Motion vector range based on motion vector precision
A method for video processing is provided to comprise: determining a motion vector precision of one or more motion vectors related to a current video block of a video; selecting, based on the motion vector precision, a motion vector range; and performing, based on the motion vector range, a conversion between the current video block and a bitstream representation of the video.
1 . A method for processing video data, comprising:
deriving, for a conversion between a current video block of a video and a bitstream of the video, at least one temporal collocated motion vector, wherein each of the at least one temporal collocated motion vector is derived at least based on a motion vector associated with a first block which has a temporal reference picture that is different from a target reference picture of the current video block;
scaling each of the at least one temporal collocated motion vector based on a first picture order count difference between the target reference picture and a current picture including the current video block and a second picture order count difference between the temporal reference picture and a first picture including the first block;
performing a first clipping operation on the at least one scaled temporal collocated motion vector to derive at least one clipped temporal collocated motion vector, wherein a clipping range of each component of the at least one clipped temporal collocated motion vector is [−(1<<(M−1)), (1<<(M−1))−1], where M is greater than 16 and << is an arithmetic left shift operation; and
performing the conversion based on the at least one clipped temporal collocated motion vector;
wherein the at least one clipped temporal collocated motion vector includes a first temporal collocated motion vector which is used to construct a subblock motion candidate list, and wherein the first temporal collocated motion vector is derived based on a temporal motion shift from a spatial neighboring block of the current video block, and a reference picture related to the temporal motion shift is a same as a collocated picture of the current video block.
2 . The method of claim 1 , wherein the at least one clipped temporal collocated motion vector is further used to construct a motion candidate list, and at least one motion predictor derived based on the motion candidate list is used in the conversion directly, or is combined with at least one motion vector difference to derive a refined motion vector.
3 . The method of claim 2 , wherein each component of the refined motion vector has a bit-depth M-bit.
4 . The method of claim 2 , wherein when the at least one motion vector difference comprising two motion vector differences and a third picture order count difference between a first reference picture and the current picture is less than a forth picture order count difference between a second reference picture and the current picture, the conversion uses a scaling process to obtain motion vector difference of the first reference picture from the motion vector difference of the second reference picture which is indicated by a filed in the bitstream, and wherein the scaling process uses a second clipping operation to constrain each component of the motion vector difference of the first reference picture to fall in a range [—((1<<(M−1))), (1<<(M−1))−1].
5 . The method of claim 1 , wherein M is 18 and the range is [−131072, 131071].
6 . The method of claim 1 , wherein the conversion is performed based on the at least one clipped temporal collocated motion vector for a spatial-temporal motion vector prediction mode of the current video block.
7 . The method of claim 1 , wherein the conversion comprises decoding the current video block from the bitstream.
8 . The method of claim 1 , wherein the conversion comprises encoding the current video block into the bitstream.
9 . 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:
derive, for a conversion between a current video block of a video and a bitstream of the video, at least one temporal collocated motion vector, wherein each of the at least one temporal collocated motion vector is derived at least based on a motion vector associated with a first block which has a temporal reference picture that is different from a target reference picture of the current video block;
scale each of the at least one temporal collocated motion vector based on a first picture order count difference between the target reference picture and a current picture including the current video block and a second picture order count difference between the temporal reference picture and a first picture including the first block;
perform a first clipping operation on the at least one scaled temporal collocated motion vector to derive at least one clipped temporal collocated motion vector, wherein a clipping range of each component of the at least one clipped temporal collocated motion vector is [−(1<<(M−1)), (1<<(M−1))−1], where M is greater than 16 and << is an arithmetic left shift operation; and
perform the conversion based on the at least one clipped temporal collocated motion vector;
wherein the at least one clipped temporal collocated motion vector includes a first temporal collocated motion vector which is used to construct a subblock motion candidate list, and wherein the first temporal collocated motion vector is derived based on a temporal motion shift from a spatial neighboring block of the current video block, and a reference picture related to the temporal motion shift is a same as a collocated picture of the current video block.
10 . The apparatus of claim 9 , wherein the at least one clipped temporal collocated motion vector is further used to construct a motion candidate list, and at least one motion predictor derived based on the motion candidate list is used in the conversion directly, or is combined with at least one motion vector difference to derive a refined motion vector.
