processing transform partitions of the predetermined block independently (i.e. the transform is executed region-wise within each transform partition)), and/or

2) sequential intra-prediction in units of transform partitions which, then, also act as prediction sub-partitions (in other words, e.g., predicting a transform-partition, coding/decoding the prediction residual for that transform partition with obtaining reconstructed samples within that transform partition and then predicting the next transform partition in the predetermined block using the reconstructed samples obtained for the former transform partition, coding/decoding the prediction residual for this next transform partition and so forth), and/or

3) sequential intra-prediction in units of groups of transform partitions (with each transform partition belonging exactly to one partition group) (in other words, e.g., predicting a group of transform-partitions, i.e. a prediction sub-partition, based on neighboring samples which are located outside that prediction sub-partition exclusively, coding/decoding the prediction residual for that prediction sub-partition with obtaining reconstructed samples within that prediction sub-partition in units of the transform partitions within that prediction sub-partition (i.e. the transform is executed region-wise within each transform partition) and then predicting the next group of transform partitions, i.e. the next prediction sub-partition, in the predetermined block using reconstructed samples including ones obtained for the former prediction sub-partition, but excluding ones located inside this next prediction sub-partition, coding/decoding the prediction residual for this next prediction sub-partition in units of transform partitions and so forth).

Accordingly, in accordance with a first aspect of the present application, a decoder for block-based decoding of a picture from a data stream, is configured to decode an intra-coding mode for a predetermined block of the picture from the data stream. The decoder is configured to decode a partition dimension flag for the predetermined block of the picture from the data stream and set a partition dimension depending on the partition dimension flag to be horizontal or vertical. In other words the partition dimension flag indicates whether the partition dimension is horizontal or vertical. The decoder is configured to partition, along the predetermined dimension (i.e. along the partition dimension), the predetermined block into transform partitions which are as wide as the predetermined block perpendicular to the predetermined dimension. The transform partitions can be associated with vertically stacked horizontal blocks, if the partition dimension is vertical, and the transform partitions can be associated with vertical blocks arranged horizontally side by side, if the partition dimension is horizontal. For each transform partition, the decoder is configured to decode a transform of a prediction residual from the data stream. Furthermore the decoder is configured to intra-predicting the predetermined block depending on one or more already reconstructed samples neighboring the predetermined block in a manner depending on the intra-coding mode to obtain a predictor for the predetermined block and reconstructing the predetermined block by correcting the predictor within each transform partition using the transform of the prediction residual decoded for the respective transform partition.

According to a first option the decoder is configured to sequentially intra-predict for one transform partition after the other a predictor and reconstruct the transform partition by correcting the predictor within each transform partition using the transform of the prediction residual decoded for the respective transform partition.

According to a second, alternative option, the decoder is configured to intra-predict for each transform partition a predictor and decode for each transform partition a transform of a prediction residual from the data stream. Then the decoder is configured to reconstruct the predetermined block by correcting the predictors using the transform of the prediction residual decoded for the respective transform partition. Thus first all predictors are intra-predicted and all transforms of prediction residuals are decoded and then all transform partitions are reconstructed by correcting the predictor within each transform partition using the transform of the prediction residual decoded for the respective transform partition. Thus, for example, in one step all predictors are corrected.

In contrast, according to the first option, one transform partition after the other is reconstructed. In other words, according to the first option, for a current transform partition a predictor is intra-predicted and corrected and afterwards for a subsequent transform partition a new predictor is intra predicted and corrected.

According to a third, alternative option, the decoder is configured to intra-predict the whole predetermined block in one step to obtain a prediction signal (i.e., a predictor) and divide this prediction signal, e.g., into predictors for transform partitions of the predetermined block. According to an embodiment, each predictor is associated with a different transform partition. The transform partitions are, for example, processed independently by the decoder. Thus, for example, the decoder is configured to decode for each transform partition a transform of a prediction residual from the data stream and reconstruct the predetermined block by correcting the predictors using the transform of the prediction residual decoded for the respective transform partition. Alternatively this is not performed for the whole block at once, but for sub-partitions of the predetermined block which can be divided further into transform partitions. In this case, for example, the decoder is configured to intra-predict a sub-partition of the predetermined block in one step to obtain a prediction signal (i.e., a predictor) and divide this prediction signal, e.g., into predictors for transform partitions of the sub-partition of the predetermined block.

According to an embodiment, the decoder is configured to divide the predetermined block depending on the block size into sub-partitions, wherein a minimum prediction width of 4 is established to reduce the hardware implementation complexity. The invention is not limited by the following examples for different partitioning’s performed by the decoder. It is clear, that also other sub-partitions and/or transform partitions can be achieved by the decoder.

• 4x4 blocks (example 1)

o Hor. Split: One 4x4 PU (prediction unit) and 4 independent 4x1 TUs (transform unit).

o Ver. Split: One 4x4 PU and 4 independent 1 x4 TUs.

In other words the whole 4x4 block is predicted at once and then divided into 4 transform partitions to be processed independently.

• 8x4 blocks (example 2)

o Hor. Split: Two 8x2 PUs and four 8x 1 TUs. The second PU is predicted using the reconstructed samples of the second TU.

o Ver. Split: Two 4x4 PUs and four 2x4 TUs. The second PU is predicted using the reconstructed samples of the second TU.

In other words the 8x4 block is divided into two sub-partitions (i.e. the PUs) and each sub-partition is divided into 2 transform partitions to be processed independently.

• 4x8 blocks (example 3)

o Hor. Split: Two 4x4 PUs and four 4x2 TUs. The second PU is predicted using the reconstructed samples of the second TU.

o Ver. Split: One 4x8 PUs and four independent 1 x8 TUs.

In other words, at the horizontal split, the 4x8 block is divided into two subpartitions (i.e. the PUs) and each sub-partition is divided into 2 transform partitions to be processed independently and, at the vertical split, the whole 4x8 block is predicted at once and then divided into 4 transform partitions to be processed independently.

• 4x8 blocks (example 3’; an alternative to example 3)

o Hor. Split (no modification compared to treatment of sub-partitions in terms of both prediction as well as transform residual coding/decoding): Two 4x4 PUs are used which concurrently form two 4x4 TUs. The second PU is predicted using the reconstructed samples of the first PU.

o Ver. Split (modified): One 4x8 PUs and two independent 2x8 Tus.

In other words, at the horizontal split, the 4x8 block Is divided into two subpartitions (i.e. the PUs) and each sub-partition is ends-up into 1 transform partition and, at the vertical split, the whole 4x8 block is predicted at once and then divided into 2 transform partitions to be processed independently.

• 4xM blocks (example 4)

The whole 4xM block is predicted at once and then divided into 4 1xM transform partitions to be processed independently.

• 4xM blocks (example 4’; with M>8)

o Hor. Split (no modification compared to treatment of sub-partitions in terms of both prediction as well as transform residual coding/decoding): The 4xM block is predicted in four PUs of 4x(M/4) each of which concurrently Is one of the four transform partitions

o Vert. Split: The whole 4xM block is predicted at once and then divided into 4 1xM transform partitions to be processed independently.

8xN blocks (example 5)

The 8xN block is divided into two 4xN sub-partitions which can be further divided into 4 1xN transform partitions.

• 8xN blocks (example 5’; with N>4)

o Hor. Split (no modification compared to treatment of sub-partitions in terms of both prediction as well as transform residual coding/decoding): The 8xN block is divided into four 8x(N/4) sub-partitions (for sake of prediction as well as transform residual coding/decoding) .

o Vert. Split: The 8xN block is divided into two 4xN sub-partitions (for sake of prediction) which can be further divided into 2 2xN transform partitions.

The examples outlined above relate to different block sizes and may apply to a codec in accordance with corresponding embodiments (i.e. to decoder and encoder, respectively) individually, altogether or combinations of two or more of them may apply. As may be seen, in accordance with an embodiment, for at least one predetermined block size (compare examples 3 to 5, for instance), there may be a difference in the decision regarding how to choose among the aforementioned options 1 to 3 (among two out of 1 to 3) depending on the split direction: While one option is selected for a horizontal split such as option 2, where each TU is also a PU and the number of PUs and the number of TUs are, thus, the same, a different option may be chosen for a vertical split such as option 1 , where the whole block acts as PU, but is divided into several Tus and the number of PUs and the number of TUs, thus, differ, or option 3, where the predetermined block is divided into PUs, each of which is further divided into TUs and the number of PUs and the number of TUs, thus, differ. Additionally or alternatively, for another block size (compare example 2), this decision may end-up into the same option irrespective of the split direction. The just-mentioned dependency of the selection among the options on the split direction may, thus, come in addition to the already mentioned block size direction, but naturally it could apply without the latter.

An embodiment according to this invention is related to an encoder for block-based encoding of a picture into a data stream, configured to encode an intra-coding mode for a predetermined block of the picture into the data stream. The encoder is configured to encode a partition dimension flag for the predetermined block of the picture into the data stream which signals that a partition dimension is to be set to be horizontal or vertical. In other words the partition dimension flag indicates whether the partition dimension is horizontal or vertical. The encoder is configured to partition, along the predetermined

dimension (i.e. along the partition dimension), the predetermined block into transform partitions which are as wide as the predetermined block perpendicular to predetermined dimension. The transform partitions can be associated with vertically stacked horizontal blocks, if the partition dimension is vertical, and the transform partitions can be associated with vertical blocks arranged horizontally side by side, if the partition dimension is horizontal. Furthermore the encoder is configured to intra-predicting the predetermined block depending on one or more already reconstructed samples neighboring the predetermined block in a manner depending on the intra-coding mode to obtain a predictor for the predetermined block. For each transform partition, the encoder is configured to encode a transform of a prediction residual into the data stream, so that the predetermined block is reconstructible by correcting the predictor within each transform partition using the transform of the prediction residual encoded for the respective transform partition.

The encoder as described above as well as methods performed by any herein described encoders and decoders and a data stream generated by the method performed by any of the herein described encoders are based on the same considerations as the above-described decoder. The methods can, by the way, be completed with all features and functionalities, which are also described with regard to the decoder and/or encoder.

