Spatiotemporal prediction for bidirectionally predictive (B) pictures and motion vector prediction for multi-picture reference motion compensation

摘要

Several improvements for use with Bidirectionally Predictive (B) pictures within a video sequence are provided. In certain improvements Direct Mode encoding and/or Motion Vector Prediction are enhanced using spatial prediction techniques. In other improvements Motion Vector prediction includes temporal distance and subblock information, for example, for more accurate prediction. Such improvements and other presented herein significantly improve the performance of any applicable video coding system/logic.

主权项

1. A computing device comprising a processor and memory, wherein the computing device implements a video encoder configured to perform operations comprising:
selectively enabling temporal motion vector (“MV”) prediction for deriving MV information for blocks of a current frame in a sequence of video frames; encoding the blocks of the current frame, including, for a given block of the blocks of the current frame, encoding the given block of the current frame using motion compensation with derived MV information relative to a reference frame to determine a motion-compensated prediction for the given block, wherein at least one of the blocks of the current frame is encoded using motion compensation with derived MV information derived using spatial MV prediction, and wherein:
if temporal MV prediction is used for the given block of the current frame, the derived MV information for the given block of the current frame is derived by performing temporal MV prediction based upon MV information of a block of another frame in the sequence of video frames; andif spatial MV prediction is used for the given block of the current frame, (1) the derived MV information for the given block of the current frame is derived by performing spatial MV prediction based upon MV information for one or more surrounding blocks of the current frame, and (2) the reference frame for the given block of the current frame is selected from among reference frames used for the one or more surrounding blocks of the current frame; and outputting, in the encoded data, information that indicates whether temporal MV prediction is enabled for the deriving the MV information for the blocks of the current frame.