11 . The apparatus of claim 10 , wherein each component of the refined motion vector has a bit-depth M-bit.
12 . The apparatus of claim 10 , wherein when the at least one motion vector difference comprising two motion vector differences and a third picture order count difference between a first reference picture and the current picture is less than a forth picture order count difference between a second reference picture and the current picture, the conversion uses a scaling process to obtain motion vector difference of the first reference picture from the motion vector difference of the second reference picture which is indicated by a filed in the bitstream, and wherein the scaling process uses a second clipping operation to constrain each component of the motion vector difference of the first reference picture to fall in a range [−((1<<(M−1))), (1<<(M−1))−1].
13 . The apparatus of claim 9 , wherein M is 18 and the range is [−131072, 131071].
14 . A non-transitory computer-readable storage medium storing instructions that cause a processor to:
derive, for a conversion between a current video block of a video and a bitstream of the video, at least one temporal collocated motion vector, wherein each of the at least one temporal collocated motion vector is derived at least based on a motion vector associated with a first block which has a temporal reference picture that is different from a target reference picture of the current video block;
scale each of the at least one temporal collocated motion vector based on a first picture order count difference between the target reference picture and a current picture including the current video block and a second picture order count difference between the temporal reference picture and a first picture including the first block;
perform a first clipping operation on the at least one scaled temporal collocated motion vector to derive at least one clipped temporal collocated motion vector, wherein a clipping range of each component of the at least one clipped temporal collocated motion vector is [−(1<<(M−1)), (1<<(M−1))−1], where M is greater than 16 and << is an arithmetic left shift operation; and
perform the conversion based on the at least one clipped temporal collocated motion vector;
wherein the at least one clipped temporal collocated motion vector includes a first temporal collocated motion vector which is used to construct a subblock motion candidate list, and wherein the first temporal collocated motion vector is derived based on a temporal motion shift from a spatial neighboring block of the current video block, and a reference picture related to the temporal motion shift is a same as a collocated picture of the current video block.
15 . The non-transitory computer-readable storage medium of claim 14 , wherein the at least one clipped temporal collocated motion vector is further used to construct a motion candidate list, and at least one motion predictor derived based on the motion candidate list is used in the conversion directly, or is combined with at least one motion vector difference to derive a refined motion vector.
16 . The non-transitory computer-readable storage medium of claim 15 , wherein when the at least one motion vector difference comprising two motion vector differences and a third picture order count difference between a first reference picture and the current picture is less than a forth picture order count difference between a second reference picture and the current picture, the conversion uses a scaling process to obtain motion vector difference of the first reference picture from the motion vector difference of the second reference picture which is indicated by a filed in the bitstream, and wherein the scaling process uses a second clipping operation to constrain each component of the motion vector difference of the first reference picture to fall in a range [−((1<<(M−1))), (1<<(M−1))−1].
17 . A method for storing a bitstream of a video, comprising:
deriving at least one temporal collocated motion vector, wherein each of the at least one temporal collocated motion vector is derived at least based on a motion vector associated with a first block which has a temporal reference picture that is different from a target reference picture of a current video block;
scaling each of the at least one temporal collocated motion vector based on a first picture order count difference between the target reference picture and a current picture including the current video block and a second picture order count difference between the temporal reference picture and a first picture including the first block;
performing a first clipping operation on the at least one scaled temporal collocated motion vector to derive at least one clipped temporal collocated motion vector, wherein a clipping range of each component of the at least one clipped temporal collocated motion vector is [−(1<<(M−1)), (1<<(M−1))−1], where M is greater than 16 and << is an arithmetic left shift operation;
generating the bitstream based on the at least one clipped temporal collocated motion vector; and
storing the bitstream in a non-transitory computer-readable recording medium;
wherein the at least one clipped temporal collocated motion vector includes a first temporal collocated motion vector which is used to construct a subblock motion candidate list, and wherein the first temporal collocated motion vector is derived based on a temporal motion shift from a spatial neighboring block of the current video block, and a reference picture related to the temporal motion shift is a same as a collocated picture of the current video block.
18 . The method of claim 17 , wherein the at least one clipped temporal collocated motion vector is further used to construct a motion candidate list, and at least one motion predictor derived based on the motion candidate list is used in the generating directly, or is combined with at least one motion vector difference to derive a refined motion vector.