On Fri, Nov 17, 2017 at 6:47 PM, <[email protected]> wrote:
> # HG changeset patch > # User Bhavna Hariharan <[email protected]> > # Date 1510901808 -19800 > # Fri Nov 17 12:26:48 2017 +0530 > # Node ID be5b0bf429d1dddf91b9a79368c8d6f434297329 > # Parent 06979c0423504a324ea05ca3de59769c6d0fba0d > analysis: evaluate CUs when inter refinement is enabled and analysis data > from > analysis save mode is unavailable. > > Earlier, smaller CU sizes were evaluated only when the current block is > decided > as "skip" by the save mode. This patch changes the behaviour of inter > refinement > by evaluating smaller CUs irrepective of the mode of the current block. > Pushed to default. Thanks! > > diff -r 06979c042350 -r be5b0bf429d1 source/encoder/analysis.cpp > --- a/source/encoder/analysis.cpp Thu Nov 16 20:23:14 2017 +0530 > +++ b/source/encoder/analysis.cpp Fri Nov 17 12:26:48 2017 +0530 > @@ -2272,8 +2272,7 @@ > bool mightNotSplit = !(cuGeom.flags & CUGeom::SPLIT_MANDATORY); > bool bDecidedDepth = parentCTU.m_cuDepth[cuGeom.absPartIdx] == depth; > > - int split = (m_param->interRefine && cuGeom.log2CUSize == > (uint32_t)(g_log2Size[m_param->minCUSize] + 1) > - && bDecidedDepth && parentCTU.m_predMode[cuGeom.absPartIdx] > == MODE_SKIP); > + int split = (m_param->interRefine && cuGeom.log2CUSize == > (uint32_t)(g_log2Size[m_param->minCUSize] + 1) && bDecidedDepth); > > if (bDecidedDepth) > { > _______________________________________________ > x265-devel mailing list > [email protected] > https://mailman.videolan.org/listinfo/x265-devel >
_______________________________________________ x265-devel mailing list [email protected] https://mailman.videolan.org/listinfo/x265-devel
