Home | Back
# Ed.Gomez: This ChangeLog is generated from a personal tree maintained
# under the arch revision control tool. That's why dates may be skewed. I
# also removed all my email adresses from the output because they are not
# relevant.


#########################################################################
# 1.0.2 (Bitstream Version 36)
#########################################################################

2004-08-29 11:24:26 GMT                                  patch-55

    Summary:
      Merged one important forgotten bugfix from head
    Revision:
      xvidcore--stable--1.0--patch-55

    Merged one important forgotten bugfix from head
    
    Patches applied:
    
     * ed.gomez@free.fr--2004-1/xvidcore--head--0.0--patch-70
       Out of bounds MVs clipping
    
     * ed.gomez@free.fr--2004-1/xvidcore--head--0.0--patch-71
       Decoder optimization (fixing regression)
    

    modified files:
     src/decoder.c

    new patches:
     ed.gomez@free.fr--2004-1/xvidcore--head--0.0--patch-70
     ed.gomez@free.fr--2004-1/xvidcore--head--0.0--patch-71


2004-08-29 10:51:58 GMT                                  patch-54

    Summary:
      Marking 1.0.2
    Revision:
      xvidcore--stable--1.0--patch-54

    From ed.gomez:
     * Marking 1.0.2

    modified files:
     ChangeLog build/generic/configure.in src/xvid.c src/xvid.h


2004-08-22 13:08:44 GMT                                  patch-53

    Summary:
      Thread safety problem in idct C version
    Revision:
      xvidcore--stable--1.0--patch-53

    From ed.gomez:
     * Fixed a thread safety problem in C version of the
       idct function. Added some comments on some static
       data not marked as RO.

    modified files:
     src/bitstream/mbcoding.c src/dct/idct.c


2004-08-21 11:45:55 GMT                                  patch-52

    Summary:
      Stupid typo+error in fdct_xxx_skal macro generator.
    Revision:
      xvidcore--stable--1.0--patch-52

    From Nicolas Boulay:
     * Found a typo mistake (ecx->eax) and an error in the same line
       But as we're lucky, the unrolled version was bugfree, and that
       is that one which is used.

    modified files:
     src/dct/x86_asm/fdct_mmx_skal.asm


2004-07-26 20:21:24 GMT                                  patch-51

    Summary:
      ChangeLog Update
    Revision:
      xvidcore--stable--1.0--patch-51

    ChangeLog Update
    

    modified files:
     ChangeLog


2004-07-24 11:33:57 GMT                                  patch-50

    Summary:
      BVOP direct/interpolated ref block rounding fix.
    Revision:
      xvidcore--stable--1.0--patch-50

    From ed.gomez:
     * BVOP direct/interpolated ref block rounding fix.
       It's been using rounding=1 for averaging stage since ever.
       The standard says it's rounding=0.
    
       See standard clause 7.6.9.4 for explicit code and Section 6.3.5
       that says "rounding=0" in bframes as they don't set the
       vop_rounding_type in VOP header. Both sections match, xvid was
       wrong

    modified files:
     src/decoder.c


2004-07-23 20:37:09 GMT                                  patch-49

    Summary:
      Removed data qualifer in .rodata
    Revision:
      xvidcore--stable--1.0--patch-49

    From ed.gomez:
     * long standing warning by yasm... data isn't a keyword for
       .(ro)data sections.

    modified files:
     src/bitstream/x86_asm/cbp_mmx.asm
     src/bitstream/x86_asm/cbp_sse2.asm
     src/dct/x86_asm/fdct_mmx_ffmpeg.asm
     src/dct/x86_asm/fdct_mmx_skal.asm
     src/dct/x86_asm/fdct_sse2_skal.asm
     src/dct/x86_asm/idct_3dne.asm src/dct/x86_asm/idct_mmx.asm
     src/dct/x86_asm/idct_sse2_dmitry.asm
     src/dct/x86_asm/simple_idct_mmx.asm
     src/image/x86_asm/colorspace_rgb_mmx.asm
     src/image/x86_asm/colorspace_yuyv_mmx.asm
     src/image/x86_asm/interpolate8x8_3dn.asm
     src/image/x86_asm/interpolate8x8_3dne.asm
     src/image/x86_asm/interpolate8x8_mmx.asm
     src/image/x86_asm/interpolate8x8_xmm.asm
     src/image/x86_asm/qpel_mmx.asm
     src/image/x86_asm/reduced_mmx.asm
     src/motion/x86_asm/sad_3dn.asm src/motion/x86_asm/sad_3dne.asm
     src/motion/x86_asm/sad_mmx.asm src/motion/x86_asm/sad_sse2.asm
     src/motion/x86_asm/sad_xmm.asm
     src/quant/x86_asm/quantize_h263_3dne.asm
     src/quant/x86_asm/quantize_h263_mmx.asm
     src/quant/x86_asm/quantize_mpeg_mmx.asm
     src/quant/x86_asm/quantize_mpeg_xmm.asm
     src/utils/x86_asm/cpuid.asm
     src/utils/x86_asm/interlacing_mmx.asm
     src/utils/x86_asm/mem_transfer_3dne.asm
     src/utils/x86_asm/mem_transfer_mmx.asm


2004-07-19 18:45:14 GMT                                  patch-48

    Summary:
      Complete previous xvid_decraw patch
    Revision:
      xvidcore--stable--1.0--patch-48

    Complete previous xvid_decraw patch
    

    modified files:
     examples/xvid_decraw.c


2004-07-18 11:58:48 GMT                                  patch-47

    Summary:
      ISO C99'ism fix
    Revision:
      xvidcore--stable--1.0--patch-47

    ISO C99'ism fix
    

    modified files:
     src/encoder.c


2004-07-17 11:32:42 GMT                                  patch-46

    Summary:
      Make sure time incr is never larger than 16bit.
    Revision:
      xvidcore--stable--1.0--patch-46

    From ed.gomez:
     * Keep both fbase and fincr under 16bit limit.

    modified files:
     src/encoder.c


2004-07-17 10:00:42 GMT                                  patch-45

    Summary:
      Future version interoperability
    Revision:
      xvidcore--stable--1.0--patch-45

    From ed.gomez:
     * Zeroing the structures is the best way not to pass
       wrong data when dealing with slightly new XviD (like
       head). If we don't zero here, then brightness is not
       initialized in HEAD leading to crash.

    modified files:
     examples/xvid_decraw.c


2004-07-10 17:30:40 GMT                                  patch-44

    Summary:
      ChangeLog update
    Revision:
      xvidcore--stable--1.0--patch-44

    ChangeLog update
    

    modified files:
     ChangeLog


2004-07-10 17:27:06 GMT                                  patch-43

    Summary:
      Small mem leak in vfw.
    Revision:
      xvidcore--stable--1.0--patch-43

    From sysKin:
     * Free zones.

    modified files:
     vfw/src/codec.c


2004-07-10 16:55:53 GMT                                  patch-42

    Summary:
      Fix wrong matrix reading logic.
    Revision:
      xvidcore--stable--1.0--patch-42

    From sysKin:
     * Fix a problem reading quantization matrix.

    modified files:
     src/bitstream/bitstream.c


2004-06-26 21:26:35 GMT                                  patch-41

    Summary:
      low delay guessing (il)logic fix.
    Revision:
      xvidcore--stable--1.0--patch-41

    From sysKin:
     * bframes were decoded wrong when trying to guess low_delay flag
       as specified in the standard when vol_control_parameters aren't
       available.

    modified files:
     src/decoder.c


2004-06-13 19:15:05 GMT                                  patch-40

    Summary:
      Small memory error in ia32 cpuid function.
    Revision:
      xvidcore--stable--1.0--patch-40

