Selection of motion vector precision
View Patent ↗Approaches to selection of motion vector (“MV”) precision during video encoding are presented. These approaches can facilitate compression that is effective in terms of rate-distortion performance and/or computational efficiency. For example, a video encoder determines an MV precision for a unit of video from among multiple MV precisions, which include one or more fractional-sample MV precisions and integer-sample MV precision. The video encoder can identify a set of MV values having a fractional-sample MV precision, then select the MV precision for the unit based at least in part on prevalence of MV values (within the set) having a fractional part of zero. Or, the video encoder can perform rate-distortion analysis, where the rate-distortion analysis is biased towards the integer-sample MV precision. Or, the video encoder can collect information about the video and select the MV precision for the unit based at least in part on the collected information.
1. A computer system comprising one or more processing units and memory, wherein the computer system implements a video decoder configured to perform operations comprising:
receiving encoded data for frames of a video sequence as part of a bitstream, each of the frames including one or more regions, the bitstream including:
an indicator in a header that applies for at least one of the frames of the video sequence, the indicator indicating:
whether or not motion vector (“MV”) precision for the at least one of the frames is controlled on a region-by-region basis; and
if the MV precision for the at least one of the frames is not controlled on a region-by-region basis, whether the MV precision for the at least one of the frames is quarter-sample precision or integer-sample precision; and
if the MV precision for the at least one of the frames is controlled on a region-by-region basis, for each region of the one or more regions of the at least one of the frames, a flag in a header for the region, the flag indicating whether MV precision for the region is quarter-sample precision or integer-sample precision; and
decoding the encoded data to reconstruct the frames of the video sequence.
2. The computer system of claim 1 , wherein the header that applies for the at least one of the frames is a sequence parameter set or picture parameter set, wherein the regions are slices, and wherein the header for the region is a slice header.
3. The computer system of claim 1 , wherein, for the indicator:
a first possible value indicates that the MV precision for the at least one of the frames is not controlled on a region-by-region basis and further indicates that the MV precision for the at least one of the frames is quarter-sample precision;
a second possible value indicates that the MV precision for the at least one of the frames is not controlled on a region-by-region basis and further indicates that the MV precision for the at least one of the frames is integer-sample precision; and
a third possible value indicates that the MV precision for the at least one of the frames is controlled on a region-by-region basis.
4. The computer system of claim 1 , wherein, for a given region of the one or more regions of the at least one of the frames, if the flag for the given region is not present in the bitstream, the flag for the given region is inferred to have a value equal to the indicator.
5. The computer system of claim 1 , wherein the indicator is a two-bit value.
6. The computer system of claim 1 , wherein interpretation of MV values changes depending on whether the MV precision is quarter-sample precision or integer-sample precision, the MV values representing integer-sample offsets if the MV precision is integer-sample precision, and the MV values representing quarter-sample offsets if the MV precision is quarter-sample precision.
7. One or more non-volatile memory or storage devices having stored therein encoded data for frames of a video sequence as part of a bitstream, each of the frames including one or more regions, the bitstream including an indicator in a header that applies for at least one of the frames of the video sequence, the indicator indicating (a) whether or not motion vector (“MV”) precision for the at least one of the frames is controlled on a region-by-region basis, and (b) if the MV precision for the at least one of the frames is not controlled on a region-by-region basis, whether the MV precision for the at least one of the frames is quarter-sample precision or integer-sample precision, the bitstream further including, if the MV precision for the at least one of the frames is controlled on a region-by-region basis, for each region of the one or more regions of the at least one of the frames, a flag in a header for the region, the flag indicating whether MV precision for the region is quarter-sample precision or integer-sample precision, wherein the encoded data is configured to facilitate reconstruction of the frames of the video sequence by operations that include:
receiving the encoded data; and
decoding the encoded data to reconstruct the frames of the video sequence.
8. The one or more non-volatile memory or storage devices of claim 7 , wherein the header that applies for the at least one of the frames is a sequence parameter set or picture parameter set, wherein the regions are slices, and wherein the header for the region is a slice header.
