| 摘要 |
A video codec is provided in which virtual reference data (e.g., virtual reference blocks) are advantageously employed by a video coder and decoder to significantly improve coding efficiency as compared to prior art codecs. Specifically, a“virtual reference,”as used herein, is defined as a group of pixels (e.g., a block) that is used as reference material for encoding portions of the video signal (e.g., a motion-compensated inter-predicted block), but that does not comprise or represent any portion of the actual video sequence to be displayed. For example, it may advantageously be determined that portions of the video sequence may be efficiently represented as a motion-compensated block of pixel values which are generated by the video coder based on the blocks being predicted but not based on any actual blocks in any of the actual video frames or slices.
|