    From ed.gomez:
     * Valgrind detected a write to suspicious stack space. To avoid
       any false reporting, added an explicit stack space allocation.

    modified files:
     src/utils/x86_asm/cpuid.asm

#########################################################################
# 1.0.1 (Bitstream Version 35)
#########################################################################

2004-06-05 22:55:56 GMT                                  patch-39

    Summary:
      Marking 1.0.1 release
    Revision:
      xvidcore--stable--1.0--patch-39

    Marking 1.0.1 release
    

    modified files:
     ChangeLog TODO build/generic/configure.in src/xvid.c
     src/xvid.h


2004-06-02 20:58:38 GMT                                  patch-38

    Summary:
      DC clipping bug for real
    Revision:
      xvidcore--stable--1.0--patch-38

    From ed.gomez:
     * patch-25  was  supposed to fix a DC clipping  bug. However  i
       added the additional  clipping  code in the  wrong place. But
       at least, my fix  didn't cause any trouble, it was just noop.
       This patch should really fix this very "unlikely bug" (i  just
       want to remind the reader that this bug isn't easy to  trigger,
       and eg: my test sequences don't trigger it at all)
    
    BS version incremented: 35

    modified files:
     src/decoder.c src/motion/estimation_rd_based.c
     src/prediction/mbprediction.c src/prediction/mbprediction.h
     src/xvid.h


2004-05-31 21:11:49 GMT                                  patch-37

    Summary:
      time fixes to decoder.
    Revision:
      xvidcore--stable--1.0--patch-37

    From ed.gomez:
     * timestamps were badly computed by teh decoder in some corner cases
       (1fps). This bug revealed that, timestamps were indeed wrong as
       expected, but that bvop blocks in direct mode (vectors interpolated)
       were somewhat compensated with wrong vectors in these same corner
       cases.

    modified files:
     src/bitstream/bitstream.c src/decoder.c src/decoder.h


2004-05-30 09:36:13 GMT                                  patch-36

    Summary:
      Wrong license header.
    Revision:
      xvidcore--stable--1.0--patch-36

    From ed.gomez:
     * Pascal did agree a plain GPL migration long ago, but this file
       remained GPL+location restriction.

    modified files:
     src/image/reduced.c


2004-05-29 09:02:25 GMT                                  patch-35

    Summary:
      More missing va_end() calls.
    Revision:
      xvidcore--stable--1.0--patch-35

    From pete:
     * portab.h is plenty of missing calls to va_end().

    modified files:
     src/portab.h


2004-05-28 21:28:21 GMT                                  patch-34

    Summary:
      FPS=1 problem in decoder.
    Revision:
      xvidcore--stable--1.0--patch-34

    From ed.gomez:
     * patch-24 did fix bad behavior in encoder, so at least, compliant
       streams were generated but the decoder was still doing the maths
       a wrong way. Apply same logic to decoder.
    
    Thanks to the patch-24 bug reporter for this followup.

    modified files:
     src/bitstream/bitstream.c


2004-05-27 20:04:01 GMT                                  patch-33

    Summary:
      Nasty typo in pvop vector lambdas.
    Revision:
      xvidcore--stable--1.0--patch-33

    From sysKin:
     * s/+/*/ in the lambda value array for vectors in the pvop
       estimation module.

    modified files:
     src/motion/estimation_pvop.c


2004-05-26 13:23:38 GMT                                  patch-32

    Summary:
      Bits/Bytes confusion in the VFW frontend.
    Revision:
      xvidcore--stable--1.0--patch-32

    From sysKin:
     * confusion between the kilo, in kilobits (1000) and
       the kilo in kilobytes (1024, should be named KiB anyway)
     * biSizeImage is in bytes, not bits according to the Win32
       API.

    modified files:
     vfw/src/codec.c vfw/src/config.c


2004-05-26 09:28:31 GMT                                  patch-31

    Summary:
      Close variable argument list.
    Revision:
      xvidcore--stable--1.0--patch-31

    From ed.gomez:
     * Close the variable argument list as specified by the ANSI C
       standard. Reported by Carsten <list4darav at gmx dot de> on
       xvid-devel.

    modified files:
     src/image/font.c


2004-05-26 09:00:26 GMT                                  patch-30

    Summary:
      ICM compatibility for VFW
    Revision:
      xvidcore--stable--1.0--patch-30

    From sysKin:
     * Makes the VFW frontend compatible with ICM applications
       (Ooo, MS Office... etc). Reported on IRC.

    modified files:
     vfw/src/config.c vfw/src/driverproc.c


2004-05-26 08:58:56 GMT                                  patch-29

    Summary:
      Small trellis bug
    Revision:
      xvidcore--stable--1.0--patch-29

    From sysKin:
     * Last coeff wasn't summed.
    
    Reported by Jean Marc.

    modified files:
     src/utils/mbtransquant.c


2004-05-26 08:46:45 GMT                                  patch-28

    Summary:
      Small bug in bframe ME.
    Revision:
      xvidcore--stable--1.0--patch-28

    From sysKin:
     * Small bug in bframe ME.

    modified files:
     src/motion/estimation_bvop.c

#########################################################################
# 1.0.0 final (Bitstream Version 34)
#########################################################################

2004-05-08 22:26:06 GMT                                  patch-27

    Summary:
      Marking 1.0.0 final
    Revision:
      xvidcore--stable--1.0--patch-27

    From ed.gomez:
     * Marking 1.0.0 final \o/

    modified files:
     ChangeLog build/generic/configure.in src/xvid.h


2004-05-06 17:56:52 GMT                                  patch-26

    Summary:
      Small mismatch in hint<->widget in VFW
    Revision:
      xvidcore--stable--1.0--patch-26

    From sysKin:
     * Small mismatch in hint<->widget.

    modified files:
     vfw/src/resource.rc


2004-05-02 22:40:50 GMT                                  patch-25

    Summary:
      DC prediction fix.
    Revision:
      xvidcore--stable--1.0--patch-25

    From ed.gomez:
     * DC predictors weren't clipped to the [-2048, 2047] range.
    
     BS version increased to 33
     
     Thanks to jnorish on our forums to point out the problem.

    modified files:
     src/bitstream/bitstream.c src/decoder.c
     src/motion/estimation_rd_based.c src/prediction/mbprediction.c
     src/prediction/mbprediction.h src/xvid.h


2004-05-02 10:30:29 GMT                                  patch-24

    Summary:
      Possible VOL header corruption.
    Revision:
      xvidcore--stable--1.0--patch-24

    From ed.gomez:
     * The VOL header could be corrupted when passing fincr=fbase=1 which
       happens for fps=1 sequences.
    
    BS version bumped up to 32
    
    Original report:
    http://www.xvid.org/modules.php?op=modload&name=phpBB2&file=viewtopic&t=2026&highlight=

    modified files:
     src/bitstream/bitstream.c src/xvid.h


2004-04-30 23:10:19 GMT                                  patch-23

    Summary:
      Some very light Unix build system changes
    Revision:
      xvidcore--stable--1.0--patch-23

    To prepare testing framework merging.
    
