Affine mode in video coding and decoding
Video coding and decoding in which the bitstream is generated using affine motion compensation mode are described. One example includes generating, for a reconstruction of a current video unit, a list of affine motion vector prediction candidates in which each candidate in the list is calculated without using a clipping operation, wherein the clipping operation limits a magnitude of the candidate motion vectors to a range; and performing the reconstruction of the current video unit using the list of affine motion vector prediction candidates.
1. A method of video processing; comprising:
determining, for a conversion between a current video unit of a video and a bitstream of the video, multiple motion vector differences (MVDs) coded with different precisions, wherein the multiple MVDs include a first MVD with a first precision and a second MVD with a second precision, the first precision is different from the second precision, and the multiple MVDs are associated with different control points of the current video unit coded in an affine mode;
wherein, due to the second MVD being predicted from the first MVD, the second precision is greater than the first precision;
performing the conversion based on the determining.
2. The method of claim 1 , wherein for a 4-parameter affine model used in the affine video coding mode, the first MVD is a MVD of a top left control point in the current video unit and the second MVD is a MVD of a top right control point in the current video unit.
3. The method of claim 1 , wherein when a 6-parameter affine model is used in the affine video coding mode, the multiple MVDs further includes a third MVD having a third precision, further comprising:
due to the second MVD and/or the third MVD is predicted from the first MVD, the second precision and/or the third precision is greater than the first precision.
4. The method of claim 3 , wherein the first MVD is a MVD of a top left control point in the current video unit, the second MVD is a MVD of a top right control point in the current video unit, the third MVD is a MVD of a left bottom control point in the current video unit.
5. The method of claim 1 , further comprising:
selectively enabling or disabling signaling, based on one or more coded mode information, an indication of usage of the different precisions of the multiple MVDs in the coded representation of the current video unit.
6. The method of claim 1 , wherein the first precision and/or the second precision is signaled in the bitstream of the video.
7. The method of claim 1 , wherein the indication of usage is signaled implicitly or explicitly in one or more of the following: a sequence parameter set (SPS), a sequence header, a video parameter set (VPS), a picture parameter set (PSP), a slice header, a picture header, a coding tree unit (CTU), a coding unit (CU), a group of CTU rows, or a group of CU rows.
8. The method of claim 1 , wherein the one or more coded mode information relate to at least one of: a size of the video current unit, a shape of the current video unit, a coding mode associated with the current video unit, a number of affine parameters associated with the current video unit, a slice type associated with the video unit, or a number of temporal layers associated with the current video unit.
9. The method of claim 5 , wherein the indication of usage is selectively disabled when a size of the current video unit is less than a threshold number of samples.
10. The method of claim 9 , wherein the threshold number of samples is 16.
11. The method of claim 1 , wherein a first motion information for motion compensation is derived at a first block size, and a second motion information for storing is derived at a second block size, the first block size being different from the second block size.
12. The method of claim 1 , wherein the conversion includes encoding the current video unit into the bitstream.
13. The method of claim 1 , wherein the conversion includes decoding the current video unit from the bitstream.
14. The method of claim 11 , wherein the first block size is M1×N1 and the second block size is M2×N2, M1, M2, N1 and N2 being positive integers, M1 being unequal to M2, and N1 being unequal to N2.
15. The method of claim 14 , wherein M1 and N1 are equal to 8, and M2 and N2 are equal to 4.
16. The method of claim 11 , wherein the current video unit includes a sub-block of a first type and/or a sub-block of a second type, and the sub-block of the first type is located in an innermost region of the current video unit and has no common boundary with the current video unit, and the sub-block of the second type shares a boundary in common with the current video unit.
17. The method of claim 16 , wherein the first block size is M1×N1, and the second block size is M2×N2 or M3×N3 respectively based on whether the sub-block of the first type or the sub-block of the second type is processed, wherein M1, M2, M3, N1, N2 and N3 are positive integers.
18. An apparatus for processing video data comprising a processor and a non-transitory memory with instructions thereon, wherein the instructions upon execution by the processor, cause the processor to:
determine, for a conversion between a current video unit of a video and a bitstream of the video, multiple motion vector differences (MVDs) coded with different precisions, wherein the multiple MVDs include a first MVD with a first precision and a second MVD with a second precision, the first precision is different from the second precision, and the multiple MVDs are associated with different control points of the current video unit coded in an affine mode;
wherein, due to the second MVD being predicted from the first MVD, the second precision is greater than the first precision;
perform the conversion based on the determining.
19. A non-transitory computer-readable recording medium storing a bitstream of a video which is generated by a method performed by a video processing apparatus, wherein the method comprises:
determining, for a conversion between a current video unit of the video and the bitstream of the video, multiple motion vector differences (MVDs) coded with different precisions, wherein the multiple MVDs include a first MVD with a first precision and a second MVD with a second precision, the first precision is different from the second precision, and the multiple MVDs are associated with different control points of the current video unit coded in an affine mode;
wherein, due to the second MVD being predicted from the first MVD, the second precision is greater than the first precision;
generating the bitstream from the current video unit based on the determining.
20. The apparatus of claim 18 , wherein the first MVD is a MVD of a top left control point in the current video unit and the second MVD is a MVD of a top right control point in the current video unit.