Decision: Potential reduction of context sizes without penalty in compression efficiency – beneficial for pipeline implementations – adopt solution from JCTVC-D260.
Decision: Conclusion on scanning: Disable/remove adaptive scan order from HM – replace by zig-zag scan implementation from JCTVC-D239.
Continue CE on other aspects of adaptive scanning. Clarify whether this is useful for inter (at least RA case seems to give little gain)? Is it justifying the additional complexity? This should include investigations in terms of complexity (beyond computation time), parallelization etc.
Decision: Adopt adaptive scanning for intra JCTVC-D393 after being verified in JCTVC-D453; from the results (JCTVC-D360, JCTVC-D393, JCTVC-D374, JCTVC-D049), as it had not been obvious whether the mode-dependent adaptive scan switching would work well for both HE and LC intra.
Continue CE on other aspects of CABAC (w.r.t. parallelization, complexity; separate encoding of last coefficient, JCTVC-D262, JCTVC-D219 etc.)