    From ed.gomez:
     * Some typos
     * Copyright updates (it's 2004 since a few months ;-)
     * Added some checking to bootstrap.sh
     * Added m4 AC_PREREQ macro to configure.in

    modified files:
     build/generic/Makefile build/generic/bootstrap.sh
     build/generic/configure.in


2004-04-20 19:40:29 GMT                                  patch-22

    Summary:
      Small visual fix.
    Revision:
      xvidcore--stable--1.0--patch-22

    From sysKin:
     * Small visual fix

    modified files:
     vfw/src/config.c


2004-04-20 19:38:24 GMT                                  patch-21

    Summary:
      Fix crash in decoder for non IFrame 1st frame.
    Revision:
      xvidcore--stable--1.0--patch-21

    From sysKin:
     * Fixed the crash caused by non IFrame 1st frame.

    modified files:
     src/decoder.c


2004-04-18 16:21:50 GMT                                  patch-20

    Summary:
      Typo
    Revision:
      xvidcore--stable--1.0--patch-20

    Typo
    

    modified files:
     vfw/src/resource.rc


2004-04-17 17:04:20 GMT                                  patch-19

    Summary:
      vfw opens audio file in shared access mode
    Revision:
      xvidcore--stable--1.0--patch-19

    vfw opens audio file in shared access mode
    

    modified files:
     vfw/src/config.c


2004-04-15 22:39:12 GMT                                  patch-18

    Summary:
      Tiny xvid_decraw cleaning
    Revision:
      xvidcore--stable--1.0--patch-18

    Tiny xvid_decraw cleaning
    

    modified files:
     examples/xvid_decraw.c


2004-04-15 19:14:31 GMT                                  patch-17

    Summary:
      Tiny minor fixes for msvc.
    Revision:
      xvidcore--stable--1.0--patch-17

    From pete:
     * Missing arch endianness define in project files.
     * Add a textual warning about win32 console EOF misreading.
     * Prevent a SIGFPE when no frames were decoded.

    modified files:
     build/win32/xvid_decraw.dsp build/win32/xvid_encraw.dsp
     examples/xvid_decraw.c


2004-04-14 22:41:07 GMT                                  patch-16

    Summary:
      Fixed missing 1st frame in dshow output.
    Revision:
      xvidcore--stable--1.0--patch-16

    From sysKin:
     * decoder flags were overwritten, this was preventing from outputing
       the first frame immediatly.

    modified files:
     dshow/src/CXvidDecoder.cpp


2004-04-14 22:39:17 GMT                                  patch-15

    Summary:
      Ressource leaking in dshow.
    Revision:
      xvidcore--stable--1.0--patch-15

    From sysKin:
     * Same kind of ressource leaking as in vfw. Same cure.

    modified files:
     dshow/src/CXvidDecoder.cpp dshow/src/CXvidDecoder.h


2004-04-14 19:45:23 GMT                                  patch-14

    Summary:
      Fixed small bug in trellis code.
    Revision:
      xvidcore--stable--1.0--patch-14

    From ed.gomez (spotted by jean marc):
     * Trellis optimization was computing the sum |coeffs| wrongly
       because the Compute_sum function wasn't doing zigzag scanning
       and stopped at last non zero coeffs in linear scanning...
       The trivial fix could have been to pass the zigzag to the compute
       sum function so it could have done its job right.
       But... Why computing stuff that is already known in the upper layer ?
       I removed the compute sum function, and just pass the original sum
       value in trellis function parameters, just in case we have to return
       it again because trellis failed optimizing the block coeffs.
    
       This fix the bug and saves cycles :-) (but should not be noticeable)

    modified files:
     src/utils/mbtransquant.c src/xvid.h


2004-04-12 12:06:12 GMT                                  patch-13

    Summary:
      Don't do SAD and RD based searches for qp.
    Revision:
      xvidcore--stable--1.0--patch-13

    From sysKin:
     * MakeGoodFlags function wasn't disabling SAD based search when
       using RD. This was causing slowdown for no gain at all. This
       patch should speed up encoding in qp mode.
     

    modified files:
     src/motion/estimation_pvop.c


2004-04-12 12:01:19 GMT                                  patch-12

    Summary:
      3dnow functions proper separation.
    Revision:
      xvidcore--stable--1.0--patch-12

    From pete (thx to a forum report):
     * Separate correctly pure 3dnow functions and 3dnow+mmxext functions.
       This fix "Illegal instruction" crash on old k6-2 CPUs.

    modified files:
     src/xvid.c


2004-04-12 11:57:20 GMT                                  patch-11

    Summary:
      Better MV clipping code.
    Revision:
      xvidcore--stable--1.0--patch-11

    From sysKin:
     * Better MV clipping. DivX 5 generates out of range vectors and clipping
       them directly borks the decoding of other MVs for which out of range
       MVs were predictors. So it's just better to clip them for the block
       decoding and keep an unclipped version for predictions.

    modified files:
     src/decoder.c


2004-04-08 20:34:54 GMT                                  patch-10

    Summary:
      PGM support back in xvid_decraw.
    Revision:
      xvidcore--stable--1.0--patch-10

    From ed.gomez:
     * pgm/pnm format added back to xvid_decraw
       + pgm/pnm formats are now default for yv12/i420/rgb24 pixel format.
       + tga is default for rgb16/32
       + use option -f to choose tga, or pnm/pgm

    modified files:
     examples/xvid_decraw.c


2004-04-07 22:30:15 GMT                                  patch-9

    Summary:
      3DNow Ext functions use MMXEXT opcodes.
    Revision:
      xvidcore--stable--1.0--patch-9

    From Soltius (XviD Forum):
     * Most of 3dnow extension functions do use MMXEXT opcodes, so
       classify these functions as 3dnowext+mmxext. Avoids K6-2
       boxes to crash with an invalid instruction error reported
       by the host OS.
    
    PS: original bug report
    http://www.xvid.org/modules.php?op=modload&name=phpBB2&file=viewtopic&t=1656

    modified files:
     src/xvid.c


2004-04-07 22:01:54 GMT                                  patch-8

    Summary:
      RGB 16bit output fix.
    Revision:
      xvidcore--stable--1.0--patch-8

    From ScarletteTout (XviD Forum):
     * Fix RGB 16bit output in C functions.
    
    From ed.gomez:
     * Replaced PGM output by TGA output so it's easy to implement
       RGB 16/24/32 and greyscale bitmaps support in a single format.
       (pgm could have supported RGB 24 and Greyscale only)
     * Added colorspace choice to xvid_decraw
       Use option -c csp, where csp is either rgb16, rgb24, rgb32, yv12 or i420
       Defaults to i420.
    
    PS: original bug report
    http://www.xvid.org/modules.php?op=modload&name=phpBB2&file=viewtopic&t=1960&highlight=

    modified files:
     examples/xvid_decraw.c src/image/colorspace.c


#########################################################################
# 1.0.0 RC4 (Bitstream Version 30)
#########################################################################

2004-04-04 20:21:38 GMT                                  patch-7

    Summary:
      DShow widget hiding.
    Revision:
      xvidcore--stable--1.0--patch-7

    From Michael:
     * No need to keep widget visibles if they won't be in 1.0.0.

    modified files:
     dshow/src/xvid.ax.rc


2004-04-04 20:17:52 GMT                                  patch-6

    Summary:
      Compiler quirk in portab.h
    Revision:
      xvidcore--stable--1.0--patch-6

    From Michael:
     * The VC.NET workaround was causing trouble. Inversed the test.

    modified files:
     src/portab.h


2004-04-04 14:19:10 GMT                                  patch-5

    Summary:
      Marking RC4
    Revision:
      xvidcore--stable--1.0--patch-5

    Marking RC4
    

    modified files:
     ChangeLog build/generic/configure.in src/xvid.h


2004-04-04 14:07:00 GMT                                  patch-4

    Summary:
      Frame dropping disabling for bframes.
    Revision:
      xvidcore--stable--1.0--patch-4

    From sysKin & Pete:
     * Disable frame dropping with bframes enabled. These two options
       do not play fine together.

