Non-separable secondary transform for video coding with reorganizing

A method for encoding video data, the method comprising:

• generating one or more predictive blocks by performing intra prediction based on an intra prediction mode and data from one or more previously encoded blocks of a current picture;
• forming a residual video block of the current picture, wherein forming the residual video block comprises subtracting the one or more predictive blocks from a video block being encoded;
• applying a first transform to at least part of the residual video block to generate a first 2-dimensional coefficient block, the first transform converting the residual video block from a pixel domain to a frequency domain;
• reorganizing, based on the intra prediction mode, the first 2-dimensional coefficient block as a first 1-dimensional coefficient vector, wherein the first 2-dimensional coefficient block is a first 4×4 coefficient group, wherein the first 1-dimensional coefficient block is a first 16×1 coefficient vector, and wherein reorganizing the first 2-dimensional coefficient block as the first 1-dimensional coefficient vector comprises reorganizing coefficients of the first 4×4 coefficient group into the first 16×1 coefficient vector;
• applying a second transform to at least part of the first 1-dimensional coefficient vector to generate a second 1-dimensional coefficient vector, the second transform being a non-separable transform;
• and
• reorganizing the second 1-dimensional coefficient vector as a second 2-dimensional coefficient block according to a coefficient scanning order, wherein the second 1-dimensional coefficient block is a second 16×1 coefficient vector, wherein the second 2-dimensional coefficient block is a second 4×4 coefficient group, and wherein reorganizing the second 1-dimensional coefficient vector as the second 2-dimensional coefficient block comprises reorganizing coefficients of the second 16×1 coefficient vector into the second 4×4 coefficient group.

An apparatus for decoding video data comprising:

• a memory configured to store the video data;
• and
• a video decoder configured to:
• reorganize a first 2-dimensional coefficient block as a first 1-dimensional coefficient vector according to a coefficient scanning order, wherein the first 2-dimensional coefficient block is a first 4×4 coefficient group, wherein the first 1-dimensional coefficient block is a first 16×1 coefficient vector, and wherein, to reorganize the first 2-dimensional coefficient block as the first 1-dimensional coefficient vector, the video decoder is configured to reorganize coefficients of the first 4×4 coefficient group into the first 16×1 coefficient vector;
• apply a first inverse transform to the first 1-dimensional coefficient vector to generate a second 1-dimensional coefficient vector, the first inverse transform being a non-separable transform;
• reorganize, based on an intra prediction mode, the second 1-dimensional coefficient vector as a second 2-dimensional coefficient block, wherein the second 1-dimensional coefficient block is a second 16×1 coefficient vector, wherein the second 2-dimensional coefficient block is a second 4×4 coefficient group, and wherein, to reorganize the second 1-dimensional coefficient vector as the second 2-dimensional coefficient block, the video decoder is configured to reorganize coefficients of the second 16×1 coefficient vector into the second 4×4 coefficient group;
• apply a second inverse transform to the second 2-dimensional coefficient block to generate a residual video block, the second inverse transform converting the second 2-dimensional coefficient block from a frequency domain to a pixel domain;
• generate one or more predictive blocks by performing intra prediction based on the intra prediction mode and data from one or more previously decoded blocks of a current picture;
• sum the residual video block with the one or more predictive blocks;
• and
• form a decoded video block of the current picture based on the sum of the residual video block with one or more predictive blocks.

An apparatus for encoding video data comprising:

• a memory configured to store the video data;
• and
• a video encoder configured to:
• generate one or more predictive blocks by performing intra prediction based on an intra prediction mode and data from one or more previously encoded blocks of a current picture;
• form a residual video block of the current picture, wherein forming the residual video block comprises subtracting the one or more predictive blocks from a video block being encoded;
• apply a first transform to at least part of the residual video block to generate a first 2-dimensional coefficient block, the first transform converting the residual video block from a pixel domain to a frequency domain;
• reorganize, based on the intra prediction mode, the first 2-dimensional coefficient block to generate a first 1-dimensional coefficient vector, wherein the first 2-dimensional coefficient block is a first 4×4 coefficient group, wherein the first 1-dimensional coefficient block is a first 16×1 coefficient vector, and wherein, to reorganize the first 2-dimensional coefficient block as the first 1-dimensional coefficient vector, the video encoder is configured to reorganize coefficients of the first 4×4 coefficient group into the first 16×1 coefficient vector;
• apply a second transform to at least part of the first 1-dimensional coefficient vector to generate a second 1-dimensional coefficient vector, the second transform being a non-separable transform;
• and
• reorganize the second 1-dimensional coefficient vector as a second 2-dimensional coefficient block according to a coefficient scanning order, wherein the second 1-dimensional coefficient block is a second 16×1 coefficient vector, wherein the second 2-dimensional coefficient block is a second 4×4 coefficient group, and wherein, to reorganize the second 1-dimensional coefficient vector, the video encoder is configured to reorganize coefficients of the second 16×1 coefficient vector into the second 4×4 coefficient group.