Preferred embodiments of the present application are described below with respect to the figures, among which:

Fig. 1 shows a block diagram of an apparatus for predictively coding a picture as an example for an encoder where an ISP concept could be implemented;

Fig. 2 shows a block diagram of an apparatus for predictively decoding a picture, which fits to the apparatus of Fig. 1 , as an example for decoder where an ISP concept could be implemented;

Fig. 3 shows a schematic diagram illustrating an example for a relationship between the prediction residual signal, the prediction signal and the reconstructed signal so as to illustrate possibilities of setting subdivisions for coding mode selection, transform selection and transform performance, respectively;

Fig. 4 shows a schematic diagram illustrating a partitioning treatment of an intra-coded block in accordance with a ISP variant allowing for a selection between different partitioning dimensions, i.e., horizontal and vertical splits;

Fig. 5 shows a schematic diagram illustrating the sequential processing of the partitions of an ISP-coded block;

Fig. 6 shows a schematic diagram illustrating the predicted derivation of filling process for a partition;

Fig. 7 illustrates examples for ISP blocks split according to horizontal and vertical split mode, respectively, and having two different intra-prediction modes associated therewith, respectively, in order to illustrate a possibility of rendering dependent determination of the partition order on the intra-prediction mode associated with the intra-predicted block;

Fig. 8 shows a schematic diagram illustrating a possible signalization spent for an intra- predicted block 80 treated using the partition option;

Fig. 9 shows a schematic diagram illustrating a possible way of transmitting the prediction residual of a partition;

Fig. 10 shows a schematic diagram illustrating the partial sum determination for coding costs involved with the partitioning of intra-prediction mode in order to be able to abort the test when it is clear that same will not become better than any of the normal intra-prediction modes; and

Fig. 11 shows a flowchart of a mode or operation of the encoder in order to perform the partition mode testing; and

Fig. 12 shows a schematic diagram illustrating a decoder for block-based decoding of a picture with an inventive ISP concept implemented;

Fig. 13 shows a schematic diagram illustrating a usage of a last position syntax element;

Fig. 14 shows a schematic diagram illustrating an encoder for block-based decoding of a picture with an inventive ISP concept implemented;

Fig. 15a-15d show schematic diagrams illustrating intra-prediction of individual transform partitions of a vertically split 4x4 block;

Fig. 16a-16d show schematic diagrams illustrating intra-prediction of individual transform partitions of a horizontally split 4x4 block;

Fig. 17 shows schematic diagrams illustrating intra-prediction of individual transform partitions of a vertically split 4x8 block;

Fig. 18 shows schematic diagrams illustrating intra-prediction of individual transform partitions of a horizontally split 8x4 block;

Fig.19a shows a vertical split of a 4xM block (M>8) in the ISP design in the WC Draft 5

(left) and in the proposed version (right); and

Fig.19b shows a vertical split of a 8xN block (N>4) in the ISP design in the WC Draft 5

(left) and in the proposed version (right).

The drawings are not necessarily to scale, emphasis instead generally being placed upon illustrating the principles of the invention.

The following description of the figures starts with a presentation of a description of encoder and decoder of a block-based predictive codec for coding pictures of a video in order to form an example for a coding framework into which embodiments for an intra-prediction codec may be built in. The former encoder and decoder are described with respect to Figs 1 to 3. Thereinafter the description of variants of the ISP concept are presented along with a description as to how such concepts could be built into the encoder and decoder of Figs. 1 and 2, respectively, although the concepts described with the subsequent Figs. 4 and following, may also be used to form encoders and decoders not operating according to the coding framework underlying the encoder and decoder of Figs. 1 and 2. Later on, embodiments are described which make use of ISP, but are improved in terms of implementation efficiency. Also, embodiments are described which make use of a variant of partition-based intra coding.

Fig. 1 shows an apparatus for predictively coding a picture 12 into a data stream 14 using exemplarily using transform-based residual coding. The apparatus, or encoder, is indicated using reference sign 10. Fig. 2 shows a corresponding decoder 20, i.e. an apparatus 20 configured to predictively decode the picture 12’ from the data stream 14 also using transform-based residual decoding, wherein the apostrophe has been used to indicate that the picture 12’ as reconstructed by decoder 20 deviates from picture 12 originally encoded by apparatus 10 in terms of coding loss introduced by a quantization of the prediction residual signal. Fig. 1 and Fig. 2 exemplarily use transform based prediction residual coding, although embodiments of the present application are not restricted to this kind of prediction residual coding. This is true for other details described with respect to Fig. 1 and 2, too, as will be outlined hereinafter.

The encoder 10 is configured to subject the prediction residual signal to spatial-to-spectral transformation and to encode the prediction residual signal, thus obtained, into the data stream 14. Likewise, the decoder 20 is configured to decode the prediction residual signal from the data stream 14 and subject the prediction residual signal thus obtained to spectral-to-spatial transformation.

Internally, the encoder 10 may comprise a prediction residual signal former 22 which generates a prediction residual 24 so as to measure a deviation of a prediction signal 26 from the original signal, i.e. the picture 12. The prediction residual signal former 22 may, for instance, be a subtractor which subtracts the prediction signal from the original signal, i.e. picture 12. The encoder 10 then further comprises a transformer 28 which subjects the prediction residual signal 24 to a spatial-to-spectral transformation to obtain a spectral-domain prediction residual signal 24’ which is then subject to quantization by a quantizer 32, also comprised by encoder 10. The thus quantized prediction residual signal 24” is coded into bitstream 14. To this end, encoder 10 may optionally comprise an entropy coder 34 which entropy codes the prediction residual signal as transformed and quantized into data stream 14. The prediction residual 24 is generated by a prediction stage 36 of encoder 10 on the basis of the prediction residual signal 24” decoded into, and decodable from, data stream 14. To this end, the prediction stage 36 may internally, as is shown in Fig. 1 , comprise a dequantizer 38 which dequantizes prediction residual signal 24” so as to gain spectral-domain prediction residual signal 24”’, which corresponds to signal 24' except for quantization loss, followed by an inverse transformer 40 which subjects the latter prediction residual signal 24”’ to an inverse transformation, i.e. a spectral-to-spatial transformation, to obtain prediction residual signal 24””, which corresponds to the original prediction residual signal 24 except for quantization loss. A combiner 42 of the prediction stage 36 then recombines, such as by addition, the prediction signal 26 and the prediction residual signal 24”” so as to obtain a reconstructed signal 46, i.e. a reconstruction of the original signal 12. Reconstructed signal 46 may correspond to signal 12’. A prediction module 44 of prediction stage 36 then generates the prediction signal 26 on the basis of signal 46 by using, for instance, spatial prediction, i.e. intra prediction, and/or temporal prediction, i.e. inter prediction.

Likewise, decoder 20 may be internally composed of components corresponding to, and interconnected in a manner corresponding to, prediction stage 36. In particular, entropy decoder 50 of decoder 20 may entropy decode the quantized spectral-domain prediction residual signal 24” from the data stream, whereupon dequantizer 52, inverse transformer 54, combiner 56 and prediction module 58, interconnected and cooperating in the manner described above with respect to the modules of prediction stage 36, recover the reconstructed signal on the basis of prediction residual signal 24” so that, as shown in Fig. 2, the output of combiner 56 results in the reconstructed signal, namely picture 12’.

Although not specifically described above, it is readily clear that the encoder 10 may set some coding parameters including, for instance, prediction modes, motion parameters and the like, according to some optimization scheme such as, for instance, in a manner optimizing some rate and distortion related criterion, i.e. coding cost. For example, encoder 10 and decoder 20 and the corresponding modules 44, 58, respectively, may support different prediction modes such as intra-coding modes and inter-coding modes. The granularity at which encoder and decoder switch between these prediction mode types may correspond to a subdivision of picture 12 and 12’, respectively, into coding segments or coding blocks. In units of these coding segments, for instance, the picture may be subdivided into blocks being intra-coded and blocks being inter-coded. Intra-coded blocks are predicted on the basis of a spatial, already coded/decoded neighborhood of the respective block as is outlined in more detail below. Several intra-coding modes may exist and be selected for a respective intra-coded segment including directional or angular intracoding modes according to which the respective segment is filled by extrapolating the sample values of the neighborhood along a certain direction which is specific for the respective directional intra-coding mode, into the respective intra-coded segment. The intracoding modes may, or instance, also comprise one or more further modes such as a DC coding mode, according to which the prediction for the respective intra-coded block assigns a DC value to all samples within the respective intra-coded segment, and/or a planar intracoding mode according to which the prediction of the respective block is approximated or determined to be a spatial distribution of sample values described by a two-dimensional linear function over the sample positions of the respective intra-coded block with driving tilt and offset of the plane defined by the two-dimensional linear function on the basis of the neighboring samples. Compared thereto, inter-coded blocks may be predicted, for instance, temporally. For inter-coded blocks, motion vectors may be signaled within the data stream, the motion vectors indicating the spatial displacement of the portion of a previously coded picture of the video to which picture 12 belongs, at which the previously coded/decoded picture is sampled in order to obtain the prediction signal for the respective inter-coded block. This means, in addition to the residual signal coding comprised by data stream 14, such as the entropy-coded transform coefficient levels representing the quantized spectral-domain prediction residual signal 24", data stream 14 may have encoded thereinto coding mode parameters for assigning the coding modes to the various blocks, prediction parameters for some of the blocks, such as motion parameters for inter-coded segments, and optional further parameters such as parameters controlling and signaling the subdivision of picture 12 and 12’, respectively, into the segments. The decoder 20 uses these parameters to subdivide the picture in the same manner as the encoder did, to assign the same prediction modes to the segments, and to perform the same prediction to result in the same prediction signal.