    modified files:
     src/encoder.c


2004-04-04 14:05:50 GMT                                  patch-3

    Summary:
      Dead code removal.
    Revision:
      xvidcore--stable--1.0--patch-3

    From sysKin:
     * FrameCodeP was always called with contanst parameters. Removed
       these parameters and associated dead code.

    modified files:
     src/encoder.c


2004-04-04 14:03:42 GMT                                  patch-2

    Summary:
      Typo in ME fast comparison.
    Revision:
      xvidcore--stable--1.0--patch-2

    From sysKin:
     * Small typo in Fast ME code.

    modified files:
     src/motion/estimation_common.c


2004-04-02 23:58:19 GMT                                  patch-1

    Summary:
      VFW Resource leak fix (try #2)
    Revision:
      xvidcore--stable--1.0--patch-1

    From Suiryc on IRC:
     * both encoder and decoder ending functions were calling
       the dll freeing code. This was an error as the first function
       called would unbind core function for the second called one.
       Thus xvidcore could not release buffers.

    modified files:
     vfw/src/codec.c vfw/src/driverproc.c


2004-04-02 20:33:02 GMT                                  base-0

    Summary:
      tag of ed.gomez@free.fr--2004-1/xvidcore--devapi4--1.0--patch-53
    Revision:
      xvidcore--stable--1.0--base-0

    (automatically generated log message)


2004-03-31 19:32:47 GMT                                  patch-53

    Summary:
      Ressources leaking in VFW.
    Revision:
      xvidcore--devapi4--1.0--patch-53

    From sysKin:
     * Storing ressources in global vars is making multithreaded/instanced
       apps leaking lot of memory. Moved these vars to codec struct.
    
    Thanks to dalox <fdalleau at hymatom d0t fr> to spot and fix the bug.

    modified files:
     vfw/src/codec.c vfw/src/codec.h vfw/src/driverproc.c


2004-03-31 19:28:51 GMT                                  patch-52

    Summary:
      Fix to bad NVOP+bframe interaction.
    Revision:
      xvidcore--devapi4--1.0--patch-52

    From syskin:
     * When generating a NVOP, it interferes with bframe flushing and
       packing.

    modified files:
     src/encoder.c


2004-03-31 19:24:42 GMT                                  patch-51

    Summary:
      Fix GMC 2 warp point.
    Revision:
      xvidcore--devapi4--1.0--patch-51

    From Skal:
     * Fix 2 warp points GMC.

    modified files:
     src/motion/gmc.c


2004-03-31 19:18:46 GMT                                  patch-50

    Summary:
      Added intra quant testing.
    Revision:
      xvidcore--devapi4--1.0--patch-50

    From Skal:
     * Added intra quant testing to xvid_bench

    modified files:
     examples/xvid_bench.c


2004-03-31 19:07:55 GMT                                  patch-49

    Summary:
      input width/height check
    Revision:
      xvidcore--devapi4--1.0--patch-49

    From Skal:
     * Input width and height aren't checked and this can cause crash.

    modified files:
     src/encoder.c


2004-03-28 01:02:21 GMT                                  patch-48

    Summary:
      $ CVS expansion removed
    Revision:
      xvidcore--devapi4--1.0--patch-48

    $ CVS expansion removed
    

    modified files:
     dshow/src/Configure.cpp dshow/src/config.c
     examples/xvid_encraw.c


2004-03-28 00:45:23 GMT                                  patch-47

    Summary:
      Allow bigger frames
    Revision:
      xvidcore--devapi4--1.0--patch-47

    Allow bigger frames
    

    modified files:
     examples/xvid_encraw.c


2004-03-28 00:41:54 GMT                                  patch-46

    Summary:
      Fix in postproc header
    Revision:
      xvidcore--devapi4--1.0--patch-46

    Fix in postproc header
    

    modified files:
     src/image/postprocessing.h


2004-03-28 00:33:02 GMT                                  patch-45

    Summary:
      Fixed xvidvfw build on real mingw+msys systems
    Revision:
      xvidcore--devapi4--1.0--patch-45

    Fixed xvidvfw build on real mingw+msys systems
    

    modified files:
     vfw/bin/Makefile


2004-03-15 21:48:48 GMT                                  patch-44

    Summary:
      VFW updates.
    Revision:
      xvidcore--devapi4--1.0--patch-44

    VFW updates.

    modified files:
     vfw/src/config.c vfw/src/resource.rc


2004-03-15 21:44:17 GMT                                  patch-43

    Summary:
      GMC bugfix.
    Revision:
      xvidcore--devapi4--1.0--patch-43

    From skal:
     * GMC bugfixes.

    modified files:
     src/motion/gmc.c


2004-03-15 21:41:18 GMT                                  patch-42

    Summary:
      Buffer overrun fix in post proc.
    Revision:
      xvidcore--devapi4--1.0--patch-42

    From sysKin:
     * Fixed buffer overrun in postproc code.

    modified files:
     src/image/postprocessing.c src/image/postprocessing.h


2004-03-15 21:33:22 GMT                                  patch-41

    Summary:
      Forgotten files.
    Revision:
      xvidcore--devapi4--1.0--patch-41

    Me:
     * Sorry i missed these file additions. Important for the packages
       as i build them from the tla archive.

    new files:
     dshow/src/.arch-ids/Configure.cpp.id
     dshow/src/.arch-ids/config.c.id
     dshow/src/.arch-ids/config.h.id dshow/src/.arch-ids/debug.h.id
     dshow/src/Configure.cpp dshow/src/config.c dshow/src/config.h
     dshow/src/debug.h


2004-03-03 21:01:09 GMT                                  patch-40

    Summary:
      MV clipping in decoder.
    Revision:
      xvidcore--devapi4--1.0--patch-40

    From sysKin:
     * clip MVs to valid ranges.

    modified files:
     src/decoder.c src/decoder.h


#########################################################################
# 1.0.0 RC3 (Bistream Version 29)
#########################################################################

2004-02-29 13:17:10 GMT                                  patch-39

    Summary:
      Marking 1.0.0 RC3
    Revision:
      xvidcore--devapi4--1.0--patch-39

    * Marking RC3

    modified files:
     ChangeLog build/generic/configure.in src/xvid.h


2004-02-29 13:08:38 GMT                                  patch-38

    Summary:
      Win32 project fixes for xvid_encraw and xvid_decraw.
    Revision:
      xvidcore--devapi4--1.0--patch-38

    From sysKin:
     * xvid_dec/encraw were linking against libxvidcore.lib. Now they link
       against xvidcore.dll.a
    
    NB: xvid_bench is left as is though it doesn't link. This program
        requires access to internal functions which aren't available
        through the dll link lib. No good solution exists for MSVC to
        build both a dll+its link lib+static lib.

    modified files:
     build/win32/xvid_decraw.dsp build/win32/xvid_encraw.dsp


2004-02-29 12:56:36 GMT                                  patch-37

    Summary:
      DShow updates
    Revision:
      xvidcore--devapi4--1.0--patch-37

    Bunch of DShow updates

    modified files:
     dshow/src/CXvidDecoder.cpp dshow/src/resource.h
     dshow/src/xvid.ax.rc


2004-02-29 12:55:41 GMT                                  patch-36

    Summary:
      VFW updates.
    Revision:
      xvidcore--devapi4--1.0--patch-36

    * Bunch of VFW frontends updates.

    modified files:
     vfw/src/codec.c vfw/src/codec.h vfw/src/config.c
     vfw/src/resource.rc vfw/src/status.c vfw/vfw.dsp


2004-02-29 12:49:24 GMT                                  patch-35

    Summary:
      Artefact workaround in bframes.
    Revision:
      xvidcore--devapi4--1.0--patch-35

