Generalized bi directional prediction mode in video processing
A method of video processing is provided. The method includes: determining, for a conversion between a first video block of a video and a bitstream of the video, that a merge mode with motion vector differences is applied into the first video block; and performing the conversion based on the determining, where the merge mode includes motion vector expression to derive motion information of the first video block, where the motion vector expression includes a first parameter representing a motion vector difference and a second parameter indicating a base candidate from a base candidate list, where the first parameter includes a first motion magnitude and a first motion direction, and where the base candidate list is based on a first merge candidate list for a merge mode which is constructed for the first video block during the conversion.
1. A method of video processing, comprising:
determining, for a conversion between a first video block of a video and a bitstream of the video, that a merge mode with motion vector differences (MVDs) is applied into the first video block; and
performing the conversion based on the determining,
wherein the merge mode with motion vector differences comprises a motion vector expression to derive motion information of the first video block, wherein the motion vector expression comprises a first parameter representing a motion vector difference and a second parameter indicating a base candidate from a base candidate list, wherein the first parameter comprises a first motion magnitude and a first motion direction,
wherein the base candidate list is based on a first merge candidate list for a merge mode which is constructed for the first video block during the conversion, wherein a size of the base candidate list is not greater than a size of the first merge candidate list for the merge mode,
wherein the method further comprises:
constructing a second merge candidate list for a second video block of the video;
generating a first motion vector from the second merge candidate list;
determining a motion vector difference offset based on a second motion distance and a second motion direction, wherein the second motion distance is determined based on a motion distance index and a sample precision;
determining, a motion vector difference (MVD) based the motion vector difference offset;
determining, a second motion vector based on the MVD and the first motion vector; and
performing a conversion between the second video block and the bitstream of the video based on the second motion vector,
wherein a first flag in the bitstream indicates whether only integer sample precision is used for indicating the second motion distance.
2. The method of claim 1 , wherein the second parameter comprises an index to the base candidate list.
3. The method of claim 1 , wherein the size of the base candidate list is set to one of 1 , or 2 .
4. The method of claim 1 , wherein a flag is present in the bitstream to indicate whether to enable the merge mode with motion vector differences.
5. The method of claim 4 , wherein the flag is present at a sequence parameter set level.
6. The method of claim 1 , wherein the size of the base candidate list is omitted from the bitstream.
7. The method of claim 1 , wherein the first motion magnitude is selected from a first table comprising at least one motion magnitude, and the first motion direction is selected from a second table comprising at least one motion direction, wherein at least one of the first table and the second table depends on a picture order count (POC) of two reference pictures or a POC of a first picture that includes the first video block, or a quantization parameter (QP) used for coding the first video block, a first slice, or the first picture.
8. The method of claim 1 , wherein the first flag is signaled in a case that a merge mode with motion vector difference enablement flag indicates that the merge mode with motion vector difference is enabled.
9. The method of claim 1 , wherein the motion distance index is binarized by a truncated rice process.
10. The method of claim 1 , wherein an element in a first motion distance table is a product of a factor and a corresponding element with a same index in a second motion distance table, wherein the first motion distance table is used to derive the second motion distance when the first flag indicate that only integer sample precision is used for indicating the second motion distance, and the second motion distance table is used to derive the second motion distance when the first flag indicates that not only integer sample precision is used for indicating the second motion distance.
11. The method of claim 10 , wherein the factor is 4.
12. The method of claim 1 , wherein the first flag is a picture header level.
13. The method of claim 12 , wherein a second flag in the bitstream indicates whether only integer sample precision is enabled for indicating the second motion distance, and wherein the second flag is a sequence parameter set (SPS) level.
14. The method of claim 13 , wherein the first flag is not included in the bitstream in response to the second flag indicating false, and wherein the first flag is included in the bitstream in resresponse to the second flag indicating true.
15. The method of claim 1 , wherein performing the conversion includes decoding the first video block from the bitstream.
16. The method of claim 1 , wherein performing the conversion includes encoding the first video block into the bitstream.
17. 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 first video block of a video and a bitstream of the video, that a merge mode with motion vector differences is applied into the first video block; and
perform the conversion based on the determining,
wherein the merge mode with motion vector differences comprises motion vector expression to derive motion information of the first video block, wherein the motion vector expression comprises a first parameter representing a motion vector difference and a second parameter indicating a base candidate from a base candidate list, wherein the first parameter comprises a first motion magnitude and a first motion direction,
wherein the base candidate list is based on a first merge candidate list for a merge mode which is constructed for the first video block during the conversion, wherein a size of the base candidate list is not greater than a size of the first merge candidate list for the merge mode,
wherein the instructions further cause the processor to:
construct a second merge candidate list for a second video block of the video;
generate a first motion vector from the second merge candidate list;
determine a motion vector difference offset based on a second motion distance and a second motion direction, wherein the second motion distance is determined based on a motion distance index and a sample precision;
determine, a motion vector difference (MVD) based the motion vector difference offset;
determine, a second motion vector based on the MVD and the first motion vector; and
perform a conversion between the second video block and the bitstream of the video based on the second motion vector,
wherein a first flag in the bitstream indicates whether only integer sample precision is used for indicating the second motion distance.
18. A non-transitory computer-readable storage medium storing instructions that cause a processor to:
determine, for a conversion between a first video block of a video and a bitstream of the video, that a merge mode with motion vector differences is applied into the first video block; and
perform the conversion based on the determining,
wherein the merge mode with motion vector differences comprises motion vector expression to derive motion information of the first video block, wherein the motion vector expression comprises a first parameter representing a motion vector difference and a second parameter indicating a base candidate from a base candidate list, wherein the first parameter comprises a first motion magnitude and a first motion direction,
wherein the base candidate list is based on a first merge candidate list for a merge mode which is constructed for the first video block during the conversion, wherein a size of the base candidate list is not greater than a size of the first merge candidate list for the merge mode,
wherein the instructions further cause the processor to:
construct a second merge candidate list for a second video block of the video;
generate a first motion vector from the second merge candidate list;
determine a motion vector difference offset based on a second motion distance and a second motion direction, wherein the second motion distance is determined based on a motion distance index and a sample precision;
determine, a motion vector difference (MVD) based the motion vector difference offset;
determine, a second motion vector based on the MVD and the first motion vector; and
perform a conversion between the second video block and the bitstream of the video based on the second motion vector, wherein a first flag in the bitstream indicates whether only integer sample precision
is used for indicating the second motion distance.
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 first video block of a video and a bitstream of the video, that a merge mode with motion vector differences is applied into the first video block; and
generating the bitstream from the first video block based on the determining,
wherein the merge mode with motion vector differences comprises motion vector expression to derive motion information of the first video block, wherein the motion vector expression comprises a first parameter representing a motion vector difference and a second parameter indicating a base candidate from a base candidate list, wherein the first parameter comprises a first motion magnitude and a first motion direction,
wherein the base candidate list is based on a first merge candidate list for a merge mode which is constructed for the first video block during the conversion, wherein a size of the base candidate list is not greater than a size of the first merge candidate list for the merge mode,
wherein the method further comprises:
constructing a second merge candidate list for a second video block of the video;
generating a first motion vector from the second merge candidate list;
determining a motion vector difference offset based on a second motion distance and a second motion direction, wherein the second motion distance is determined based on a motion distance index and a sample precision;
determining, a motion vector difference (MVD) based the motion vector difference offset;
determining, a second motion vector based on the MVD and the first motion vector; and
performing a conversion between the second video block and the bitstream of the video based on the second motion vector,
wherein a first flag in the bitstream indicates whether only integer sample precision is used for indicating the second motion distance.