Fig. 3 illustrates the relationship between the reconstructed signal, i.e. the reconstructed picture 12’, on the one hand, and the combination of the prediction residual signal 24”” as signaled in the data stream, and the prediction signal 26, on the other hand. As already denoted above, the combination may be an addition. The prediction signal 26 is illustrated in Fig. 3 as a subdivision of the picture area into intra-coded blocks which are illustratively indicated using hatching, and inter-coded blocks which are illustratively indicated not-hatched. The subdivision may be any subdivision, such as a regular subdivision of the picture area into rows and columns of blocks or blocks, or a multi-tree subdivision of picture 12 into leaf blocks of varying size, such as a quadtree subdivision or the like, into blocks, wherein a mixture thereof is illustrated in Fig. 3 where the picture area is first subdivided into rows and columns of tree-root blocks which are then further subdivided in accordance with a recursive multi-tree subdivisioning. Again, data stream 14 may have an intra-coding mode coded thereinto for intra-coded blocks 80, which assigns one of several supported intra-coding modes to the respective intra-coded block 80. Further details are described below. For inter-coded blocks 82, the data stream 14 may have one or more motion

parameters coded thereinto. Generally speaking, inter-coded blocks 82 are not restricted to being temporally coded. Alternatively, inter-coded blocks 82 may be any block predicted from previously coded portions beyond the current picture 12 itself, such as previously coded pictures of a video to which picture 12 belongs, or picture of another view or an hierarchically lower layer in the case of encoder and decoder being scalable encoders and decoders, respectively. The prediction residual signal 24”” in Fig. 3 is also illustrated as a subdivision of the picture area into blocks 84. These blocks might be called transform blocks in order to distinguish same from the coding blocks 80 and 82. In effect, Fig. 3 illustrates that encoder 10 and decoder 20 may use two different subdivisions of picture 12 and picture 12’, respectively, into blocks, namely one subdivisioning into coding blocks 80 and 82, respectively, and another subdivision into blocks 84. Both subdivisions might be the same, i.e. each coding block 80 and 82, may concurrently form a transform block 84, but Fig. 3 illustrates the case where, for instance, a subdivision into transform blocks 84 forms an extension of the subdivision into coding blocks 80/82 so that any border between two blocks of blocks 80 and 82 overlays a border between two blocks 84, or alternatively speaking each block 80/82 either coincides with one of the transform blocks 84 or coincides with a cluster of transform blocks 84. However, the subdivisions may also be determined or selected independent from each other so that transform blocks 84 could alternatively cross block borders between blocks 80/82. As far as the subdivision into transform blocks 84 is concerned, similar statements are thus true as those brought forward with respect to the subdivision into blocks 80/82, i.e. the blocks 84 may be the result of a regular subdivision of picture area into blocks/blocks, arranged in rows and columns, the result of a recursive multi-tree subdivisioning of the picture area, or a combination thereof or any other sort of blockation. Just as an aside, it is noted that blocks 80, 82 and 84 are not restricted to being of quadratic, rectangular or any other shape.

Fig. 3 illustrates that the combination of the prediction signal 26 and the prediction residual signal 24”” directly results in the reconstructed signal 12'. However, it should be noted that more than one prediction signal 26 may be combined with the prediction residual signal 24”” to result into picture 12’ in accordance with alternative embodiments.

In Fig. 3, the transform segments 84 shall have the following significance. Transformer 28 and inverse transformer 54 perform their transformations in units of these transform segments 84. For instance, many codecs use some sort of DST or DCT for all transform blocks 84. Some codecs allow for skipping the transformation so that, for some of the transform segments 84, the prediction residual signal is coded in in the spatial domain

directly. However, in accordance with embodiments described below, encoder 10 and decoder 20 are configured in such a manner that they support several transforms. For example, the transforms supported by encoder 10 and decoder 20 could comprise:

o DCT-II (or DCT-III), where DCT stands for Discrete Cosine T ran storm

o DST-IV, where DST stands for Discrete Sine Transform

o DCT-IV

o DST-VII

o Identity Transformation (IT)

Naturally, while transformer 28 would support all of the forward transform versions of these transforms, the decoder 20 or inverse transformer 54 would support the corresponding backward or inverse versions thereof:

o Inverse DCT-II (or inverse DCT-III)

o Inverse DST-IV

o Inverse DCT-IV

o Inverse DST-VII

o Identity Transformation (IT)

The subsequent description provides more details on which transforms could be supported by encoder 10 and decoder 20. In any case, it should be noted that the set of supported transforms may comprise merely on transform such as one spectral-to-spatial or spatial-to-spectral transform.

As already outlined above, Figs. 1 - 3 have been presented as an example where the intraprediction concepts described further below may be implemented. Insofar, the encoder and decoder of Figs. 1 and 2, respectively, represent possible implementations of the encoders and decoders described herein below. As will be outlined in more detail below, when having the subsequently explained embodiments for intra-prediction according to the present application built into the encoder and decoder of Fig. 1 and 2, the encoder of Fig. 1 and the decoder of Fig. 2 support, at least as one option, to process an intra-predicted block 80 in the manner outlined in more detail below. Thus, the embodiments described hereinafter refer to an encoder which equals the encoder 10 of Fig. 1 which treats intra-coded blocks 80 in the manner outlined in more detail below and the same applies with respect to the decoder of Fig. 2 which, thus, represents an example for a decoder according to an embodiment where intra-coded blocks are treated in the manner outlined in more detail below. Fig 1 and 2 are, however, only specific examples. An encoder according to embodiments of the present application may, however, perform block-based encoding of a picture 12 using the concept outlined in more detail below and being different from the encoder of Fig. 1 such as, for instance, in that same is no video encoder, in that same does not support inter-prediction, or in that the sub-division into blocks 80 is performed in a manner different than exemplified in Fig. 3, or - depending on the embodiment - even in that this encoder does not use transform prediction residual coding with coding the prediction residual, for instance, in spatial domain directly instead. Likewise, decoders according to embodiments of the present application may perform block-based decoding of picture 12’ from data stream 14 using the intra-prediction coding concept further outlined below, but may differ, for instance, from the decoder 20 of Fig. 2 in that same is no video decoder, but a still picture decoder, in that same does not support intra-prediction, or in that same sub-divides picture 12’ into blocks in a manner different than described with respect to Fig. 3 and/or in that same does not derive the prediction residual from the data stream 14 in transform domain, but in spatial domain, for instance.

Having said this, the following description first concentrates on the description of ISP based intra-prediction. According to ISP intra-prediction, intra-predicted blocks such as block 80 in Fig. 4 are allowed to be split into one-dimensional horizontal partitions or one-dimensional vertical partitions. The availability of treating blocks in that manner may be offered for intra-predicted blocks 80 of any size or be restricted to blocks 80 within a predefined range of blocks sizes only such as blocks greater than a certain size. One-dimensional’’ refers to the fact that - when related to partitions being the result of partitioning - the partitions are merely one sample wide along partition dimension. One-dimensionality of the partitioning modes discussed herein, however, refers to the fact that the partitioning takes place along a certain dimension, with the resulting partitions being like stripes extending completely over the block in a direction transverse to the partitioning direction. See, for instance, Fig. 4. Fig. 4 shows the intra-predicted block 80, i.e. , the block to be decoded or the block to be encoded, at the left hand side. It has dimension W x H. That is, it is a W x H dimensional block where H is the height and W is the width of block 80 measured in samples, respectively. According to Fig. 4, there are two split or partitioning options available, namely a horizontal split 100 according to which block 80 is split or partitioned into a number of partitions 102i, 1022, 1023 and 1024 along a vertical axis, i.e., the partition dimension 104. According to the example of Fig. 4, which is the example applied in the following description, each partition 102i to 102 is one sample wide as illustrated by the double headed arrow 106 so that the number of partitions 102i to 102 resulting from block 80 equals H, i.e., the height of block 80 in units of samples 108 of block 80, but it should be clear that the partitioning may be performed by encoder and decoder also according to a different manner agreed between encoder and decoder such as, for instance, a partitioning of block 80 along

dimension 104 may be done in a manner leading to a predefined number of partitions 102 , the predefined number being greater than two, for instance, or a mixture thereof, with distributing the size of block 80 along the partition dimension evenly onto the predefined number of partitions.

The other coding option depicted in Fig. 4 and indicated by reference sign 1 10 corresponds to splitting block 80 into vertical partitions 1 12i, 1122, ... 1 12s. That is, according to option 1 10, block 80 is partitioned into partitions 1 12, along the horizontal axis, i.e., a horizontal partition dimension 104. In case of option 100, each partition 102, is as wide as block 80, i.e., has the block’s width W, whereas the partitions 1 12, adopt the height H of block 80, i.e. have height H. Summarizing, in a manner similar to the description of option 100, the vertical split 1 10 may split block 80 into a numberW of partitions 1 12, with W denoting the horizontal width of block 80 measured in samples 108 so that each partition 1 12, is one sample wide in horizontal direction, wherein, however, the partitioning according to option 1 10 may also be performed in another manner agreed between encoder and decoder.

Thus, according to Fig. 4, the encoder is free to partition block 80 into H Wx1 partitions 102, according to the horizontal split option 100 orW 1xH partitions 1 12, according to the vertical split option 1 10, and the split option chosen by the encoder for block 80 may be signaled in the data stream 14 for block 80 such as, for instance, by way of a corresponding partition dimension flag 1 14 in data stream 14. It should be clear, however, that embodiments of the present application also cover encoders and decoders which, by default, merely use one of options 100 and 1 10 without the need for flag 1 14 in the data stream. Even further, flag 1 14 may be conveyed in data stream 80 in other examples depending on the intra-coding mode 1 16 signaled in the data stream 14 for block 80 from encoder to decoder. The intra-coding mode may, as outlined above, indicate one out of a set of available/supported intra-coding modes including, for instance, angular modes and, optionally, one or more non-angular modes such as a DC mode or a planar mode. That is, flag 1 14 may, in accordance with alternative embodiments not further discussed hereinafter, be conveyed in data stream 14 in a manner conditionally depending on the intra-coding mode 1 16. According to the embodiments described hereinafter, flag 1 14 is present in data stream 14 for block 80 independent on the intra-coding mode 1 16 signaled for block 80 in data stream 14. A dependency may be, however, be present relative to a flag switching between the partitioning handling of intra-coded block 80 just-discussed and a different way of handling the intra-coding of block 80 as will be outlined hereinafter.