    From sysKin:
     * Workaround for some artefacts appearing in bframes.

    modified files:
     src/motion/estimation_bvop.c


2004-02-29 12:46:13 GMT                                  patch-34

    Summary:
      Decoder buffer overflow fix.
    Revision:
      xvidcore--devapi4--1.0--patch-34

    From sysKin:
     * Avoids buffer overflow when reading the last align byte. Our
       bitstream do 4 byte reading because of some platform contraints
       (ARM) and can cause buffer overflow reads.

    modified files:
     src/decoder.c


2004-02-29 11:53:47 GMT                                  patch-33

    Summary:
      Compatibility decoding for old bitstreams.
    Revision:
      xvidcore--devapi4--1.0--patch-33

    From syskin:
     * old core versions used in dev-api-3 distributed by nearly all win32 bin
       builders used to have a edging bug. So when this information is known,
       workaround the bug.

    modified files:
     src/decoder.c src/encoder.c src/image/image.c
     src/image/image.h

#########################################################################
# 1.0.0 RC2 (Bistream Version 28)
#########################################################################

2004-02-08 01:06:40 GMT                                  patch-32

    Summary:
      Marking RC2
    Revision:
      xvidcore--devapi4--1.0--patch-32

    Marking RC2
    

    modified files:
     ChangeLog build/generic/configure.in


2004-02-07 13:54:24 GMT                                  patch-31

    Summary:
      Win32 project outputs dll lib for linking.
    Revision:
      xvidcore--devapi4--1.0--patch-31

    From pete:
     * output a lib to link against the dll (xvidcore.dll.a).

    modified files:
     build/win32/libxvidcore.dsp


2004-02-07 13:51:01 GMT                                  patch-30

    Summary:
      DShow update.
    Revision:
      xvidcore--devapi4--1.0--patch-30

    From pete:
     * cmd line driving
    
    From sysKin(?):
     * bugfixes related to video flipping
     * bugfix for the 'crash at the end" bug

    modified files:
     TODO dshow/dshow.dsp dshow/src/CAbout.cpp dshow/src/CAbout.h
     dshow/src/CXvidDecoder.cpp dshow/src/CXvidDecoder.h
     dshow/src/xvid.ax.def


2004-02-07 13:47:45 GMT                                  patch-29

    Summary:
      VFW updates
    Revision:
      xvidcore--devapi4--1.0--patch-29

    Sorry feeling lazy about splitting this patch...
    
    From peter:
     * bitrate calculator
    
    From sysKin:
     * WMP9 bugfix

    modified files:
     TODO vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/resource.h vfw/src/resource.rc


2004-02-07 13:43:26 GMT                                  patch-28

    Summary:
      Implicit overflow tuning for 2nd pass.
    Revision:
      xvidcore--devapi4--1.0--patch-28

    From sysKin:
     * When doing a bigger 2nd pass, the overflow loop must be more
       aggressive else no bonus bits are reinjected. So we can auto
       tune the overlow values in that case.

    modified files:
     src/plugins/plugin_2pass2.c


2004-02-07 13:38:33 GMT                                  patch-27

    Summary:
      GMC+interlaced bugfix in decoder.
    Revision:
      xvidcore--devapi4--1.0--patch-27

    From sysKin:
     * GMC+interlaced bugfix in decoder.

    modified files:
     src/decoder.c


2004-02-07 13:35:16 GMT                                  patch-26

    Summary:
      Reverted patch-23
    Revision:
      xvidcore--devapi4--1.0--patch-26

    From christoph:
     * reverted patch-23, old code was right.
     * Important typo for the YVYU csp (passing the y plane instead of u).

    modified files:
     src/image/image.c src/xvid.h


2004-01-31 11:20:36 GMT                                  patch-25

    Summary:
      DShow support for more mpeg4 fourccs.
    Revision:
      xvidcore--devapi4--1.0--patch-25

    From sysKin:
     * Added support for the MP4V fourcc.

    modified files:
     dshow/src/CAbout.cpp dshow/src/CAbout.h
     dshow/src/CXvidDecoder.cpp dshow/src/CXvidDecoder.h
     dshow/src/resource.h dshow/src/xvid.ax.rc


2004-01-31 11:12:38 GMT                                  patch-24

    Summary:
      DivX decoder compatibility
    Revision:
      xvidcore--devapi4--1.0--patch-24

    From sysKin:
     * DivX decoder compatibility improved for packed bitstreams.
       It should now detect them and play them fine.

    modified files:
     src/bitstream/bitstream.c


2004-01-31 11:10:26 GMT                                  patch-23

    Summary:
      YV12/I420 confusion fixed.
    Revision:
      xvidcore--devapi4--1.0--patch-23

    From christoph:
     * I420/YV12 were swapped since ... ages.
     * CSP_USER renamed to CSP_PLANAR

    modified files:
     src/encoder.c src/image/image.c src/xvid.h


2004-01-31 10:53:20 GMT                                  patch-22

    Summary:
      Arch separation for mem transfer functions
    Revision:
      xvidcore--devapi4--1.0--patch-22

    Arch separation for mem transfer functions
    

    modified files:
     src/utils/mem_transfer.h


2004-01-27 14:47:08 GMT                                  patch-21

    Summary:
      Write to registry Flip video flag in dshow
    Revision:
      xvidcore--devapi4--1.0--patch-21

    From sysKin:
     * The flip video flag is now saved in registry.
     * Changed internal flags name convention (use n prefix for all now)

    modified files:
     dshow/src/CAbout.cpp dshow/src/CAbout.h
     dshow/src/CXvidDecoder.cpp


2004-01-27 14:42:52 GMT                                  patch-20

    Summary:
      Added bitrate calc to VFW
    Revision:
      xvidcore--devapi4--1.0--patch-20

    From Pete:
     * Added bitrate calculator.
     * Changed up a few function calls to static type.

    modified files:
     vfw/src/codec.c vfw/src/codec.h vfw/src/config.c
     vfw/src/config.h vfw/src/resource.h vfw/src/resource.rc
     vfw/vfw.dsp


2004-01-27 14:29:49 GMT                                  patch-19

    Summary:
      Bugfix in decoder
    Revision:
      xvidcore--devapi4--1.0--patch-19

    From sysKin:
     * when stats are not used, don't write to the stats pointer.

    modified files:
     src/decoder.c

#########################################################################
# 1.0.0 RC1 (Bistream Version 26)
#########################################################################

2004-01-25 16:01:06 GMT                                  patch-18

    Summary:
      Marking RC1
    Revision:
      xvidcore--devapi4--1.0--patch-18

    Marking RC1

    modified files:
     ChangeLog build/generic/configure.in src/xvid.h


2004-01-25 15:37:57 GMT                                  patch-17

    Summary:
      VFW update (again²)
    Revision:
      xvidcore--devapi4--1.0--patch-17

    From sysKin:
     * Added Constant Quant encoding.