9. The one or more non-volatile memory or storage devices of claim 7 , wherein, for the indicator:
a first possible value indicates that the MV precision for the at least one of the frames is not controlled on a region-by-region basis and further indicates that the MV precision for the at least one of the frames is quarter-sample precision;
a second possible value indicates that the MV precision for the at least one of the frames is not controlled on a region-by-region basis and further indicates that the MV precision for the at least one of the frames is integer-sample precision; and
a third possible value indicates that the MV precision for the at least one of the frames is controlled on a region-by-region basis.
10. The one or more non-volatile memory or storage devices of claim 7 , wherein, for a given region of the one or more regions of the at least one of the frames, if the flag for the given region is not present in the bitstream, the flag for the given region is inferred to have a value equal to the indicator during the decoding.
11. The one or more non-volatile memory or storage devices of claim 7 , wherein the indicator is a two-bit value.
12. The one or more non-volatile memory or storage devices of claim 7 , wherein interpretation of MV values changes during the decoding depending on whether the MV precision is quarter-sample precision or integer-sample precision, the MV values representing integer-sample offsets if the MV precision is integer-sample precision, and the MV values representing quarter-sample offsets if the MV precision is quarter-sample precision.
13. One or more non-volatile memory or storage devices having stored therein computer-executable instructions for causing one or more processing units, when programmed thereby, to perform operations comprising:
encoding frames of a video sequence to produce encoded data, each of the frames including one or more regions, wherein the encoding includes:
setting a value of an indicator in a header that applies for at least one of the frames of the video sequence, the indicator indicating:
whether or not motion vector (“MV”) precision for the at least one of the frames is controlled on a region-by-region basis; and
if the MV precision for the at least one of the frames is not controlled on a region-by-region basis, whether the MV precision for the at least one of the frames is quarter-sample precision or integer-sample precision; and
if the MV precision for the at least one of the frames is controlled on a region-by-region basis, for each region of the one or more regions of the at least one of the frames, setting a value of a flag in a header for the region, the flag indicating whether MV precision for the region is quarter-sample precision or integer-sample precision; and
outputting the encoded data as part of a bitstream, the bitstream including the indicator and, if the MV precision for the at least one of the frames is controlled on a region-by-region basis, a flag for each region of the one or more regions of the at least one of the frames that indicates the MV precision for the region.
14. The one or more non-volatile memory or storage devices of claim 13 , wherein the header that applies for the at least one of the frames is a sequence parameter set or picture parameter set, wherein the regions are slices, and wherein the header for the region is a slice header.
15. The one or more non-volatile memory or storage devices of claim 13 , wherein, for the indicator:
a first possible value indicates that the MV precision for the at least one of the frames is not controlled on a region-by-region basis and further indicates that the MV precision for the at least one of the frames is quarter-sample precision;
a second possible value indicates that the MV precision for the at least one of the frames is not controlled on a region-by-region basis and further indicates that the MV precision for the at least one of the frames is integer-sample precision; and
a third possible value indicates that the MV precision for the at least one of the frames is controlled on a region-by-region basis.
16. The one or more non-volatile memory or storage devices of claim 13 , wherein, for a given region of the one or more regions of the at least one of the frames, if the flag for the given region is not present in the bitstream, the flag for the given region is inferred to have a value equal to the indicator during the encoding.
17. The one or more non-volatile memory or storage devices of claim 13 , wherein the indicator is a two-bit value.
18. The one or more non-volatile memory or storage devices of claim 13 , wherein interpretation of MV values changes during the encoding depending on whether the MV precision is quarter-sample precision or integer-sample precision, the MV values representing integer-sample offsets if the MV precision is integer-sample precision, and the MV values representing quarter-sample offsets if the MV precision is quarter-sample precision.
19. The one or more non-volatile memory or storage devices of claim 13 , wherein the setting the value of the indicator is based on source of the frames of the video sequence.
20. The one or more non-volatile memory or storage devices of claim 13 , wherein the setting the value of the indicator is based on measurement of a performance heuristic or historical data.