According to ISP, each of the partitions 102/1 12 is predicted, transformed, quantized and coded individually, with sequentially processing the partitions in this manner. Therefore, the reconstructed samples of a certain partition will be able to be used to predict any following partition 102/112 in partition order among the partitions into which block 80 has been partitioned, and in this manner, the process of intra-prediction cycles through the partitions 102/1 12 into which block 80 has been partitioned. Fig. 5 exemplarily shows an intra-predicted block 80 split according to option 100. Each partition 102i to 1024 of block 80 is subject to prediction, i.e., derivation of the predictor of the respective partition 102,, and prediction residual related task, namely the correction of the predictor using the prediction residual. The latter task may be performed by combining the prediction residual and the predictor. This is done in the decoder for reconstruction. The encoder performs as prediction residual related task the determination of the prediction residual involving, for instance, transformation and quantization, as well as the correction of the predictor using the prediction residual, namely in order to keep the prediction loop synchronized to the decoder by filling the decoded picture buffer in the encoder with the reconstruction of the picture. The just-mentioned tasks, i.e. prediction and residual handling, are performed individually for, and sequentially among, partitions 102i to 1024. After those two steps for a currently processed partition, the next partition 102,· according to the partition order is processed the same way. The partition order is exemplarily illustrated in Fig. 5 using the three arrows 126.

Fig. 5 illustrates that the partition including the upper most left pixel of block 80 would be treated first before proceeding with the immediately lower neighbor partition 1022 and so forth, corresponding to the assignment of indices to partitions 102i to 1024 in Fig. 5, but this order is merely an example and as the following description will render clear, this partition order may be chosen in a manner depending on other settings such as the intra-coding mode and/or the size of block 80 with the former dependency being discussed hereinafter.

In the examples discussed further below, the partition order 126 merely varies between ones traversing the partitions 102/1 12 in a manner so that immediately succeeding partitions immediately neighbor each other so that, in case of split type 100, the partition order leads from top to bottom or bottom to top, and in case of partition type 1 10 from left to right or right to left, respectively. It should be mentioned, however, that other examples are imaginable too. For instance, the partition order could be chosen in a manner so that the partitions are scanned in the just-outlined neighboring order in two scans with, in the first scan, processing every second partition from top to bottom, bottom to top, left to right or right to left, whatever applies, and then processing in the same direction of order, or in the opposite direction, the remaining partitions therebetween.

In any case, Fig. 5 illustrates the first partition 102i to be processed first and to be the currently processed partition. For the first partition, here exemplarily 102i, the set of neighboring samples 1 18i used to form the predictor for partition 102i may merely be chosen on the basis of samples lying outside the boarders of block 80 as at the time of processing the first partition of block 80, no sample of block 80 has been processed yet, i.e., reconstructed or encoded. That is, the samples in set 1 18i are already reconstructed in the encoder using any prediction and correction of the corresponding predictor using a prediction residual transmitted in the data stream. They belong to previously coded/decoded picture blocks and may be inter-coded or intra-coded or any other coded block. As to number and exact position of the samples of the set 1 18i of neighboring samples which are used for forming the predictor of the first partition 102i, same depend on the intra-coding mode assigned to block 80. This intra-coding mode is jointly, or equally, used for the processing of every partition of block 80 as will be discussed in the following. In order to finish the processing of the first partition 102i, the predictor for this partition 102i derived in decoder and encoder by filling this partition 102i depending on the one or more already reconstructed/encoded samples in set 1 18i , its prediction residual is determined as far the encoder is concerned, namely by transformation and quantization as outlined above, and then this prediction residual is - in the version transmitted in the data stream, i.e. including the quantization loss - used for reconstruction of this partition 102i by correcting the predictor using the prediction residual in the data stream 14. Fig. 5, for instance, shows the prediction residual for partition 102i exemplarily at 120i . That is, 120i comprises the transform coefficients corresponding to the transform of the prediction residual of partition 102i with a description of data 120i being discussed in more detail below.

Turning now to the next partition in partition order, namely partition 1022 in the example of Fig. 5. The situation has changed insofar as the set of neighboring already reconstructed/encoded samples used for deriving the predictor for partition 102å may now be composed of samples located outside block 80 and/or samples within block 80, namely ones located in any already processed partition, here currently partition 102i in the example of Fig. 5, as for these samples the prediction residual has already been determined and is already available in data stream 14. That is, encoder and decoder derive the predictor for this partition 1022 followed by prediction residual determination in the encoder and prediction residual usage for correction of the predictor in encoder and decoder,

respectively. This process is then continued with the next partition in line, i.e. , the next partition in partition order, thereby sequentially handling all partitions of block 80.

As has already been mentioned above, it could be possible that the partition order 126 is chosen in another manner than traversing the partition so that immediately consecutive partitions are immediate partition neighbors. That is, the partition order may jump from one partition to the next partition. This implies that the sets 118, of neighboring samples used for deriving the respective predictor by filling the respective partitioning 102, is not restricted to immediate sample neighbors of the respective partition as illustrated in Fig. 5. This also pertains the selection of the start of the partition order 126. Imagine, for instance, partition 102 was the first partition in partition order. Then, its predictor could be derived by filling same depending on a set of neighboring samples 1184, not illustrated in Fig. 5, which collects samples located alongside the circumference of block 80 to the left and to the top of block 80. Some of samples in set 1 184 would not immediately neighbor partition 1024. This would, by the way, correspond to the situation of filling the last sample row in the usual intra-prediction filling of block 80 en bloc. The just-mentioned possibility is also true with respect to any subsequently processed partition, i.e., the second and further partitions in partition order. That is, their neighbor sample set 118, may also contain samples not immediately neighboring the respective partition 102,. And even further, in case of not restricting the partition order to traverse the partitions in the manner so that consecutive partitions are immediate neighbors of each other, then the set of reference samples 118, of any second or subsequently processed partition 102, may not only collect samples lying to the left and to the top of the respective partition 102i, but may also be samples lying below the respective partition 102i depending on whether any partition of block 80 has been processed earlier than partition 102i according to the partition order. That is, set 180, may comprise samples located on more than two sides of partition 102,.

Briefly summarizing, Fig. 5 showed the sequential processing of the partitions 102/112 of a block 80 here exemplarily with respect to horizontal partitions, but the same description also applies to the vertical mode 110 with respect to vertical partitions 112,. For each partition 102j, corresponding prediction residual 102i is contained in data stream 14. Data 120i to I2O4 together forms the prediction residual for block 80, namely 120. It should be recalled that transform residual coding may, in accordance with an alternative embodiment of the present application, not be used, i.e., the prediction residual 120 of block 80 may be signaled in the data stream 14, for instance, in spatial domain directly. In this case, the data 120i to 12O4 for the various partitions 102i to 1024 may not contain partition separate fields in data stream 14 as illustrated in Fig. 5 where each data portion 120i represents the signaling for a certain transform of the respective partition 102,. Rather, the prediction residual 120 for block 80 could, in that case, form one field of data 14. The decoder would, at the time of processing a certain partition 102i, collect the information on the prediction residual for this partition 102i from field 120 in this alternative embodiment. This procedure could also be applied when using an exactly reversible version of the transformation so that the quantization may be done in spatial domain.

Thus, Fig. 5 showed that in an encoder and decoder, two tasks are performed for each partition 102i, namely: 1) the prediction derivation task 122 yielding the prediction or predictor for the respective partition 102i, i.e., a predicted sample value for each sample of the respective partition 102, and 2) the prediction residual related task performed afterwards, namely the prediction residual derivation at the encoder including the quantization of the prediction residual for sake of its entry into data stream 14, and the reconstruction of the samples of the respective partition 102, by combining or correcting the prediction residual and the predictor so as to gain reconstructed samples for this partition 102j. The latter reconstructed samples may serve as a reservoir for neighboring sample sets 118j of subsequently processed partitions 10¾ following in partition order 126 for sake of the prediction derivation task.

Before proceeding with the further description of details of ISP concept possibilities, Fig. 6 illustrates the process of prediction derivation 122 by filling a currently processed partition 102i wherein it should be recalled that the illustration with respect to a horizontal partition 102 has been chosen merely illustratively and that the same description also relates to vertical partitions 112. Fig. 6 shows the currently processed partition 102, and its corresponding set of neighboring samples 1 18, already reconstructed/encoded. As already denoted above with respect to Fig. 5, set 1 18, may not be restricted to samples 128 directly neighboring, or being adjacent to, partition 102,. However, due to the partitioning, the mean distance 130 between samples of partition 102i and samples 128 of set 118i, when averaged over all samples of block 80, is lower than compared to performing intra-prediction of block 80 as known from, for instance, H.264 or HEVC. As described with respect to Fig. 5, the predictor derivation or filling 122 is performed for each partition 102, using the intraprediction mode associated with block 80, with this mode indicating one of a set of available intra-prediction modes. This set may comprise angular or directional modes differing from each other in the angle or direction 132 along which the sample content of the neighboring sample set 118i is copied into samples 134 of partition 102,. In order to perform this copying, the prediction of each sample 134 of partition 102, may derived on the basis of a number of neighboring samples 134 out of set 1 18, located relative to the sample 134 in the direction facing opposite to direction 132. The number is, for instance, defined by a kernel of an interpolation filter used to derive inter-pel positions between the samples 128 of sample set 118j. Fig. 6 illustrates, for instance, that three samples 128 out of set 118, are used to compute the prediction of one sample 134 out of currently processed partition 102,. Owing to the relatively small mean distance 130, the number of reference samples 134 per sample 134 of partition 102, may be kept low. More details will be presented hereinafter. For sake of completeness, however, it should be noted that the set of available intra-prediction modes may also comprise a DC mode according to which one DC value is assigned to all samples 134 of partition 102i, with deriving this DC value by performing an averaging on the set of neighboring samples 118i. Further, a planar mode may exist according to which the prediction values for samples 134 are defined by a linear function over the sample positions within partition 102, with deriving slope and offsets of this linear function on the basis of the neighboring samples 118,. Further, it should be noted that the neighboring set 118, may be different depending on the intra-prediction mode chosen for block 80 and may be, for instance, particularly different between angular modes and the non-angular modes DC/planar.