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/resource.rc


2004-01-25 15:35:38 GMT                                  patch-16

    Summary:
      Missing MB quants for PP.
    Revision:
      xvidcore--devapi4--1.0--patch-16

    From sysKin:
     * MB Quants are used by the PP code, so don't forget to update
       them even if the block is skipped or not coded.

    modified files:
     src/decoder.c


2004-01-23 13:25:52 GMT                                  patch-15

    Summary:
      VFW update (again)
    Revision:
      xvidcore--devapi4--1.0--patch-15

    From sysKin:
     * Status window updates.
     * Big resource.h cleanup, it seems msvc isn't able to do it automatically.

    modified files:
     vfw/src/resource.h vfw/src/resource.rc vfw/src/status.c


2004-01-23 11:17:20 GMT                                  patch-14

    Summary:
      VFW gcc warnings
    Revision:
      xvidcore--devapi4--1.0--patch-14

    VFW gcc warnings
    

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h


2004-01-22 20:54:53 GMT                                  patch-13

    Summary:
      DShow updates.
    Revision:
      xvidcore--devapi4--1.0--patch-13

    From sysKin:
     * Fixed registry params type. Bool cannot be used or something weird
       happens when writing to registry
     * Defaults set to what the Reset widget sets.

    modified files:
     dshow/src/CAbout.h dshow/src/CXvidDecoder.cpp


2004-01-22 20:35:27 GMT                                  patch-12

    Summary:
      VFW updates.
    Revision:
      xvidcore--devapi4--1.0--patch-12

    From sysKin:
     * GUI improvements.
     * Stats fixing.
     * Automatic config clear upon installation.
     * Added postprocessing options in there too.

    modified files:
     vfw/bin/xvid.inf vfw/src/codec.c vfw/src/codec.h
     vfw/src/config.c vfw/src/config.h vfw/src/resource.h
     vfw/src/resource.rc vfw/src/status.c vfw/src/status.h


2004-01-22 20:28:54 GMT                                  patch-11

    Summary:
      Minor updates to text files
    Revision:
      xvidcore--devapi4--1.0--patch-11

    Minor updates to text files
    

    modified files:
     AUTHORS TODO


2004-01-22 20:27:10 GMT                                  patch-10

    Summary:
      2pass plugin changes.
    Revision:
      xvidcore--devapi4--1.0--patch-10

    From sysKin:
     * Disabled QPel during first pass as well.
     * Fix a mistaking condition when enabling largers 2nd passes.
    
    From ed.gomez:
     * Fix the fix logic. The previous fix enclosed a condition it should
       not have touched. So i removed the mistaking condition, which was
       wrong anyway as stated in the comment, and got back the sane condition
       test.

    modified files:
     src/plugins/plugin_2pass1.c src/plugins/plugin_2pass2.c


2004-01-17 13:03:11 GMT                                  patch-9

    Summary:
      Unitialized pointers during plugin creation.
    Revision:
      xvidcore--devapi4--1.0--patch-9

    From sysKin:
     - plugins which do not require private data were leaving the param2
       unitialized. Just init it to NULL. This bug wasn't causing any
       trouble anyway...

    modified files:
     src/plugins/plugin_dump.c src/plugins/plugin_psnr.c
     vfw/src/codec.c


2004-01-17 01:09:01 GMT                                  patch-8

    Summary:
      DShow forwwards AR information.
    Revision:
      xvidcore--devapi4--1.0--patch-8

    From syskin(?):
     - forward AR information to DShow framework.
     - grayed some widgets.

    modified files:
     dshow/src/CXvidDecoder.cpp dshow/src/CXvidDecoder.h
     dshow/src/xvid.ax.rc


2004-01-07 13:50:29 GMT                                  patch-7

    Summary:
      Scaled zones fix.
    Revision:
      xvidcore--devapi4--1.0--patch-7

    From sysKin (ideas from Koepi iirc):
     * Fix scaled zones computing prescaled data and so on instead of
       guessing everything with a global zone weight factor.

    modified files:
     src/plugins/plugin_2pass2.c


2004-01-06 01:06:39 GMT                                  patch-6

    Summary:
      Tab->Spaces in header only
    Revision:
      xvidcore--devapi4--1.0--patch-6

    Tab->Spaces in header only
    

    modified files:
     src/xvid.h


2004-01-04 18:35:35 GMT                                  patch-5

    Summary:
      Typo fixed
    Revision:
      xvidcore--devapi4--1.0--patch-5

    Typo fixed
    

    modified files:
     build/generic/configure.in


2004-01-04 13:40:51 GMT                                  patch-4

    Summary:
      VFW safer code.
    Revision:
      xvidcore--devapi4--1.0--patch-4

    From sysKin:
     * Protects some parts of the code depending on a previous
       xvidcore opening. Avoids resources leaking.

    modified files:
     vfw/src/codec.c vfw/src/driverproc.c


2004-01-04 13:33:28 GMT                                  patch-3

    Summary:
      Fixes VC debug target name
    Revision:
      xvidcore--devapi4--1.0--patch-3

    Fixes VC debug target name
    

    modified files:
     vfw/vfw.dsp


2004-01-02 23:10:56 GMT                                  patch-2

    Summary:
      Win32 linking policy revised.
    Revision:
      xvidcore--devapi4--1.0--patch-2

    Finnaly Win32 linking policy is to separate all XviD components:
     - xvidcore.dll exports XviD API
     - xvidvfw.dll links against xvidcore DLL
     - xviddshow.dll links against xvidcore DLL
    
    From sysKin:
     * Changed DShow linking policy in VS project file.
     * Changed VFW linking policy in VS project file.
     * Added runtime xvidcore.dll loading in DShow and VFW.
     * Installs xvidcore.dll along side with xvidvfw.dll.
    
    From ed.gomez:
     * Changed libxvidcore.dll mingw32/cygwin target name to
       xvidcore.dll in the configure script.
     * Changed xvid.dll VFW target name to xvidvfw.dll in the
       generic Makefile.
    
    PS: unlike CVS, i reverted back to MS build tools in VS project files

    modified files:
     build/generic/configure.in build/win32/libxvidcore.dsp
     dshow/dshow.dsp dshow/src/CXvidDecoder.cpp
     dshow/src/CXvidDecoder.h vfw/bin/Makefile vfw/bin/sources.inc
     vfw/bin/xvid.inf vfw/src/codec.c vfw/src/codec.h
     vfw/src/config.c vfw/src/config.h vfw/vfw.dsp


2004-01-02 22:02:07 GMT                                  patch-1

    Summary:
      Better seeking in dshow
    Revision:
      xvidcore--devapi4--1.0--patch-1

    From Michael:
     * Better DShow seeking.

    modified files:
     dshow/src/CXvidDecoder.cpp


2004-01-02 12:28:39 GMT                                  base-0

    Summary:
      tag of ed.gomez@free.fr--2003-1/xvidcore--devapi4--1.0--patch-162
    Revision:
      xvidcore--devapi4--1.0--base-0

    Archive cycling...

#########################################################################
# 1.0.0 beta3 (Bistream Version 25)
#########################################################################

2003-12-26 22:21:35 GMT                                  patch-162

    Summary:
      Marking 1.0.0 beta3
    Revision:
      xvidcore--devapi4--1.0--patch-162

    Marking beta3

    modified files:
     ChangeLog TODO build/generic/configure.in src/xvid.h


2003-12-25 20:57:52 GMT                                  patch-161

    Summary:
      Thread safe PP.
    Revision:
      xvidcore--devapi4--1.0--patch-161

    From Michael:
     * Thread safe PP, context is now stored in DECODER struct.

    modified files:
     src/decoder.c src/decoder.h src/image/postprocessing.c
     src/image/postprocessing.h


2003-12-25 20:49:36 GMT                                  patch-160

    Summary:
      Added Turbo option to VFW GUI
    Revision:
      xvidcore--devapi4--1.0--patch-160

    From Michael:
     * Added turbo mode that enables all fast ME flags.

