Methods and devices for refining motion vector candidates
This disclosure relates generally to video coding/decoding and particularly to methods and systems for refining motion vector candidates. An example decoding method includes receiving a coded video bitstream; extracting, from the coded video bitstream, a first syntax element indicating whether a compound reference mode is applied for a current video block; extracting, from the coded video bitstream, a second syntax element indicating whether a motion vector difference (MVD) is joined signaled for the current video block; in response to the first syntax element indicating that the compound reference mode is applied for the current video block and the second syntax element indicating that the MVD is joined signaled for the current video block, refining, by the device, at least one MV predictor (MVP) in a MV candidate list based on whether the at least one MVP is symmetric; and decoding, by the device based on the MV candidate list, the current video block.
1. A method for decoding a current video block of a video bitstream, comprising;
receiving, by a device comprising a memory storing instructions and a processor in communication with the memory, a coded video bitstream;
extracting, by the device from the coded video bitstream, a first syntax element indicating whether a compound reference mode is applied for a current video block;
extracting, by the device from the coded video bitstream, a second syntax element indicating whether a motion vector difference (MVD) is joined signaled for the current video block;
when the first syntax element indicates that the compound reference mode is applied for the current video block and the second syntax element indicates that the MVD is joined signaled for the current video block, refining, by the device, at least one MV predictor (MVP) in a MV candidate list based on whether the at least one MVP is symmetric, wherein:
the at least one MVP in the MV candidate list comprises an MVP pair, and
the refining the at least one MVP in the MV candidate list comprises:
refining the MVP pair in the MV candidate list based on whether the MVP pair is symmetric or non-symmetric; and
decoding, by the device based on the MV candidate list, the current video block,
wherein:
the MVP pair comprises a first MVP and a second MVP,
the first MVP has a first POC distance between a first reference frame corresponding to the first MVP and a current frame,
the second MVP has a second POC distance between a second reference frame corresponding to the second MVP and the current frame,
the first POC distance is greater than the second POC distance, and
the refining the at least one MVP in the MV candidate list comprises:
when the MVP pair is non-symmetric, refining the MVP pair to be a symmetric MVP pair, by:
projecting the first MVP to the second reference frame to obtain a projected MVP, and
replacing the second MVP with the projected MVP.
2. The method of claim 1 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric, moving the MVP pair before all non-symmetric MVPs in the MV candidate list.
3. The method of claim 1 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric, moving the MVP pair to a front of the MV candidate list.
4. The method of claim 1 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric with equal picture order count (POC) distance between two reference frames and a current frame, moving the MVP pair to a front of the MV candidate list.
5. A method of video encoding, the method comprising:
receiving, by a device comprising a memory storing instructions and a processor in communication with the memory, a video bitstream;
determining, by the device based on the video bitstream, a first syntax element indicating whether a compound reference mode is applied for a current video block;
determining, by the device based on the video bitstream, a second syntax element indicating whether a motion vector difference (MVD) is joined signaled for the current video block;
encoding, by the device, the first syntax element in a coded bitstream;
encoding, by the device, the second syntax element in the coded bitstream;
when the first syntax element indicates that the compound reference mode is applied for the current video block and the second syntax element indicates that the MVD is joined signaled for the current video block, refining, by the device, at least one MV predictor (MVP) in a MV candidate list based on whether the at least one MVP is symmetric, wherein:
the at least one MVP in the MV candidate list comprises an MVP pair, and
the refining the at least one MVP in the MV candidate list comprises:
refining the MVP pair in the MV candidate list based on whether the MVP pair is symmetric or non-symmetric; and
encoding, by the device based on the MV candidate list, the current video block in the coded bitstream,
wherein:
the MVP pair comprises a first MVP and a second MVP,
the first MVP has a first POC distance between a first reference frame corresponding to the first MVP and a current frame,
the second MVP has a second POC distance between a second reference frame corresponding to the second MVP and the current frame,
the first POC distance is greater than the second POC distance, and
the refining the at least one MVP in the MV candidate list comprises:
when the MVP pair is non-symmetric, refining the MVP pair to be a symmetric MVP pair, by:
projecting the first MVP to the second reference frame to obtain a projected MVP, and
replacing the second MVP with the projected MVP.
6. The method of claim 5 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric, moving the MVP pair before all non-symmetric MVPs in the MV candidate list.
7. The method of claim 5 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric, moving the MVP pair to a front of the MV candidate list.
8. The method of claim 5 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric with equal picture order count (POC) distance between two reference frames and a current frame, moving the MVP pair to a front of the MV candidate list.
9. A method of processing video data, the method comprising:
performing, by a device comprising a memory storing instructions and a processor in communication with the memory, a conversion between video data and a bitstream by:
determining, by the device based on the video data, a first syntax element indicating whether a compound reference mode is applied for a current video block;
determining, by the device based on the video data, a second syntax element indicating whether a motion vector difference (MVD) is joined signaled for the current video block;
encoding, by the device, the first syntax element in the bitstream;
encoding, by the device, the second syntax element in the bitstream;
when the first syntax element indicates that the compound reference mode is applied for the current video block and the second syntax element indicates that the MVD is joined signaled for the current video block, refining, by the device, at least one MV predictor (MVP) in a MV candidate list based on whether the at least one MVP is symmetric, wherein:
the at least one MVP in the MV candidate list comprises an MVP pair, and
the refining the at least one MVP in the MV candidate list comprises:
refining the MVP pair in the MV candidate list based on whether the MVP pair is symmetric or non-symmetric; and
encoding, by the device based on the MV candidate list, the current video block in the bitstream,
wherein:
the MVP pair comprises a first MVP and a second MVP,
the first MVP has a first POC distance between a first reference frame corresponding to the first MVP and a current frame,
the second MVP has a second POC distance between a second reference frame corresponding to the second MVP and the current frame,
the first POC distance is greater than the second POC distance, and
the refining the at least one MVP in the MV candidate list comprises:
when the MVP pair is non-symmetric, refining the MVP pair to be a symmetric MVP pair, by:
projecting the first MVP to the second reference frame to obtain a projected MVP, and
replacing the second MVP with the projected MVP.
10. The method of claim 9 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric, moving the MVP pair before all non-symmetric MVPs in the MV candidate list.
11. The method of claim 9 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric, moving the MVP pair to a front of the MV candidate list.
12. The method of claim 9 , wherein the refining the MVP pair in the MV candidate list comprises:
when the MVP pair is symmetric with equal picture order count (POC) distance between two reference frames and a current frame, moving the MVP pair to a front of the MV candidate list.