For example, in the state-of-the-art JEM decoder, there are 67 intra-prediction modes available: 65 of them are angular modes and two of them, DC and planar, model non-directional textures. The 1 D partitioning (briefly called 1 D partition mode) mode, i.e. , the predictor derivation 122 performed for the partitions 102/112 outlined above and hereinafter, according to which block 80 is partitionOd/split into partitions along dimension 104 with the resulting partitions extending over the complete width of the block transversely to dimension 104 with being one sample wide or more than one sample wide along direction 104, can be combined with any of them or, differently speaking, could be implemented using any of them. As already described with respect to Fig. 5, all partitions 102/112 of one block 80, such as a coding unit CU, use the same associated intra-prediction mode of block 80, thereby avoiding an excessive overhead in signalization as the intra-prediction mode 116 needs to be transmitted in the data stream 14 merely once for block 80.

That is, the prediction 122 may be carried out in the same way as in the two-dimensional case outlined in the JEM decoder. However, merely one line, be it horizontal or vertical, is calculated for a currently processed partition 102/112 so that compared to JEM, the prediction process 122 would be adjusted accordingly. In case of choosing the partition

order for traversing the partitions in a manner so that consecutive partitions immediately neighbor each other, the prediction process 122 may correspond to the two-dimensional case of JEM, but merely with respect to the first line, i.e., the one being nearest to the already reconstructed/encoded neighborhood. In some cases, both HEVC and JEM allow the usage of certain filters applied on the reference samples 128 or on the resulting predictor. This is useful in the two-dimensional case to better predict samples within the predict block 80 that are far away from the reference samples 128 to reduce boundary discontinuities. However, by using the partitioning into partitions 102/112, it is possible, and it should be the aim, to exploit the high correlations among nearby pixels.

That is, the reduced mean distance 130 should be exploited. Excessive smoothing would reduce this quality. Accordingly, should the encoder or decoder be able to perform both kinds of intra-predictions, namely intra-prediction using partitioning as outlined with respect of Figs. 4 to 6 and in the following, then intra-filters, i.e., filters involved in the predictor derivation 122, are either disabled or at least the number of contributing samples 134 per partition sample 134 is reduced relative to the number of samples contributing to one sample in the two-dimensional case where the intra-prediction for block 80 is performed on block or performed according to HEVC, namely decomposed into leaf blocks of a hierarchical quadtree subdivisioning of block 80 into rectangular blocks.

As became clear from the discussion brought forward above, in order to perform the prediction residual related tasks 124, the decoder decodes, for instance, a transform of the respective prediction residual of the currently processed partition from the data stream 14, and performs an inverse transform such as a spectral-to-spatial transformation onto this transform in order yield the prediction residual which is then used to correct the predictor obtained at 122 by combination/addition. The same is done in the encoder in order to keep the prediction loop synchronized with the decoder. In addition, the encoder performs the transformation of the prediction error of the predictor determined using 122 for a currently processed partition, subjects same to a transformation such as a spatial-to-spectral transformation, followed by a quantization of the transform coefficients with then coding the transform into the data stream 14 to yield the corresponding data 120, of the currently processed partition 102,. As to the transformation, all partitions 102/112 within block 80 may be treated using this same transformation. It may be a DCT-II except in the case of the planar mode, for instance, where the DST-VII may be used. For this reason, all tools related to the transformation and inverse transformation which encoder and decoder may use for other blocks such as transform skipping, i.e., coding in spatial domain, EMT (EMT = Explicit multiple core transform), NSST (NSST = Mode dependent non-separable secondary transforms) or others, may be disabled if block 80 is coded using intra-prediction mode in the partitioning manner outlined so far with respect to Figs. 4 to 7 and further outlined below, to avoid unnecessary overhead bits. Even alternatively, the transformation may be a linear transform a type of which is selected based on one or more of the intra prediction mode, an dedicated syntax element and the predetermined partition order.

Some words have already been spent with respect to the partition order 126 using which the partitions 102/1 12 of currently processed block 80 are sequentially processed. It should be emphasized that this embodiment is merely an example and that partition order may be static in accordance with alternative embodiments, or may be varied in a different manner in accordance with examples are set out below. Fig. 7 indicates by inscribed numbering possible partitions/processing orders illustrated in using arrows 126 in Fig. 5. Here, this order follows the inscribed numbers in ascending order. Fig. 5 represented an example, where the order 126 started with a partition containing the top-left pixel/sample 140 of block 80 with leading downward to the lowest partition. Similarly, if the split type were vertical, then the processing order would start with a left most partition containing the top-left pixel/sample again with leading rightwards. However, this is not the optimal case for all existing intra-prediction modes. This is exemplified in Fig. 7 which shows the vertical and horizontal partitioning of a block 80 for the diagonal modes too, i.e., the copy angle/direction 132 points at 45° from lower-left to the upper-right-hand side, and 34, i.e., the copy angle/direction 132 points at 45 from upper-left to lower-right. In the former case, if the split is horizontal, then starting at the top-left corner of the block 80 would produce partitions whose reconstructed samples would not have any influence on the prediction of the following partition. Consequently, it is more reasonable to start at the bottom-left corner of the block, so that the reconstructed samples of each partition can be used to predict the next partition in partition order. In the vertical split, nevertheless, this is not necessary, as it can be observed in the aforementioned figures. On the other hand, mode 34 does not have any of these problems, given that samples come in both horizontal and vertical splits from both sides. Therefore, the normal processing order may be employed in both splits.

Table 1 shows the complete list of processing orders according to the intra-prediction mode and the split type.

Table 1: Processing order according to the intra mode and the split type. HOR_DIR AND VER_DIR are the horizontal and vertical modes respectively and VDIA_DIR is the vertical diagonal mode

Summarizing the ISP concepts described so far with respect to signalization overhead, reference is made to Fig. 8. Fig. 8 illustrates as to what is transmitted for a block 80 in accordance with an embodiment of the present application. In particular, there is the intraprediction mode signalization 116 signaling as to which intra-prediction mode is to be applied to block 80. Thus, signalization 116 indicates one of the angular modes, for instance, or one of the available modes including the angular modes and non-angular mode(s) such as DC and planar. In addition to this signalization 116, there is a partitioning flag 160 coded by the encoder into data stream 14 and decoded therefrom for block 80 by the decoder, which indicates whether the partitioning treatment according to Figs. 4 to 7 is applied to block 80, or whether same is treated“normally”, such as en block or in one piece or two-dimensionally, i.e., merely samples outside block 80 are used to form the reference sample reservoir 118 to predict each sample within block 80. Alternatively, flag 160 may switch between the partitioning treatment discussed with respect to Figs. 4 to 7 on the one hand, and the decomposition of block 80 using a quadtree subdivisioning into transform blocks which are then sequentially treated with a disadvantage, however, of having to signal the decomposition in data stream 14. If the partitioning flag 160 indicates the partitioning according to Fig. 4, then data stream 14 contains for block 80 the partitioning dimension flag 114 switching between the partitioning types 100 and 110 discussed with respect to Fig. 4. And then, for each partition of block 80 into which block 80 is sub-divided/partitioned if partitioning flag 160 indicates this partitioning option, data stream 14 comprises a signaling/data 120i having the prediction residual of the respective partition encoded there into such as, as indicated above, in transform domain.

With respect to Fig. 8, it should be noted that the prediction residual data 120i, 1202 ... may be coded into data stream 14 in an order Corresponding to the partition/coding order 126. The latter may, as discussed above, be uniquely determined by the intra-prediction mode

indicated by signalization 1 16. An alternative, however, would be that the partitioning order 126 is at least partially determined on the basis of an optional additionally signalization in data stream 14.

A further alternative to the description brought forward herein is the fact that signalization 1 16 may alternatively be used in order to indicate whether the partitioning option is used or not. In other words, one syntax element may commonly assume responsibility for the signalization of 1 16 and 160. Such a syntax element would assume one out of a range of values each corresponding to a combination of intra-prediction mode and the indication whether block partitioning is used or not. In such a case, I would also be possible to offer the partitioning option merely for a subset of the intra-prediction modes. And lastly, it should be noted that partitioning flag 160 may also be conveyed in data stream 14 conditionally merely in case of the intra-prediction mode indicated by signalization 1 16 assumes a certain subset out of the available intra-prediction modes.

Fig. 9 shows exemplarily as to how the data 120, having the prediction residual of a certain partition 102/1 12, could look like. According to Fig. 9, the prediction residual is coded into data stream 14 in transform domain. That is, the encoder generates, by way of transformation 180, a transform 182 of the prediction residual, with a decoder deriving the prediction residual and spatial domain by way of inverse transformation 184. Fig. 9 illustrates the transform coefficients 186 of transform 182 corresponding, for instance, to different spectral frequencies f. Data 120, could comprise a coded block flag CBF data 120i could comprise a coded block flag CBF 188 indicating whether transform 182 comprises any significant transform coefficient 186, i.e., whether transform 182 is completely zero or not. If CBF 188 is set, the transform 182 is not zero, and data 120, may comprise a last position (LP) syntax element 190 indicating 192 the last position along increasing spectral frequency (see axis 194) of a significant transform coefficient, i.e., a non-zero transform coefficient 186, starting from the lowest or DC coefficient 196. Then, data 120, comprises signaling 198 signaling the transform coefficients from 196 to 192.