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/resource.h vfw/src/resource.rc


2003-12-25 20:46:16 GMT                                  patch-159

    Summary:
      Cartoon mode usage written to stream.
    Revision:
      xvidcore--devapi4--1.0--patch-159

    From Michael:
     * @encoding: write cartoon usage in the version user data.
     * @decoding: detect cartoon flag appended to version string
                  and force FILM PP disabling.

    modified files:
     src/bitstream/bitstream.c src/bitstream/bitstream.h
     src/decoder.c src/decoder.h src/encoder.c


2003-12-21 13:34:03 GMT                                  patch-158

    Summary:
      Removed unused var in VFW GUI
    Revision:
      xvidcore--devapi4--1.0--patch-158

    Removed unused var in VFW GUI
    

    modified files:
     vfw/src/config.c


2003-12-21 13:32:52 GMT                                  patch-157

    Summary:
      Two pass small update
    Revision:
      xvidcore--devapi4--1.0--patch-157

    from syskin:
     * allow second pass to be bigger than 1st one
       (not tested, the quant mapping formula may not be adapted for this
        usage, so take this change as experimental, and prefer doing
        second pass still smaller than 1st one)
     * let ivops benefit from positive overflow.
    
    from me:
     * set frame type in quant zones (was a buglet)

    modified files:
     src/plugins/plugin_2pass2.c


2003-12-20 22:28:07 GMT                                  patch-156

    Summary:
      New VFW defaults
    Revision:
      xvidcore--devapi4--1.0--patch-156

    From michael:
     * new VFW default values

    modified files:
     vfw/src/config.c


2003-12-20 22:12:38 GMT                                  patch-155

    Summary:
      Added ARGB colorspace.
    Revision:
      xvidcore--devapi4--1.0--patch-155

    Christoph Nägeli naegelic(at)ee{dot}ethz<dot>ch asked me to add C
    support for ARGB colorspace. He provided the encoding part, and i
    extended the original patch in order to have full support for ARGB
    both for encoding and decoding (though it's C only, read *slow*).

    modified files:
     src/image/colorspace.c src/image/colorspace.h
     src/image/image.c src/xvid.c src/xvid.h


2003-12-20 21:29:37 GMT                                  patch-154

    Summary:
      Added 2pass1 comment about fast 1st pass.
    Revision:
      xvidcore--devapi4--1.0--patch-154

    Just added a comment on fast 1st pass, so it explains why
    we do it that way and why some things are left aside.

    modified files:
     src/plugins/plugin_2pass1.c


2003-12-20 20:03:51 GMT                                  patch-153

    Summary:
      Win32 VC6 wrong libc linking.
    Revision:
      xvidcore--devapi4--1.0--patch-153

    From sysKin:
     * Changed single thread libc linking to multithreaded version.

    modified files:
     vfw/vfw.dsp


2003-12-20 15:28:53 GMT                                  patch-152

    Summary:
      VOL flags updating -- take #2
    Revision:
      xvidcore--devapi4--1.0--patch-152

    From sysKin:
     * VOL flags updates fix take #2

    modified files:
     src/encoder.c


2003-12-20 15:10:30 GMT                                  patch-151

    Summary:
      Fast ME tunings.
    Revision:
      xvidcore--devapi4--1.0--patch-151

    From michael:
     * fast refinement for 8x8 blocks
     * more reliable behavior for all fast ME decisions

    modified files:
     src/motion/estimation_bvop.c src/motion/estimation_pvop.c
     src/xvid.h


2003-12-20 14:59:58 GMT                                  patch-150

    Summary:
      VFW AR revamping -- take #2
    Revision:
      xvidcore--devapi4--1.0--patch-150

    From sysKin:
     * more AR revamping

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/resource.h vfw/src/resource.rc


2003-12-20 14:57:40 GMT                                  patch-149

    Summary:
      Unbuffured IO for 1st pass stat files
    Revision:
      xvidcore--devapi4--1.0--patch-149

    Unbuffured IO for 1st pass stat files
    

    modified files:
     src/plugins/plugin_2pass1.c


2003-12-18 17:44:07 GMT                                  patch-148

    Summary:
      Forgotten bit of patch-141
    Revision:
      xvidcore--devapi4--1.0--patch-148

    Damn i forgot to merge the 1st pass changes... so lame, i tested the
    original patch but not the merged one.

    modified files:
     src/plugins/plugin_2pass1.c


2003-12-18 14:45:39 GMT                                  patch-147

    Summary:
      More postprocessing.
    Revision:
      xvidcore--devapi4--1.0--patch-147

    From michael:
     * added film noise effect.
     * moved postproc initialization to decoder initialization.
     * added support for this postproc filter into DShow.

    modified files:
     dshow/src/CAbout.cpp dshow/src/CAbout.h
     dshow/src/CXvidDecoder.cpp dshow/src/resource.h
     dshow/src/xvid.ax.rc src/decoder.c src/image/postprocessing.c
     src/image/postprocessing.h src/xvid.c src/xvid.h


2003-12-18 14:38:19 GMT                                  patch-146

    Summary:
      BFrames ME speed up flags.
    Revision:
      xvidcore--devapi4--1.0--patch-146

    From michael:
     * Added 3 ME flags to skip some bvop ME steps and thus speed up
       ME for bvops (at the expense of quality loss).
        - skip delta search
        - fast interpolate mode
        - early stop

    modified files:
     src/motion/estimation.h src/motion/estimation_bvop.c
     src/motion/estimation_common.c src/motion/estimation_pvop.c
     src/xvid.h


2003-12-18 14:11:01 GMT                                  patch-145

    Summary:
      VFW defaults changed
    Revision:
      xvidcore--devapi4--1.0--patch-145

    From michael:
     * disabled DXN profiles. Better not say we are compatible to avoid
       problems with DXN. Mostly because of VBV lacking.
     * Changed overflow defaults to 5/5/5 with new 2pass code.

    modified files:
     vfw/src/config.c


2003-12-17 16:45:59 GMT                                  patch-144

    Summary:
      Forgotten bit for AR support in VFW
    Revision:
      xvidcore--devapi4--1.0--patch-144

    Koepi might have forgotten to send me this change:
     * added resource id to the resource header

    modified files:
     vfw/src/resource.h


2003-12-17 15:11:37 GMT                                  patch-143

    Summary:
      Lower starting quantizer for CBR encoding
    Revision:
      xvidcore--devapi4--1.0--patch-143

    From christoph:
     * Lower starting quantizer for CBR encodings.

    modified files:
     src/plugins/plugin_single.c


2003-12-17 15:04:33 GMT                                  patch-142

    Summary:
      Small xvid_encraw updates
    Revision:
      xvidcore--devapi4--1.0--patch-142

    From christoph:
     * set upper frame size limit to 4096 pixels
     * GME refinement flag set where it belongs to.

    modified files:
     examples/xvid_encraw.c


2003-12-17 15:01:52 GMT                                  patch-141

    Summary:
      Two pass update.
    Revision:
      xvidcore--devapi4--1.0--patch-141

    This patch improves the two pass code, quantizer distribution is
    smoother and results seem to be better.
    
    * Two pass now scales only a specific part of the frame length.
      This required changing the stats file format (added a header+MV
      length field) and the xvid_plg_data_t structure (binary compatible).
    * Overflow improvement and degradation set to 10% instead of 60%
      Assymetric values may help...
    * Some cleanup work done on the encoder part of the API header.
    