An apparatus for decoding video data comprising:

• means for reorganizing a first 2-dimensional coefficient block as a first 1-dimensional coefficient vector according to a coefficient scanning order, wherein the first 2-dimensional coefficient block is a first 4×4 coefficient group, wherein the first 1-dimensional coefficient block is a first 16×1 coefficient vector, and wherein the means for reorganizing the first 2-dimensional coefficient block comprises means for reorganizing coefficients of the first 4×4 coefficient group into the first 16×1 coefficient vector;
• means for applying a first inverse transform to the first 1-dimensional coefficient vector to generate a second 1-dimensional coefficient vector, the first inverse transform being a non-separable transform;
• means for reorganizing, based on an intra prediction mode, the second 1-dimensional coefficient vector as a second 2-dimensional coefficient block, wherein the second 1-dimensional coefficient block is a second 16×1 coefficient vector, wherein the second 2-dimensional coefficient block is a second 4×4 coefficient group, and wherein the means for reorganizing the second 1-dimensional coefficient block as the second 2-dimensional coefficient vector comprises means for reorganizing coefficients of the second 16×1 coefficient vector into the second 4×4 coefficient group;
• means for applying a second inverse transform to the second 2-dimensional coefficient block to generate a residual video block, the second inverse transform converting the second 2-dimensional coefficient block from a frequency domain to a pixel domain;
• means for generating one or more predictive blocks by performing intra prediction based on the intra prediction mode and data from one or more previously decoded blocks of a current picture;
• and
• means for forming a decoded video block of the current picture, wherein the means for forming the decoded video block comprises means for summing the residual video block with the one or more predictive blocks.

An apparatus for encoding video data comprising:

• means for generating one or more predictive blocks by performing intra prediction based on an intra prediction mode and data from one or more previously encoded blocks of a current picture;
• means for forming a residual video block of the current picture, wherein forming the residual video block comprises subtracting the one or more predictive blocks from a video block being coded;
• means for applying a first transform to at least part of the residual video block to generate a first 2-dimensional coefficient block, the first transform converting the residual video block from a pixel domain to a frequency domain;
• means for reorganizing, based on the intra prediction mode, the first 2-dimensional coefficient block to generate a first 1-dimensional coefficient vector, wherein the first 2-dimensional coefficient block is a first 4×4 coefficient group, wherein the first 1-dimensional coefficient block is a first 16×1 coefficient vector, and wherein the means for reorganizing the first 2-dimensional coefficient block as the first 1-dimensional coefficient vector comprises means for reorganizing coefficients of the first 4×4 coefficient group into the first 16×1 coefficient vector;
• means for applying a second transform to at least part of the first 1-dimensional coefficient vector to generate a second 1-dimensional coefficient vector, the second transform being a non-separable transform;
• and
• means for reorganizing the second 1-dimensional coefficient vector as a second 2-dimensional coefficient block according to a coefficient scanning order, wherein the second 1-dimensional coefficient block is a second 16×1 coefficient vector, wherein the second 2-dimensional coefficient block is a second 4×4 coefficient group, and wherein the means for reorganizing the second 1-dimensional coefficient vector comprises means for reorganizing coefficients of the second 16×1 coefficient vector into the second 4×4 coefficient group.

A non-transitory computer-readable storage medium storing instructions that, when executed, cause one or more processors of a device for decoding video data to:

• reorganize a first 2-dimensional coefficient block as a first 1-dimensional coefficient vector according to a coefficient scanning order, wherein the first 2-dimensional coefficient block is a first 4×4 coefficient group, wherein the first 1-dimensional coefficient block is a first 16×1 coefficient vector, and wherein, to reorganize the first 2-dimensional coefficient block, the instructions further cause the one or more processors to reorganize coefficients of the first 4×4 coefficient group into the first 16×1 coefficient vector;
• apply a first inverse transform to the first 1-dimensional coefficient vector to generate a second 1-dimensional coefficient vector, the first inverse transform being a non-separable transform;
• reorganize, based on an intra prediction mode, the second 1-dimensional coefficient vector as a second 2-dimensional coefficient block, wherein the second 1-dimensional coefficient block is a second 16×1 coefficient vector, wherein the second 2-dimensional coefficient block is a second 4×4 coefficient group, and wherein, to reorganize the second 1-dimensional coefficient block as the second 2-dimensional coefficient vector, the instructions further cause the one or more processors to reorganize coefficients of the second 16×1 coefficient vector into the second 4×4 coefficient group;
• apply a second inverse transform to the second 2-dimensional coefficient block to generate a residual video block, the second inverse transform converting the second 2-dimensional coefficient block from a frequency domain to a pixel domain;
• generate one or more predictive blocks by performing intra prediction based on the intra prediction mode and data from one or more previously decoded blocks of a current picture;
• and
• form a decoded video block of the current picture, wherein forming the decoded video block comprises summing the residual video block with the one or more predictive blocks.