Method and apparatus for video coding
Aspects of the disclosure provide methods and apparatuses for video encoding/decoding. In some examples, an apparatus for video decoding includes processing circuitry. The processing circuitry can decode coding information of a coefficient block from a coded video bitstream. The coding information can indicate a size of the coefficient block. The processing circuitry can determine, based on the size of the coefficient block, an order in which inverse horizontal and inverse vertical transforms of an inverse primary transform are to be performed on transform coefficients of the coefficient block to obtain residual data of a residual block. When the size of the coefficient block satisfies a condition, the inverse vertical transform is performed after the inverse horizontal transform is performed on the transform coefficients of the coefficient block. The processing circuitry can reconstruct a sample in the residual block based on the residual data.
1. A method for video decoding in a decoder, comprising:
decoding coding information of a coefficient block from a coded video bitstream, the coding information indicating a size of the coefficient block;
determining, based on the size of the coefficient block, whether to reduce a number of calculations in one of inverse horizontal and inverse vertical transforms of an inverse primary transform, the inverse vertical transform transforming transform coefficients of the coefficient block to intermediate data of an intermediate block, and the inverse horizontal transform transforming the intermediate data to residual data in a residual block;
performing the inverse primary transform, wherein:
when the number of calculations in the inverse vertical transform is determined to be reduced, top 16 rows of the intermediate data in the intermediate block are calculated by the inverse vertical transform and the remaining intermediate data in the intermediate block are zero; and
when the number of calculations in the inverse horizontal transform is determined to be reduced, left 16 columns of the residual data in the residual block are calculated by the inverse horizontal transform and the remaining residual data in the residual block are zero; and
reconstructing a sample in the residual block based on the residual data.
2. The method of claim 1 , wherein
the size of the coefficient block is 32×64;
the one of the inverse horizontal and inverse vertical transforms is the inverse vertical transform;
the determining includes determining that the number of calculations in the inverse vertical transform is to be reduced when the size of the coefficient block is 32×64; and
the performing the inverse primary transform includes performing the inverse primary transform, the top 16 rows of the intermediate data in the intermediate block being calculated by the inverse vertical transform and the remaining intermediate data in the intermediate block being zero.
3. The method of claim 1 , wherein
the size of the coefficient block is 32×64;
the one of the inverse horizontal and inverse vertical transforms is the inverse horizontal transform;
the determining includes determining that the number of calculations in the inverse horizontal transform is to be reduced when the size of the coefficient block is 32×64; and
the performing the inverse primary transform includes performing the inverse primary transform, the left 16 columns of the residual data in the residual block being calculated by the inverse horizontal transform and the remaining residual data in the residual block being zero.
4. The method of claim 1 , wherein:
the size of the coefficient block is 32×32;
the one of the inverse horizontal and inverse vertical transforms is the inverse horizontal transform;
the determining includes determining that the number of calculations in the inverse horizontal transform is to be reduced when the size of the coefficient block is 32×32; and
the performing the inverse primary transform includes performing the inverse primary transform, the left 16 columns of the residual data in the residual block being calculated by the inverse horizontal transform, the remaining residual data in the residual block being zero, and the intermediate data in the intermediate block being calculated by the inverse vertical transform.
5. The method of claim 1 , wherein
the size of the coefficient block is 32×32;
the one of the inverse horizontal and inverse vertical transforms is the inverse vertical transform;
the determining includes determining that the number of calculations in the inverse vertical transform is to be reduced when the size of the coefficient block is 32×32; and
the performing the inverse primary transform includes performing the inverse primary transform, the top 16 rows of the intermediate data in the intermediate block being calculated by the inverse vertical transform, the remaining intermediate data in the intermediate block being zero, and the residual data in the residual block being calculated by the inverse horizontal transform.
6. An apparatus for video decoding, comprising processing circuitry configured to:
decode coding information of a coefficient block from a coded video bitstream, the coding information indicating a size of the coefficient block;
determine, based on the size of the coefficient block, whether to reduce a number of calculations in one of inverse horizontal and inverse vertical transforms of an inverse primary transform, the inverse vertical transform transforming transform coefficients of the coefficient block to intermediate data of an intermediate block, and the inverse horizontal transform transforming the intermediate data to residual data in a residual block;
perform the inverse primary transform, wherein:
when the number of calculations in the inverse vertical transform is determined to be reduced, top 16 rows of the intermediate data in the intermediate block are calculated by the inverse vertical transform and the remaining intermediate data in the intermediate block are zero; and
when the number of calculations in the inverse horizontal transform is determined to be reduced, left 16 columns of the residual data in the residual block are calculated by the inverse horizontal transform and the remaining residual data in the residual block are zero; and
reconstruct a sample in the residual block based on the residual data.
7. The apparatus of claim 6 , wherein
the size of the coefficient block is 32×64;
the one of the inverse horizontal and inverse vertical transforms is the inverse vertical transform; and
the processing circuitry is further configured to:
determine that the number of calculations in the inverse vertical transform is to be reduced when the size of the coefficient block is 32×64; and
perform the inverse primary transform, wherein
the top 16 rows of the intermediate data in the intermediate block are calculated by the inverse vertical transform and the remaining intermediate data in the intermediate block are zero.
8. The apparatus of claim 6 , wherein
the size of the coefficient block is 32×64;
the one of the inverse horizontal and inverse vertical transforms is the inverse horizontal transform; and
the processing circuitry is further configured to:
determine that the number of calculations in the inverse horizontal transform is to be reduced when the size of the coefficient block is 32×64; and
perform the inverse primary transform, wherein
the left 16 columns of the residual data in the residual block are calculated by the inverse horizontal transform and the remaining residual data in the residual block are zero.
9. The apparatus of claim 6 , wherein:
the size of the coefficient block is 32×32;
the one of the inverse horizontal and inverse vertical transforms is the inverse horizontal transform; and
the processing circuitry is further configured to:
determine that the number of calculations in the inverse horizontal transform is to be reduced when the size of the coefficient block is 32×32; and
perform the inverse primary transform, wherein
the left 16 columns of the residual data in the residual block are calculated by the inverse horizontal transform, the remaining residual data in the residual block are zero, and the intermediate data in the intermediate block are calculated by the inverse vertical transform.
10. The apparatus of claim 6 , wherein
the size of the coefficient block is 32×32;
the one of the inverse horizontal and inverse vertical transforms is the inverse vertical transform; and
the processing circuitry is further configured to:
determine that the number of calculations in the inverse vertical transform is to be reduced when the size of the coefficient block is 32×32; and
perform the inverse primary transform, wherein
the top 16 rows of the intermediate data in the intermediate block are calculated by the inverse vertical transform, the remaining intermediate data in the intermediate block are zero, and the residual data in the residual block are calculated by the inverse horizontal transform.