That is, Fig. 9 illustrates that each partition 102/1 12, may have its prediction residual coded into data stream 14 by way of CBF 188, LP 190 and transform coefficient data 198. That is, for a block 80 with n partitions 102/1 12, there will be n CBFs 188, and one LP 190 for each partition with a non-zero CBF 188, and with the transform coefficient data 198 merely for these partitions having a non-zero CBF 188 associated therewith. The coefficient data 198 may be coded in the same manner as is done for the normally treated intra-predicted blocks, i.e., blocks 80 for which the partitioning flag 160 indicates the non-partitioning option, with the following exceptions: each IP 190 requires only one coordinate if the partition is one-sample wide (otherwise it requires 2 coordinates, as usual), namely x for horizontal splits 100 and y for vertical ones 1 10. In case of two-dimensional partitions, though, LP 190 indicates the last position along a scan direction or path either using rank indication, or using x and y coordinates. The context of each CBF 188 may be chosen to be the value of previously coded CBF, i.e., the CBF of the previous partition in partition order 126. Further, owing to the partitioning, the transform coefficient data 198 relates to a different shape. That is, the transform 182 has a different shape, too. The transform 182 is a one-dimensional transform in case of the partitions being one-dimensional partitions as discussed with respect to Fig. 4. That is, the transform 182 may be a W/H long vector of transform coefficients 186 depending on the split type 100 or 1 10.

With respect to the flags 160 and 1 14 of Fig. 8, and their coding, the following is noted. The flag 160, which indicates whether block 80 is to be divided into partitions 102/1 12, defines the condition to be checked whether flag 1 14 is conveyed in data stream 14 for block 80. In particular, if flag 160 indicates the partitioning into partitions 102/1 12, flag 1 14 is present in data stream 14 and sent to the decoder in order to signal as to which type of split 100/1 10 is to be performed, i.e., either horizontal or vertical. Just as the flag CBF, also flag 1 14 may be coded using context-dependent entropy coding/decoding. The context of flag 1 14 may have three possibilities according to the intra-prediction modes of block 80: 0 for non-angular modes, 1 for horizontal modes and 2 for vertical modes.

While Fig. 9 illustrated that that a CBF 188 might be present once per partition i of the current block 80, additionally or alternatively, the transform 182 of the partitions 120i of the current block might each be portioned into one or more sub-blocks with a coded sub-block flag signaled per sub-block within data 120i indicating whether the transform coefficients 186 within that sub-block are all zero or at least one coefficient thereof is non-zero. Thus, only coefficients 186 within sub-blocks for which the coded sub-block flag signals the presence of non-zero coefficients, would be coded, The other coefficients within sub-blocks for which the coded sub-block flag signals the absence of any non-zero coefficients, would be inferred to be zero at decoder side. Note that as each partition 120i is transformed separately, sub-blocks belonging to one partition differ in spectral components of the transform 182 of that partition and differ in the transform coefficients 186 they comprise out of that transform. For example, sub-blocks can be set in such a way that they are 4x4 coefficient blocks as long as the respective partition 102i/1 12, has dimensions in x (partition width) and y (partition height) which are both equal to or larger than 4 samples 140, and consequently as long as the transform 180 of the respective partition 102I/1 12I has dimensions in x and y which are both equal to or larger than 4 coefficients 186. In case of a 4xN partition, the sub-blocks form a column of m 4x4 sub-blocks with m*4=N and m being an integer. In case of a Nx4 partition, the sub-blocks form a row of m 4x4 sub-blocks with m*4=N and m being an integer. For broader partitions, an array of 4x4 sub-blocks, arranged in rows and columns, could result. Depending on the embedment it might be, however, that such partitions, i.e. broader than 4 samples and/or as broad as 4 samples, do not occur. Irrespective of the occurrence or not, for partitions being narrower, i.e. for which one of its dimensions is less than 4 samples, i.e. vyhich are less then 4 sample wide in at least one dimension x or y, the sub-block partitioning of its transform 180 into sub-blocks each which gathering different groups of coefficients of that transform 180 may be done so that the subblocks have a minimum number M of coefficients in all possible cases for the current block’s size. That is, the partitions might be set to be as large as the block width N along one dimension, with the partitioning taking place along the other dimension 104. The transform 180 of each partition may, thus, be of size 1 xN, 2xN, Nx1 or Nx2. In fact, the transform 180 of a certain partition may have a number of coefficients equaling the number of samples in this partition. In case of a 1xN partition/transform, the sub-blocks may then form a column of m 1xM sub-blocks with m*M=N and m being an integer. In case of a Nx1 partition, the sub-blocks form a row of m Mx1 sub-blbcks with m*M=N and m being an integer. In case of a 2xN partition/transform, the sub-blocks may then form a column of m 2x(M/2) sub-blocks with m*(M/2)=N and m being an integer. In case of a Nx2 partition, the sub-blocks may form a row of m (M/2)x2 sub-blocks with m*(M/2)=N and m being an integer. This is exemplarily shown in Table 2 for the exemplary case of M=16 for the minimum number of coefficients.

Table 2: Entropy coding coefficient group sizes

Block (partition/transform) Size Coefficient group (sub-block) Size

All other possible M x N cases 4 x 4

While Fig. 9 illustrated that that a CBF 188 might be present once per partition i of the current block 80, it might be agreed between decoder and encoder that at least one of the n partitions among the partitions for a current block 80 has a non-zero CBF 188. For this reason, ifn is the number of sub-partitions and the first n— 1 sub-partitions in coding order have produced a zero CBF, then the CBF of the n-th partition will be inferred to be 1. Therefore, it is not necessary to decode it and it is not encoded. Thus, the CBF of data 120n would be missing is the CBF in data 120i to 120n-i signalled zeroness and the decoder would infer this CBF to signal that at least one non-zero coefficient is present in the transform of that partition.

As far as the intra-coding mode signalization 1 16 is concerned, the following may hold true. It might be the coding mode signalization 1 16 is sent as a pointer or index which points to one out of a list of most probable modes (MPM). The latter MPM list, in turn, might be determined by encoder and encoder in the same manner based on intra prediction modes used for previously coded/decoded intra-predicted blocks such as spatially and/or temporally neighboring intra prediction modes. Thus, the MPM list may represent a proper subset of available/supported intra-prediction modes, namely the afore-mentioned angular modes and/or one or more of DC and planar modes. As mentioned above, it might be that there are intra-predicted blocks using the LIP or ISP scheme such as block 80 in the figures, besides ones which are intra-predicted classically, i.e. en block or in units of transform blocks into which such intra-predicted blocks are partitioned using recursive quadtree partitioning. Both types of intra-predicted blocks might support the same set of available/supported intra-prediction modes. While for the later normal/classical intra-predicted blocks, a MPM flag may be signaled in the data stream - with the decoder decoding same and the encoder encoding same - indicating whether the mode of that block is selected out of the MPM list, in which case a pointer/index into this MPM list is transmitted - with the decoder decoding same and the encoder encoding same - the MPM flag would be inferred to signal the MPM list restriction in case of intra-predicted blocks using the LIP or ISP scheme such as block 80. If, for a certain normal/classical intra-predicted block the MPM flag signals that none of the MPM modes is used, no index/pointer is present for that block in the data stream and a substitute pointer/index into a remainder list of intraprediction modes is transmitted in the data stream for that block instead. The remainder list may also be a proper subset of the set of available/supported intra-prediction modes, and may, in particular, be the complementary set of the MPM list compared to the set of available/supported intra-prediction modes, i.e. every member of the set of available/supported intra-prediction modes would either be member of the MPM list or the remainder set. The pointer/index into the MPM list might be VLC coded, while the pointer/index into the remainder set might be coded using a fixed-length code. Naturally, it might be that even for intra-predicted blocks of the LIP or ISP scheme, the MPM flag is

transmited and that the encoder would be free to select any mode out of the set of available/supported intra-prediction modes with setting the MPM flag depending on whether the selected mode is thin the MPM list or the remainder set.

The MPM list might be the same, i.e. might be determined in the same manner by encoder and encoder, for the normal/classical intra-predicted blocks as well as for the ISP/LIP intra-predicted blocks. However, irrespective of whether restriction to the MPM list and inference of the MPM flag to signal MPM list usage for ISP/LIP intra-predicted blocks applying or not, alternatively, the MPM list may be determined differently for ISP/LIP intra-predicted blocks in order to adapt to the statistics of the ISP/LIP mode. For example, it could be altered to exclude the DC intra mode from the MP list and to prioritize horizontal intra modes for the ISP horizontal split, i.e. horizontal direction 104, and vertical intra modes for the vertical one, i.e. vertical direction 104. That is, for a normal/classical intra-predicted block, the MPM list could form a proper subset of the set of available/supported intra-prediction modes, the modes being selected and ordered in accordance with a certain concept. For ISP/LIP intra-predicted blocks 80, the MPM index could point to an MPM list which depends on the partitioning direction 104 signaled by flag 1 14 and/or which forms a proper subset of set of available/supported intra-prediction modes less the DC mode or less the DC and planar modes, i.e. a proper subset of the angular modes in the set of available/supported intraprediction modes. The MPM list construction based on previously used intra-prediction modes of previously coded/decoded could prefer angular modes of an angular intra prediction direction being closer to the horizontal dimension in case of the flag 1 14 indicating the partitioning direction 104 to be horizontal and prefer angular modes of an angular intra prediction direction being closer to the vertical dimension in case of the flag 1 14 indicating the partitioning direction 104 to be vertical.

With respect to the description just brought forward, it is again noted that the juxtaposition between intra-prediction modes normally treated and intra-prediction modes treated using partitioning as outlined herein, needs not to be. That is, encoder and decoder may treat intra-predicted block 80 inevitably using the partitioning presented herein with then, accordingly the partitioning flag 160, for instance, becoming obsolete. If, however, the partitioning option signaled by flag 160 is available as one decision for the encoder, then the following description reveals a possibility as to how the encoder performs the decision, or finds out, whether the partition mode should be used for a certain block 80 and which split type, namely horizontal or vertical, is the best one. In order to perform this, the encoder should test both options for different intra-prediction modes for each block. Compared to

the case where the encoder would merely have one option such as the normal option, the encoder will, thus, be slower since more options have to be tested. In order to reduce this impact, the partition mode signaled by flag 160 may be tested by the encoded according to the following strategy, wherein reference is made to Figs. 10 and 11.

