Method and apparatus for video coding
Aspects of the disclosure provide methods and apparatuses for video coding. In some examples, an apparatus includes processing circuitry configured to encode a first block in a picture according to one of bi-directional prediction and uni-directional prediction. The processing circuitry is configured to add motion information and a bi-prediction weight index of a History-based Motion Vector Prediction (HMVP) candidate to an HMVP list based on whether the first block is encoded according to the bi-directional prediction or the uni-directional prediction, the bi-prediction weight index indicating bi-prediction weights of the bi-directional prediction for the first block when the first block is encoded according to the bi-directional prediction, and the bi-prediction weight index indicating a default value when the first block is encoded according to the uni-directional prediction. The processing circuitry is configured to encode a second block in the picture based on a plurality of candidates that includes the HMVP candidate.
1. A method for video encoding in an encoder, the method comprising:
encoding a first block in a picture;
when the first block is encoded according to bi-directional prediction,
adding motion information of the first block and a first bi-prediction weight index of the first block as a History-based Motion Vector Prediction (HMVP) candidate to an HMVP list, the first bi-prediction weight index indicating bi-prediction weights used for performing the bi-directional prediction for the first block;
when the first block is encoded according to uni-directional prediction,
the first bi-prediction weight index indicates equal bi-prediction weights;
encoding a second block in the picture based on a plurality of candidates that includes the HMVP candidate; and
when the second block is encoded according to the bi-directional prediction,
storing the motion information of the HMVP candidate as a merge candidate, and
setting a second bi-prediction weight index to the first bi-prediction weight index.
2. The method of claim 1 , wherein the setting the second bi-prediction weight index further comprises:
when the second block in the picture is encoded based on another motion information of a candidate of the plurality of candidates that is neither a regular spatial merge candidate nor from the HMVP list, and the second block is encoded according to the bi-directional prediction,
setting the second bi-prediction weight index for the second block to the first bi-prediction weight index.
3. The method of claim 1 , wherein the setting the second bi-prediction weight index further comprises:
when the first block is in a Coding Tree Unit (CTU) row different from that of a current CTU in which the second block is included, the second block in the picture being encoded based on another motion information stored in a regular merge candidate or an affine merge candidate, and the second block is encoded according to the bi-directional prediction,
setting the second bi-prediction weight index for the second block to the first bi-prediction weight index.
4. The method of claim 1 , wherein the setting the second bi-prediction weight index further comprises:
when the first block is outside a current CTU in which the second block is included, the second block in the picture is encoded based on another motion information stored in a translational merge candidate or an inherited affine merge candidate, and the second block being encoded according to the bi-directional prediction,
setting the second bi-prediction weight index for the second block to the first bi-prediction weight index.
5. The method of claim 1 , wherein
the first block is encoded according to one of the bi-directional prediction and the uni-directional prediction with the picture as a reference picture.
6. The method of claim 1 , wherein the first block is encoded according to the bi-directional prediction, and both of a first weight applicable to a first reference picture in a first list and a second weight applicable to a second reference picture in a second list are positive when the first and second reference pictures correspond to different reference pictures.
7. The method of claim 1 , wherein
the first block is encoded according to the bi-directional prediction, and
one of a first weight applicable to a first reference picture in a first list and a second weight applicable to a second reference picture in a second list is negative when the first and second reference pictures correspond to a same reference picture.
8. The method of claim 1 , wherein
the bi-prediction weights are used for performing the bi-directional prediction for the first block and are determined according to a weight parameter indicated by the bi-prediction weight index of the HMVP candidate and a precision factor F, and the precision factor F is 8.
9. A method for video decoding in a decoder, the method comprising:
receiving prediction information of a first block in a picture from a coded video bitstream;
generating reconstructed samples of the first block for output according to the prediction information and one of bi-directional prediction and uni-directional prediction;
when the first block is coded according to the bi-directional prediction,
adding motion information of the first block and a first bi-prediction weight index of the first block as a History-based Motion Vector Prediction (HMVP) candidate to an HMVP list, the first bi-prediction weight index indicating bi-prediction weights used for performing the bi-directional prediction for the first block;
when the first block is coded according to the uni-directional prediction, the first bi-prediction weight index indicates equal bi-prediction weights;
decoding a second block in the picture based on a plurality of candidates that includes the HMVP candidate; and
when the second block is coded according to the bi-directional prediction,
storing the motion information of the HMVP candidate as a merge candidate, and
setting a second bi-prediction weight index to the first bi-prediction weight index.
10. The method of claim 9 , wherein
the second bi-prediction weight index is set to the first bi-prediction weight index when the first block is in a Coding Tree Unit (CTU) row different from that of a current CTU in which the second block is included, the second block in the picture is determined to be coded based on another motion information stored in a regular merge candidate or an affine merge candidate, and the second block is coded according to the bi-directional prediction.
11. The method of claim 9 , wherein
the second bi-prediction weight index is set to the first bi-prediction weight index when the first block is outside a current CTU in which the second block is included, the second block in the picture is determined to be coded based on another motion information stored in a translational merge candidate or an inherited affine merge candidate, and the second block is coded according to the bi-directional prediction.
12. The method of claim 9 , wherein
the first block is coded according to the one of the bi-directional prediction and the uni-directional prediction with the picture as a reference picture.
13. The method of claim 9 , wherein
the first block is coded according to the bi-directional prediction, and
both of a first weight applicable to a first reference picture in a first list and a second weight applicable to a second reference picture in a second list are positive when the first and second reference pictures correspond to different reference pictures.
14. The method of claim 9 , wherein
the bi-prediction weights are used for performing the bi-directional prediction for the first block and are determined according to a weight parameter indicated by the bi-prediction weight index of the HMVP candidate and a precision factor F, and the precision factor F is 8.
15. The method of claim 9 , wherein
the first block is coded according to the bi-directional prediction, and
one of a first weight applicable to a first reference picture in a first list and a second weight applicable to a second reference picture in a second list is negative when the first and second reference pictures correspond to a same reference picture.
16. The method of claim 9 , wherein
when the first block is coded according to the bi-directional prediction,
a first weight w 1 applicable to a first reference picture in a first list is determined according to
w
1
=
w
F
,
and
a second weight w 0 applicable to a second reference picture in a second list is determined according to
w 0 =1− w 1 ,
where w and F are integers, w represents a weight parameter indicated by the first bi-prediction weight index, and F represents a precision factor.
17. A method of processing visual media data, the method comprising:
processing a bitstream that includes the visual media data according to a format rule, wherein
the bitstream includes prediction information of a first block in a picture,
the format rule specifies that:
reconstructed samples of the first block are generated for output according to the prediction information and one of bi-directional prediction and uni-directional prediction;
when the first block is coded according to the bi-directional prediction,
motion information of the first block and a first bi-prediction weight index of the first block are added as a History-based Motion Vector Prediction (HMVP) candidate to an HMVP list, the first bi-prediction weight index indicating bi-prediction weights used for performing the bi-directional prediction for the first block;
when the first block is coded according to the uni-directional prediction, the first bi-prediction weight index indicates equal bi-prediction weights;
a second block in the picture is decoded based on a plurality of candidates that includes the HMVP candidate; and
when the second block is coded according to the bi-directional prediction,
the motion information of the HMVP candidate is stored as a merge candidate, and
a second bi-prediction weight index is set to the first bi-prediction weight index.
18. The method of claim 9 , wherein
the second bi-prediction weight index is set to the first bi-prediction weight index when the second block in the picture is determined to be coded based on another motion information stored in a candidate that is neither a regular spatial merge candidate nor from the HMVP list, and the second block is coded according to the bi-directional prediction.