Abstract:

Supporting digital Video Cassette Recording (VCR) trick-play functionalities (e.g. random access, false-forward play, fast-reverse play) is desirable for compressed video streams. However, due to strong Inter-frame dependencies introduced by motion compensated prediction (MCP), the computational complexity and memory requirement is drastically increased. In this paper, we address the problem of digital VCR trick-play functionalities by investigating different prediction schemes. Specifically, we propose prediction schemes named G-Group and Binary Reference GOP Structure (BRGS), for the H.264/AVC video coding standard to achieve the trick-play functionalities while keeping low decoder complexity and memory requirement. We show that different prediction schemes represent different tradeoffs between the prediction distance (which affects the video quality) and the decoder complexity. We also address the problem of how to compare different prediction schemes by formulating it into an unconstrained optimization problem so that the minimal costs of different schemes can be compared. By comparing the solution offered by G-Group and BRGS to the convex-hull of the operation points, we show that G-Group and BRGS are close to the global optimal solutions. The same approach can be used to evaluate other prediction schemes. Our proposed approaches are flexible and general, and can be easily adapted to evaluate and to achieve a good tradeoff between compression performance and complexity saving for the trick play modes.