Scalable video coding using base-layer hints for enhancement layer motion parameters

1. A scalable video decoder comprising:

an entropy decoder configured to perform, using a processor, entropy decoding of information from a data stream comprising encoded data of a video to obtain a base residual signal and an enhancement residual signal; and

a predictive decoder configured to process, using the processor, the entropy-decoded information to reconstruct an enhancement layer signal of the video based on block-wise predictive coding that includes:

gathering, for a current block of a frame of the enhancement layer signal, a set of motion parameter candidates from already reconstructed neighboring blocks of the frame, the reconstructed neighboring blocks neighboring the current block,

gathering a set of one or more base layer motion parameters of a block of a base layer signal, the block of the base layer signal being collocated to the current block of the frame of the enhancement layer signal,

obtaining an extended motion parameter candidate set for the current block by incorporating the set of one or more base layer motion parameters or a scaled version thereof into the set of motion parameter candidates,

extracting, from the data stream, a syntax element corresponding to the current block identifying one of the motion parameter candidates of the extended motion parameter candidate set to be selected for the current block,

selecting, for the current block, the one of the motion parameter candidates of the extended motion parameter candidate set based on the syntax element,

generating, by motion compensated prediction, a prediction of the current block of the enhancement layer signal using the selected one of the motion parameter candidates of the extended motion parameter candidate set, and

reconstructing at least a portion of the enhancement layer signal based on the prediction of the current block and the enhancement residual signal,

wherein the entropy decoder is configured to, in decoding a transform coefficient at a current position in a transform coefficient block, determine a context model for decoding the transform coefficient using a template or information regarding spectral decomposition of the base residual signal, the template involving positions in one or more transform blocks in the base residual signal that correspond to the current position in terms of spectral frequency.

2. The scalable video decoder according to claim 1 , wherein the base layer motion parameters are scaled according to a spatial resolution ratio between the base layer signal and the enhancement layer signal to acquire the scaled version of the base layer motion parameters.

3. The scalable video decoder according to claim 1 , the gathering, for the block of the frame of the enhancement layer signal, of the set of enhancement layer motion parameter candidates is performed from temporally and/or spatially neighboring already reconstructed blocks of the enhancement layer by copying and/or combining motion parameters of the neighboring already reconstructed blocks.

4. The scalable video decoder according to claim 1 , wherein the block of the base layer signal, collocated to the block of the frame of the enhancement layer signal, is identified by checking as to which block of a temporally corresponding frame of the base layer signal overlays a position spatially corresponding a predetermined sample position of the block of the frame of the enhancement layer signal.

5. The scalable video decoder according to claim 1 , the extended motion parameter candidate set is ordered depending on the base layer motion parameters.

6. A scalable video decoding method comprising:

entropy decoding of information from a data stream comprising encoded data of a video to obtain a base residual signal and an enhancement residual signal; and

reconstructing an enhancement layer signal of the video using block-wise predictive coding comprising:

gathering, for a current block of a frame of the enhancement layer signal, a set of motion parameter candidates from already reconstructed neighboring blocks of the frame, the reconstructed neighboring blocks neighboring the current block,

gathering a set of one or more base layer motion parameters of a block of a base layer signal, the block of the base layer signal being collocated to the current block of the frame of the enhancement layer signal,

obtaining an extended motion parameter candidate set for the current block by incorporating the set of one or more base layer motion parameters or a scaled version thereof into the set of motion parameter candidates,

extracting, from the data stream, a syntax element corresponding to the current block identifying one of the motion parameter candidates of the extended motion parameter candidate set to be selected for the current block,

selecting, for the current block, the one of the motion parameter candidates of the extended motion parameter candidate set based on the syntax element,

generating, by motion compensated prediction, a prediction of the current block of the enhancement layer signal using the selected one of the motion parameter candidates of the extended motion parameter candidate set, and

reconstructing at least a portion of the enhancement layer signal based on the prediction of the current block and the enhancement residual signal,

wherein the entropy decoding includes decoding a transform coefficient at a current position in a transform coefficient block by determining a context model for decoding the transform coefficient using a template or information regarding spectral decomposition of the base residual signal, the template involving positions in one or more transform blocks in the base residual signal that correspond to the current position in terms of spectral frequency.

7. A non-transitory computer readable medium including a computer program comprising a program code for performing, when running on a computer, a method according to claim 6 .

8. A scalable video encoder comprising:

an encoder configured to encode, using a processor, an enhancement layer signal of a video based on block-wise predictive coding that includes:

gathering, for a current block of a frame of the enhancement layer signal, a set of motion parameter candidates from already reconstructed neighboring blocks of the frame, the reconstructed neighboring blocks neighboring the current block,

gathering a set of one or more base layer motion parameters of a block of a base layer signal, the block of the base layer signal being collocated to the current block of the frame of the enhancement layer signal,

obtaining an extended motion parameter candidate set for the current block by incorporating the set of one or more base layer motion parameters or a scaled version thereof into the set of motion parameter candidates,

selecting one of the motion parameter candidates of the extended motion parameter candidate set based on a syntax element corresponding to the current block to be signaled in a data stream, the syntax element specifically indicating the one of the motion parameter candidates of the extended motion parameter candidate set to be selected for the current block,

generating, by motion compensated prediction, a prediction of the current block of the enhancement layer signal using the selected one of the motion parameter candidates of the extended motion parameter candidate set, and

obtaining an enhancement residual signal for the current block based on a difference of the current block and the prediction of the current block; and

an entropy encoder configured to entropy encode, using the processor, a base residual signal and the enhancement residual signal and the syntax element to generate the data stream including encoded data of the video,

wherein the entropy encoder is configured to, in encoding a transform coefficient at a current position in a transform coefficient block, determine a context model for encoding the transform coefficient using a template or information regarding spectral decomposition of the base residual signal, the template involving positions in one or more transform blocks in the base residual signal that correspond to the current position in terms of spectral frequency.

9. A scalable video encoding method comprising:

encoding an enhancement layer signal of a video based on block-wise predictive coding that includes:

gathering, for a current block of a frame of the enhancement layer signal, a set of motion parameter candidates from already reconstructed neighboring blocks of the frame, the reconstructed neighboring blocks neighboring the current block,

gathering a set of one or more base layer motion parameters of a block of a base layer signal, the block of the base layer signal being collocated to the current block of the frame of the enhancement layer signal,

obtaining an extended motion parameter candidate set for the current block by incorporating the set of one or more base layer motion parameters or a scaled version thereof into the set of motion parameter candidates,

selecting one of the motion parameter candidates of the extended motion parameter candidate set based on a syntax element corresponding to the current block to be signaled in a data stream, the syntax element specifically indicating the one of the motion parameter candidates of the extended motion parameter candidate set to be selected for the current block,

generating, by motion compensated prediction, a prediction of the current block of the enhancement layer signal using the selected one of the motion parameter candidates of the extended motion parameter candidate set, and

obtaining an enhancement residual signal for the current block based on a difference of the current block and the prediction of the current block; and

entropy encoding a base residual signal and the enhancement residual signal and the syntax element to generate the data stream including encoded data of the video, wherein the entropy encoding includes, in encoding a transform coefficient at a current position in a transform coefficient block, determining a context model for encoding the transform coefficient using a template or information regarding spectral decomposition of the base residual signal, the template involving positions in one or more transform blocks in the base residual signal that correspond to the current position in terms of spectral frequency.

10. A non-transitory computer readable medium including a computer program comprising a program code for performing, when running on a computer, a method according to claim 9 .