1 ) The 1 D partitions mode is the last intra mode to be tested.

2) Let Cmin be the minimal cost so far at the moment when the 1 D partitions mode is going to be tested.

3) Select a combination of intra mode and split type to be tested.

4) The block is split into N 1 D partitions and let i denote the index of each of these partitions, where i = [1, N],

5) After every partition is coded, its subcost is calculated. Therefore, we can know the sum of all the subcosts that are available after the partition i has been coded, which is Si =
This procedure is depicted in Fig. 10 which, thus, illustrates the accumulation of the 1 D partitions subcosts to obtain the final cost of the whole block.

6) After every partition is processed, the expression St < Cmin is evaluated. If it is true, we continue coding partitions till the end. Otherwise, it is guaranteed that this test mode is not going to yield a lower RD cost than Cmin, so the process is interrupted and we move on to the next combination of intra mode and split type.

7) If all 1 D partitions are coded, then the test mode is the best mode and Cmin is updated accordingly.

The advantage of this procedure is that it avoids processing 1 D partitions that are not necessary, since it can be already known that the 1 D partitions mode is not going to yield a better cost than the already existing minimal cost. Besides, it does not have any drawbacks in terms of RD loss. The whole process is illustrated as a flow chart in Fig. 11.

It is noted again that all of the above ISP examples illustrating the partitioning as being made in one sample wide stripes transverse to direction 104, the partitioning may alternatively be made in a manner leading to partitions being wider, thereby leading to two-dimensional partitions. Further alternatives with respect to the partitioning are set out below.

In a specific ISP example, the width along splitting direction 104 is defined based on 1) whether the intra-prediction mode is an angular or non-angular mode, and 2) the width of the intra-predicted block along that direction 104.

) A W x H (where it is assumed W and H are powers of 2) block 80 can be divided horizontally or vertically (indicated, for instance, with a syntax element 114 that is sent to the decoder) into K equal partitions 102/1 12 with w h dimensions, whose values are described in Table . According to Table 3, a block with W = 16, H = 8 predicted using a non-angular intra mode and subject to a vertical split (i.e. with direction 104 being vertical), for instance, would be split into 4 partitions 102 all of which would have dimensions w = 16 and h = 2. If the same block 80 was predicted using an angular intra mode, it would be split into 8 partitions 102 each of which had dimensions w = 16 and h = 1.

Table 3: Values of w, h and K for the extra layout example 1

) A W x H (where it is assumed W and H are powers of 2) block 80 can alternatively be divided horizontally or vertically (indicated, for instance, with a syntax element 1 14 that is sent to the decoder) into K equal partitions with w x h dimensions, where the value of K is not fixed (and therefore it is transmitted to the decoder with a syntax element) and its range can be any power of 2 between 2 and S, where S is value of the dimension that is being split (width for the vertical split and height for the horizontal one). The values of w and h are obtained as described in Table 4.

Table 4: Values of w and h for the extra layout example 2

Instead, the width of the partitions along the dimension 104 could be signaled for block 80 directly.

3) A W x H (where it is assumed W and H are powers of 2) block 80 can alternatively be divided horizontally or vertically (indicated, for instance, with a syntax element 1 14 that is sent to the decoder) into K partitions (where K depends on W and H) with wi x hi dimensions with i = 1, 2, Let S = H, s¾ = ht if the split is horizontal and S = W, Si = Wi if it is vertical). Various options of values of
are described in Table 5 for different values of S which measures the width of block 80 along dimension 104 and s( measures the width of partition i along dimension 104.

The option that is used by the decoder is fixed or it can be determined implicitly according to the value of existing parameters at the decoder side.

4) A W x H (where it is assumed W and H are powers of 2) block 80 can alternatively be divided horizontally or vertically (indicated, for instance, with a syntax element 114 that is sent to the decoder) into K partitions (where K depends on W and H) with Wi x ht dimensions with i = 1, 2, ... , K. Let S = H, st = ht if the split is horizontal and S = W, Si = wt if it is vertical). The value of st will be determined through a syntax element that indicates which of the three options presented in Example 3) is to be used to divide the block into sub-partitions.

Thus, as exemplified in above examples 1 to 4, the partitioning may be done along one dimension 104 so that the partitions are as wide as the predetermined block perpendicular to predetermined dimension, while a width of the partitions, measured along the predetermined dimension 104, is selected out of at least two different width settings or options. Explicit or implicit signaling concepts may be used to keep the selection synchronous between encoder and decoder. The selection, thus, enables that, while partitioning may be varied between blocks of the same size and shape, the overhead associated with this variation is kept reasonably low. The selection may, for instance, be

done depending on the intra-coding mode for the predetermined block such as depending on whether the intra-coding mode for the predetermined block is an angular mode or not. The selection may also be made depending on an index in the data stream for the predetermined block indexing on of the at least two different width settings as shown in example No. 4. The partitions may be one or more samples wide along the partitioning dimension. Within one block, the partitions width along the partitioning/predetermined direction may vary. One may be one sample wide, i.e. is one-dimensional stripe, while another is more than one sample wide, is a two-dimensional field of samples.

WE CLAIMS

1. Decoder for block-based decoding of a picture (12) from a data stream (14), configured to

decode an intra-coding mode (116) for a predetermined block (80) of the picture from the data stream (14);

decode a partition dimension flag (114) for the predetermined block of the picture from the data stream and set a partition dimension (104) depending on the partition dimension flag to be horizontal or vertical,

partition, along the predetermined dimension (104), the predetermined block (80) into transform partitions (300) which are as wide as the predetermined block perpendicular to predetermined dimension;

decode, for each transform partition, a transform (182) of a prediction residual from the data stream;

intra-predicting the predetermined block depending on one or more already reconstructed samples neighboring the predetermined block in a manner depending on the intra-coding mode to obtain a predictor for the predetermined block; and

reconstructing the predetermined block by correcting (122) the predictor within each transform partition using the transform of the prediction residual decoded for the respective transform partition.

2. Decoder of claim 1 ,

wherein the number of transform partitions is greater than 2 and/or the transform partitions are one sample wide along the predetermined dimension.

3. Decoder of claim 1 or 2,

wherein the decoder is configured to re-transform for each transform partition the transform into spatial domain for usage in correcting the predictor within the respective transform partition.

4. Decoder of any of claims 1 to 3,

wherein the decoder is configured to decode, for each partition, the transform from the data stream by

decoding a coded transform partition flag (188) from the data stream;

if the coded transform partition flag is not set, set the prediction residual for the respective transform partition to zero, and

if the coded transform partition flag is set, decode transform coefficients of the transform of the prediction residual of the respective transform partition from the data stream.

5. Decoder of claim 4,

configured to decode the coded transform partition flags for the transform partitions from the data stream sequentially and infer that the coded transform partition flag for a last transform partition in transform partition order is set, if all preceding coded transform partition flags are not set.

6. Decoder of claim 5,

configured to decode the coded transform partition flag for a respective transform partition from the data stream by use of context-dependent entropy decoding using a context which depends on the coded transform partition flag decoded for a preceding transform partition preceding the respective transform partition in the predetermined transform partition order.

7. Decoder of any of claims 1 to 6,

wherein the decoder is configured to

decode, for a predetermined transform partition, the transform of the prediction residual of the predetermined partition from the data stream by

decoding a last position indication (190) form the data stream indicating a last transform coefficient position of the transform along a predetermined scan order scanning transform coefficients of the one-dimensional transform; and

decoding transform coefficients (198) of the transform up to the last transform coefficient position along the predetermined scan order from the data stream and inferring that transform coefficients of the transform beyond the last transform coefficient position along the predetermined scan order are zero.

8. Decoder of claim 7, wherein the transform partitions are one sample wide along to the predetermined dimension and the transform is a one-dimensional transform.

9. Decoder of any of claims 1 to 8,

wherein

the transform is a DCT transform in case of the intra prediction mode not being a planar mode, and a DST transform in case of the intra prediction mode being the planar mode, or

the transform is a linear transform a type of which is selected based on the intra prediction mode, a block size of the predetermined block and/or an dedicated syntax element.

10. Decoder of any of claims 1 to 9,

configured to

decode a split mode flag (160) for the predetermined block of the picture from the data stream;

if the split mode flag indicates a first split mode, perform the decoding the partition dimension flag, the partitioning and the decoding of the transform for each transform partition;

if the split mode flag indicates a second split mode, instead of the decoding the partition dimension flag, the partitioning and the decoding of the transform for each transform partition, decode one transform of the prediction residual within the predetermined block.

1 1. Decoder of any of the previous claims,

configured to

decode the partition dimension flag (114) by use of context-dependent entropy decoding using a context which depends on the intra-coding mode.

12. Decoder of any of the previous claims ,

configured to

decode the partition dimension flag (114) by use of context-dependent entropy decoding using one of three contexts comprising

the intra-coding mode signaling a non-angular mode,

the intra-coding mode signaling a horizontal mode,

the intra-coding mode signaling a vertical mode.

13. Decoder of any of claims 1 to 12,

configured to set a width of the transform partitions, measured along the predetermined dimension (104),

depending on a size of the predetermined block (80) along predetermined dimension (104) and/or

depending on the intra-coding mode for the predetermined block and/or

depending on whether the intra-coding mode for the predetermined block is an angular mode or not.

14. Decoder of any of claims 1 to 13,

wherein the decoder is configured to perform the intra-predicting the predetermined block depending on one or more already reconstructed samples neighboring the predetermined block in a manner depending on the intra-coding mode to obtain a predictor for the predetermined block and the reconstructing the predetermined block by correcting (122) the predictor within each transform partition using the transform of the prediction residual decoded for the respective transform partition by

sequentially reconstructing groups of transform partitions into which the transform partitions of the predetermined block are grouped to form a sub-partition (102; 112) for each group of transform partitions according to a predetermined sub-partition order (126) which sequentially traverses the sub-partitions along the predetermined dimension by, for a current sub-partition and before proceeding with a subsequent sub-partition,

deriving (122) a predictor for the current sub-partition by filling the current sub-partition depending on one or more already reconstructed samples neighboring the current sub-partition in a manner depending on the intracoding mode;

reconstructing the current sub-partition by correcting (122) the predictor within each transform partition comprised by the group of transform partitions forming the current sub-partition using the transform of the respective transform partition.

