--- branches/dev-api-4/xvidcore/ChangeLog 2003/11/29 17:26:12 1226 +++ branches/dev-api-4/xvidcore/ChangeLog 2003/11/29 17:59:28 1227 @@ -6,6 +6,286 @@ # +2003-11-29 17:21:08 GMT patch-111 + + Summary: + First beta marking + Revision: + xvidcore--devapi4--1.0--patch-111 + + First beta marking + + + modified files: + build/generic/configure.in src/xvid.c src/xvid.h + + +2003-11-29 16:59:14 GMT patch-110 + + Summary: + Catching up with CVS. + Revision: + xvidcore--devapi4--1.0--patch-110 + + Changes from sysKin: + * dquant optimization. + * CBR fix + + modified files: + src/encoder.c src/plugins/plugin_single.c + + +2003-11-24 22:05:38 GMT patch-109 + + Summary: + Big level handling in trellis. + Revision: + xvidcore--devapi4--1.0--patch-109 + + Trellis was treating big levels exactly the same way as lower ones. + In some cases, trellis was doing wild optimizations favoring a 0 + because the distortion introduced by that big coeff change was acceptable. + But visually this could result in some nasty blocks with wrong chroma + information or similar brutal changes in other planes as well. + + Skal added big levels handling where trellis just tries to minimize + the cost varying the run value only. No level modification is done + anymore. + + modified files: + TODO src/utils/mbtransquant.c + + +2003-11-23 16:42:55 GMT patch-108 + + Summary: + Trellis for MPEG. + Revision: + xvidcore--devapi4--1.0--patch-108 + + * Added trellis support for MPEg quantization type. + * Changed RD fixed point precision, should help avoiding overflow + (see the constant TL_SHIFT) + + NB: we still have some problems when trellis optimizes DC for big DC + values. + + modified files: + src/utils/mbtransquant.c + + +2003-11-22 00:53:59 GMT patch-107 + + Summary: + Win32 lib project fix (bis) + Revision: + xvidcore--devapi4--1.0--patch-107 + + * nasm >= 0.98.37 support in project file got reverted in a previous + patch, push it back. Nota bene /O3 changed to /O2 for proper + compilation with msvc compiler (everyone is not supposed to compile + stuff with icc) + * TODO update. + + modified files: + TODO build/win32/libxvidcore.dsp + + +2003-11-19 21:26:34 GMT patch-106 + + Summary: + updated bench crc + Revision: + xvidcore--devapi4--1.0--patch-106 + + updated bench crc + + + modified files: + examples/xvid_bench.c + + +2003-11-19 16:00:00 GMT patch-105 + + Summary: + Lumimasking fixes. + Revision: + xvidcore--devapi4--1.0--patch-105 + + from sysKin: + * New plugin hook entry XVID_PLG_FRAME that happens inside FrameCodeIPB + when both type and quant are known. Added hook handling in all plugins. + * Fixed lumimasking. + + from me: + * small reverse commit in pvop estimation fixed. + + modified files: + src/bitstream/bitstream.c src/bitstream/bitstream.h + src/encoder.c src/motion/estimation_pvop.c + src/plugins/plugin_2pass1.c src/plugins/plugin_2pass2.c + src/plugins/plugin_dump.c src/plugins/plugin_lumimasking.c + src/plugins/plugin_psnr.c src/plugins/plugin_single.c + src/xvid.h + + +2003-11-19 15:37:16 GMT patch-104 + + Summary: + Removed indirections from SearchData structure. + Revision: + xvidcore--devapi4--1.0--patch-104 + + Patch from sysKin: + * removed indirections in SearchData structure. CheckCandidate functions + don't use a const SearchData pointer anymore, but they should be a bit + faster because of the less numerous indirections. + + modified files: + src/motion/estimation.h src/motion/estimation_bvop.c + src/motion/estimation_common.c src/motion/estimation_gmc.c + src/motion/estimation_pvop.c src/motion/estimation_rd_based.c + src/motion/vop_type_decision.c + + +2003-11-19 15:33:55 GMT patch-103 + + Summary: + Formula error in twopass code. + Revision: + xvidcore--devapi4--1.0--patch-103 + + The reversing bframe formula in 2pass 2 was not right. This was in fact + a test code i used when i was maintaining the code on its own branch. + I should not have commited it :\ + + Fixed :-) + + modified files: + src/plugins/plugin_2pass2.c + + +2003-11-18 21:41:08 GMT patch-102 + + Summary: + Another problem with mis/unitialized reads. + Revision: + xvidcore--devapi4--1.0--patch-102 + + Michael introduced a fast subpel refine that uses a (iMinSAD2, currentQMV2) + couple of data. The problem is that he plugged this in CheckCandidate16_qpel + that is used outside this context, thus a if statement was traversed with + garbage data in the standard subpel case. For perfection sake, using a + iMinSAD=256*4096 value collects correct data even if it will not be used in + the normal subpel case. + + modified files: + src/motion/estimation_pvop.c + + +2003-11-16 17:29:39 GMT patch-101 + + Summary: + The MEanalysis patch assumed bvops were always used + Revision: + xvidcore--devapi4--1.0--patch-101 + + The MEanalysis patch assumed bvops were always used + + + modified files: + src/encoder.c + + +2003-11-16 15:12:15 GMT patch-100 + + Summary: + MEanalysis using wrong mvs + bframe search using wrong mvs + Revision: + xvidcore--devapi4--1.0--patch-100 + + Valgrind reported lot of unitialized reads. These unitialized reads + helped sysKin finding three bugs: + - ZeroMacroblock did not reset the cbp field. So for some skipped + blocks, a test was done on the cbp value... + - MEanalysis was using wrong mvs from the current bvop (unitialized + or just wrong in current context). That's because in devapi3, bframes + used to share the same mvs array wheras now, it's one array a bvop. + - Collocated skipped MBs for a bvop didn't reset mvs[0] and b_mvs[0]. + + modified files: + src/encoder.c src/motion/estimation_bvop.c src/motion/motion.h + src/motion/motion_inlines.h src/motion/vop_type_decision.c + + +2003-11-15 15:21:09 GMT patch-99 + + Summary: + Small fixes + Revision: + xvidcore--devapi4--1.0--patch-99 + + Small fixes + + + modified files: + src/encoder.c src/motion/vop_type_decision.c + + +2003-11-15 15:02:47 GMT patch-98 + + Summary: + DShow update + libxvidcore project file update + Revision: + xvidcore--devapi4--1.0--patch-98 + + From peter: + * DShow now links against libxvidcore.lib + * Some updates to the libxvidcore project file + + modified files: + TODO build/win32/libxvidcore.dsp dshow/dshow.dsp + dshow/src/CXvidDecoder.cpp dshow/src/CXvidDecoder.h + + +2003-11-15 01:51:28 GMT patch-97 + + Summary: + Small updates to doc files + Revision: + xvidcore--devapi4--1.0--patch-97 + + Small updates to doc files + + + modified files: + AUTHORS CodingStyle TODO doc/INSTALL doc/README + + renamed files: + .arch-ids/authors.txt.id + ==> .arch-ids/AUTHORS.id + .arch-ids/todo.txt.id + ==> .arch-ids/TODO.id + authors.txt + ==> AUTHORS + todo.txt + ==> TODO + + +2003-11-14 11:23:55 GMT patch-96 + + Summary: + Updated ChangeLog + Revision: + xvidcore--devapi4--1.0--patch-96 + + Updated ChangeLog + + + modified files: + ChangeLog + + 2003-11-13 23:09:34 GMT patch-95 Summary: