]> git.sesse.net Git - x264/history - encoder/me.c
implement multithreaded me=esa
[x264] / encoder / me.c
2007-07-17 Loren Merrittmove os/compiler specific defines to their own header
2007-07-06 Loren Merrittcosmetics
2007-05-26 Alex Izvorskireplace alloca with malloc everywhere. per manpage...
2007-04-02 Loren Merrittin hpel search, merge two 16x16 mc calls into one 16x17...
2006-12-15 Loren MerrittNew threading method:
2006-09-24 Loren Merrittworkaround: on some systems, alloca() isn't aligned
2006-09-12 Loren Merrittfaster ESA
2006-09-12 Loren Merrittfaster ESA
2006-04-24 Loren MerrittBefore evaluating the RD score of any mode, check satd...
2006-04-24 Loren MerrittWhen evaluating predictors to start fullpel motion...
2006-04-10 Loren Merrittmore interleaved SAD.
2006-04-10 Loren Merrittmore interleaved SAD.
2006-04-10 Loren Merrittinterleave multiple calls to SAD.
2006-03-20 Loren MerrittRD subpel motion estimation (--subme 7)
2006-02-27 Loren Merrittcopy current macroblock to a smaller buffer, to improve...
2006-01-18 Loren Merrittearly termination in UMH search
2006-01-18 Loren Merrittsplit mv_range enforcement from edge-of-frame clipping...
2006-01-17 Loren Merrittcosmetics: simplify #includes
2006-01-09 Loren Merrittearly termination within large SADs. ~1% faster UMH...
2005-12-30 Loren MerrittMSVC compatibility fixes
2005-12-30 Loren Merrittjoint bidirectional motion refinement (--bime)
2005-12-20 Loren MerrittSuccessive elimination motion search: same as exhaustiv...
2005-12-03 Loren Merrittfaster subpel motion search.
2005-11-22 Loren Merrittsubpel search: always check mvp.
2005-11-03 Loren Merrittskip a few duplicate candidates in qpel search.
2005-11-03 Loren Merrittskip a few duplicate candidates in fullpel hex&umh...
2005-11-03 Loren MerrittX264_ABS => abs
2005-10-15 Loren MerrittUse SAD instead of SATD for halfpel motion search.
2005-10-08 Loren Merrittnew option: --mixed-refs
2005-06-17 Loren Merrittlossless mode enabled at qp=0
2005-06-14 Loren Merrittrate-distortion optimized MB types in I- and P-frames...
2005-06-05 Loren MerrittSATD-based decision for 8x8 transform in inter-MBs.
2005-05-24 Loren MerrittUMHex motion seach (but no early termination yet)
2005-05-11 Loren MerrittNew ratecontrol options:
2005-05-11 Loren Merrittuse a predicted mv as starting point for subpel refinement.
2005-04-19 Loren MerrittAllow manual selection of fullpel ME method. New method...
2005-03-15 Loren Merrittmacroblock_analyse: simplify cost comparisons. (cosmetic)
2005-03-14 Loren MerrittChroma ME (P-frames only).
2005-03-02 Loren MerrittFix a buffer overrun with very long MVs.
2005-02-23 Loren MerrittPreliminary adaptive B-frame decision (not yet tuned).
2005-02-14 Loren MerrittPrecalculate lambda*bits for all allowed mvs. 1-2%...
2005-02-14 Loren MerrittSave some memcopies in halfpel ME.
2005-02-13 Loren MerrittCache half-pixel interpolated reference frames, to...
2005-02-10 Loren Merrittcosmetics
2005-02-04 Loren Merrittr114 didn't completely fix the problem, trying again.
2005-02-03 Loren MerrittAnother MV clipping fix.
2005-01-31 Loren MerrittMore accurate clipping rectangle for motion search...
2005-01-18 Loren Merrittfinish subpixel motion refinement for B-frames (up...
2004-12-31 Loren Merrittremove relative include paths, to avoid conflicts with...
2004-12-30 Loren Merrittreplace EPZS diamond with a hexagon search pattern.
2004-12-14 Loren Merrittrename 'core/' to 'common/', which avoids conflicts...
2004-11-22 Loren Merrittmultiple hypothesis mv prediction:
2004-09-22 Laurent Aimar * all: Patches by Loren Merritt:
2004-07-28 Eric Petit BeOS fixes (no stdint.h, no libm)
2004-06-03 Laurent Aimar * all: re-import of the CVS.