15. Decoder of claim 1 ,

Configured so that

a number of transform partitions (300) per sub-partition

depends on dimensions of the predetermined block.

16. Decoder of claim 15,

Configured so that the number of transform partitions (300) per sub-partition is one in case of dimensions of the predetermined block exceeding a predetermined threshold, and larger than one in case of dimensions of the predetermined block not exceeding a predetermined threshold.

17. Decoder of any of claims 14 to 16,

Configured so that

a number of sub-partitions in the predetermined block,

depends on dimensions of the predetermined block and/or the predetermined dimension

(104).

18. Decoder of claim 17,

Configured so that the number of sub-partitions (300) in the predetermined block is one in case of dimensions of the predetermined block falling below a further predetermined threshold, and larger than one in case of dimensions of the predetermined block not falling below the further predetermined threshold.

19. Decoder of any of claims 17 to 18,

Configured so that

a number of sub-partitions in the predetermined block,

depends on dimensions of the predetermined block in that same equals a first number in case of the dimensions of the predetermined block assuming a first width and first height and a second number different from the first in case of the dimensions of the predetermined block assuming a second width equaling the first height and second height equaling the first width.

21. Encoder for block-based encoding of a picture (12) into a data stream (14), configured to

encode an intra-coding mode (116) for a predetermined block (80) of the picture into the data stream (14);

encode a partition dimension flag (114) for the predetermined block of the picture into the data stream which signals that a partition dimension (104) is to be set to be horizontal or vertical,

partition, along the predetermined dimension (104), the predetermined block (80) into transform partitions (300) which are as wide as the predetermined block perpendicular to predetermined dimension;

intra-predicting the predetermined block depending on one or more already reconstructed samples neighboring the predetermined block in a manner depending on the intra-coding mode to obtain a predictor for the predetermined block;

encode, for each transform partition, a transform (182) of a prediction residual into the data stream, so that the predetermined block is reconstructible by correcting (122) the predictor within each transform partition using the transform of the prediction residual encoded for the respective transform partition.

22. Encoder of claim 1 ,

wherein the number of transform partitions is greater than 2 and/or the transform partitions are one sample wide along the predetermined dimension.

23. Encoder of claim 21 or 22,

configured to transform for each transform partition the prediction residual in the respective partition into spectral domain for usage in correcting the predictor within the respective transform partition.

24. Encoder of any of claims 21 to 23

wherein the encoder is configured to encode, for each partition, the transform into the data stream by

encoding a coded transform partition flag (188) into the data stream;

wherein, if the coded transform partition flag is not set, the coded transform partition flag (188) signals that the prediction residual for the respective transform partition is zero, and

if the coded transform partition flag is set, the encoder is configured to encode transform coefficients of the transform of the prediction residual of the respective transform partition into the data stream.

25. Encoder of claim 24,

configured to encode the coded transform partition flags for the transform partitions into the data stream sequentially except the coded transform partition flag for a last transform partition in transform partition order, if all preceding coded transform partition flags are not set, which is then inferred to be set.

26. Encoder of claim 25,

configured to encode the coded transform partition flag for a respective transform partition into the data stream by use of context-dependent entropy encoding using a context which depends on the coded transform partition flag encoded for a preceding transform partition preceding the respective transform partition in the predetermined transform partition order.

27. Encoder of any of claims 21 to 26,

configured to

encode, for a predetermined transform partition, the transform of the prediction residual of the predetermined partition into the data stream by

encoding a last position indication (190) into the data stream indicating a last transform coefficient position of the transform along a predetermined scan order scanning transform coefficients of the one-dimensional transform; and

encoding transform coefficients (198) of the transform up to the last transform coefficient position along the predetermined scan order into the data stream, wherein transform coefficients of the transform beyond the last transform coefficient position along the predetermined scan order are zero are inferred to be zero.

28. Encoder of claim 27, wherein the transform partitions are one sample wide along to the predetermined dimension and the transform is a one-dimensional transform.

29. Encoder of any of claims 21 to 28,

wherein

the transform is a DCT transform in case of the intra prediction mode not being a planar mode, and a DST transform in case of the intra prediction mode being the planar mode, or

the transform is a linear transform a type of which is selected based on the intra prediction mode, a block size of the predetermined block and/or an dedicated syntax element.

30. Encoder of any of claims 21 to 29,

configured to

encode a split mode flag (160) for the predetermined block of the picture into the data stream;

if the split mode flag indicates a first split mode, perform the encoding the partition dimension flag, the partitioning and the encoding of the transform for each transform partition;

if the split mode flag indicates a second split mode, instead of the encoding the partition dimension flag, the partitioning and the encoding of the transform for each transform partition, encode one transform of the prediction residual within the predetermined block.

31. Encoder of any of the previous claims 21 to 30,

configured to

encode the partition dimension flag (114) by use of context-dependent entropy encoding using a context which depends on the intra-coding mode.

32. Encoder of any of the previous claims 21 to 31 ,

configured to

encode the partition dimension flag (114) by use of context-dependent entropy encoding using one of three contexts comprising

the intra-coding mode signaling a non-angular mode,

the intra-coding mode signaling a horizontal mode,

the intra-coding mode signaling a vertical mode.

33. Encoder of any of claims 21 to 32,

configured to set a width of the transform partitions, measured along the predetermined dimension (104),

depending on a size of the predetermined block (80) along predetermined dimension (104) and/or

depending on the intra-coding mode for the predetermined block and/or

depending on whether the intra-coding mode for the predetermined block is an angular mode or not.

34. Encoder of any of claims 21 to 33

configured to perform the intra-predictirig the predetermined block depending on one or more already reconstructed samples neighboring the predetermined block in a manner depending on the intra-coding mode to obtain a predictor for the predetermined block and determine prediction residual of the predetermined block for correcting (122) the predictor within each transform partition using the transform of the prediction residual to be encoded for the respective transform partition by

sequentially subjecting groups of transform partitions into which the transform partitions of the predetermined block are grouped to form a subpartition (102; 112) for each group of transform partitions according to a predetermined sub-partition order (126) which sequentially traverses the sub-partitions along the predetermined dimension, to a prediction so as to derive a predictor for a current sub-partition by filling the current sub-partition depending on one or more already reconstructed samples neighboring the current sub-partition in a manner depending on the intra-coding mode, determining the transform of the prediction residual within each transform partition comprised by the group of transform partitions forming the current sub-partition so as to serve, before proceeding with a subsequent subpartition, for reconstructing the current sub-partition by correcting (122) the predictor within each transform partition comprised by the group of transform partitions forming the current sub-partition using the transform of the respective transform partition.

35. Encoder of claim 34,

Configured so that

a number of transform partitions (300) per sub-partition

depends on dimensions of the predetermined block.

36. Encoder of claim 35,

Configured so that the number of transform partitions (300) per sub-partition is one in case of dimensions of the predetermined block exceeding a predetermined threshold, and larger than one in case of dimensions of the predetermined block not exceeding a predetermined threshold

37. Encoder of any of claims 34 to 36,

Configured so that

a number of sub-partitions in the predetermined block,

depends on dimensions of the predetermined block and/or the predetermined dimension

(104),

38. Encoder of claim 37,

Configured so that the number of sub-partitions (300) in the predetermined block is one in case of dimensions of the predetermined block falling below a further predetermined threshold, and larger than one in case of dimensions of the predetermined block not falling below the further predetermined threshold.

39. Encoder of any of claims 37 to 38,

Configured so that

a number of sub-partitions in the predetermined block,

depends on dimensions of the predetermined block in that same equals a first number in case of the dimensions of the predetermined block assuming a first width and first height and a second number different from the first in case of the dimensions of the predetermined block assuming a second width equaling the first height and second height equaling the first width.

40. Method for block-based decoding of a picture (12) from a data stream (14), comprising

decoding an intra-coding mode (116) for a predetermined block (80) of the picture from the data stream (14);

decoding a partition dimension flag (1 14) for the predetermined block of the picture from the data stream and setting a partition dimension (104) depending on the partition dimension flag to be horizontal or vertical,

partitioning, along the predetermined dimension (104), the predetermined block (80) into transform partitions (300) which are as wide as the predetermined block perpendicular to predetermined dimension;

decoding, for each transform partition, a transform (182) of a prediction residual from the data stream;

intra-predicting the predetermined block depending on one or more already reconstructed samples neighboring the predetermined block in a manner depending on the intra-coding mode to obtain a predictor for the predetermined block; and

reconstructing the predetermined block by correcting (122) the predictor within each transform partition using the transform of the prediction residual decoded for the respective transform partition.

41 Method for block-based encoding of a picture (12) into a data stream (14), comprising

encoding an intra-coding mode (116) for a predetermined block (80) of the picture into the data stream (14);

encoding a partition dimension flag (114) for the predetermined block of the picture into the data stream which signals that a partition dimension (104) is to be set to be horizontal or vertical,

partitioning, along the predetermined dimension (104), the predetermined block (80) into transform partitions (300) which are as wide as the predetermined block perpendicular to the predetermined dimension;

intra-predicting the predetermined block depending on one or more already reconstructed samples neighboring the predetermined block in a manner depending on the intra-coding mode to obtain a predictor for the predetermined block;

encoding, for each transform partition, a transform (182) of a prediction residual into the data stream, so that the predetermined block is reconstructive by correcting (122) the predictor within each transform partition using the transform of the prediction residual encoded for the respective transform partition.

42. Data stream generated by the method of claim 41.

43. A computer program having a program code for performing, when running on a computer, a method according to claim 40 and/or 41.