    NB: plg data struct will be cleaned up before 1.0 so if you rely on it
        please read the header file to know which part will disappear.

    modified files:
     src/encoder.c src/plugins/plugin_2pass1.c
     src/plugins/plugin_2pass2.c src/xvid.h


2003-12-17 13:53:34 GMT                                  patch-140

    Summary:
      VOL flags fix.
    Revision:
      xvidcore--devapi4--1.0--patch-140

    From sysKin:
     * VOL flags updated as they should

    modified files:
     src/encoder.c


2003-12-17 11:07:15 GMT                                  patch-139

    Summary:
      VFW GUI Update.
    Revision:
      xvidcore--devapi4--1.0--patch-139

    From Koepi:
     * typo in GMC description
     * Added AR widget (experimental)

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/resource.rc


2003-12-14 12:44:36 GMT                                  patch-138

    Summary:
      Fix PP crash
    Revision:
      xvidcore--devapi4--1.0--patch-138

    Fix PP crash
    

    modified files:
     src/decoder.c


2003-12-14 12:43:21 GMT                                  patch-137

    Summary:
      Texture bit counting for bframes
    Revision:
      xvidcore--devapi4--1.0--patch-137

    Texture bit counting for bframes
    

    modified files:
     src/bitstream/mbcoding.c


2003-12-12 23:58:18 GMT                                  patch-136

    Summary:
      SSE2 code enabled.
    Revision:
      xvidcore--devapi4--1.0--patch-136

    * SSE2 code enabled by default (only sane ones, _no_ idct)
    * Aligned data in xvid_bench to avoid crashes with SSE2 code
      because of unaligned read accesses.

    modified files:
     examples/xvid_bench.c src/xvid.c


2003-12-12 22:50:33 GMT                                  patch-135

    Summary:
      DShow update.
    Revision:
      xvidcore--devapi4--1.0--patch-135

    From michael (from nic):
     * Dshow updates (colorspace etc...)
     * Deblocking option.

    modified files:
     dshow/src/CAbout.cpp dshow/src/CAbout.h
     dshow/src/CXvidDecoder.cpp dshow/src/CXvidDecoder.h
     dshow/src/resource.h dshow/src/xvid.ax.rc


2003-12-12 14:18:13 GMT                                  patch-134

    Summary:
      Added missing postproc init
    Revision:
      xvidcore--devapi4--1.0--patch-134

    Added missing postproc init
    

    modified files:
     src/xvid.c


2003-12-12 14:16:29 GMT                                  patch-133

    Summary:
      YUV space clarifications and fix.
    Revision:
      xvidcore--devapi4--1.0--patch-133

    From christoph:
     * Put some comments
     * fixed UV swapping in USER case.

    modified files:
     src/image/image.c


2003-12-12 14:01:52 GMT                                  patch-132

    Summary:
      Default compilation flags change.
    Revision:
      xvidcore--devapi4--1.0--patch-132

    from christoph:
     - -fgcse was causing trouble on Suse9 gcc
     - -01->-02

    modified files:
     build/generic/configure.in


2003-12-12 13:43:35 GMT                                  patch-131

    Summary:
      Pixel Aspect Ratio support improvement.
    Revision:
      xvidcore--devapi4--1.0--patch-131

    * 1:1 VGA is default now (old behavior)
    * When passed EXT PAR type, we now (try to) sanityse the
      par_width/par_height value:
       - make it positive
       - using 0 (typical memset reseting) defaults to 1
       - simplify the PAR (using gcd)
       - then range it in [1..255] (can be lossy)
    * Specify valid range in API header
    
    Bitstream version bumped to 24.
    
    PS: this patch supersedes sysKin's one in CVS.

    modified files:
     src/encoder.c src/xvid.h


2003-12-10 22:57:50 GMT                                  patch-130

    Summary:
      xvid_decraw cmdline changes.
    Revision:
      xvidcore--devapi4--1.0--patch-130

    * -nframes -> -frames (why did i put a 'n' there ?)
    * -save changed its meaning, it now controls per frame Elementary
      Stream saving.
    * -o string is now independent from -save.
    
    So now it's possible to save both a ES file per frame + a ES file for the
    sequence.

    modified files:
     examples/xvid_encraw.c


2003-12-10 15:08:20 GMT                                  patch-129

    Summary:
      Decoder bugfixes.
    Revision:
      xvidcore--devapi4--1.0--patch-129

    From syskin:
     * bvops MBs were going banana from time to time because they were
       referencing wrong future ref MBs.
     * decoder now informs the client app about bvop lag, returning
       XVID_TYPE_NOTHING, up to the client app to display (or not)
       the bvop lag frame (black with error message).
    
    Fixing previous patch a bit:
     * added $Id$ fields
     * Fixed copyright

    modified files:
     dshow/src/CXvidDecoder.cpp src/decoder.c
     src/image/postprocessing.c src/image/postprocessing.h


2003-12-10 14:53:58 GMT                                  patch-128

    Summary:
      Deblocking code.
    Revision:
      xvidcore--devapi4--1.0--patch-128

    Patch from michael:
     * added deblocking code
    
    Merge work:
     * Added postprocessing.[ch] to project files
     * added #include "image/postprocessing.h" directive in decoder.c
     * new lines missing (gcc is so pedantic)
    
    NB: slice rendering + postprocessing is impossible. Slice rendering is
        somewhat abandoned.

    new files:
     src/image/.arch-ids/postprocessing.c.id
     src/image/.arch-ids/postprocessing.h.id
     src/image/postprocessing.c src/image/postprocessing.h

    modified files:
     build/generic/sources.inc build/win32/libxvidcore.dsp
     src/decoder.c src/xvid.h


2003-12-08 18:33:26 GMT                                  patch-127

    Summary:
      Don't read out of bounds
    Revision:
      xvidcore--devapi4--1.0--patch-127

    Don't read out of bounds
    

    modified files:
     src/plugins/plugin_2pass2.c


2003-12-08 18:31:41 GMT                                  patch-126

    Summary:
      Macroblock structure cleanup
    Revision:
      xvidcore--devapi4--1.0--patch-126

    Macroblock structure cleanup
    

    modified files:
     src/global.h


2003-12-07 15:09:41 GMT                                  patch-125

    Summary:
      Small fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-125

    From gruel:
     * xvid.h: Minor color space correction.
    
    From sysKin:
     * codec.c: Zones fix

    modified files:
     src/xvid.h vfw/src/codec.c


2003-12-07 14:57:14 GMT                                  patch-124

    Summary:
      HUGE file handling in twopass.
    Revision:
      xvidcore--devapi4--1.0--patch-124

    Because of a lacking cast, two pass did not handle well some very
    large target size (bitrate mode is not affected).
    
    It should now be safe specifying target sizes up to 2^31kB which
    represents 2TB. Someone using XviD in studios ? ;-)

    modified files:
     src/plugins/plugin_2pass2.c

#########################################################################
# 1.0.0 beta2 (Bitstream Version 23)
#########################################################################

2003-12-05 14:43:53 GMT                                  patch-123

    Summary:
      Marking 1.0.0 Beta2
    Revision:
      xvidcore--devapi4--1.0--patch-123

    Marking 1.0.0 Beta2
    

    modified files:
     ChangeLog TODO build/generic/configure.in


2003-12-05 14:35:22 GMT                                  patch-122

    Summary:
      Cap quants correctly (the best we can at least)
    Revision:
      xvidcore--devapi4--1.0--patch-122

    Cap quants correctly (the best we can at least)
    

    modified files:
     src/plugins/plugin_single.c


2003-12-05 14:33:48 GMT                                  patch-121

    Summary:
      Small glitch
    Revision:
      xvidcore--devapi4--1.0--patch-121

    Small glitch
    

    modified files:
     src/motion/vop_type_decision.c


2003-12-05 14:06:19 GMT                                  patch-120

    Summary:
      KFthresholding changes.
    Revision:
      xvidcore--devapi4--1.0--patch-120

    As user reports proved, the logic behind the min_key_interval was
     1/ misleading because the parameter is kfthreshold indeed and not
        a minimum keyframe interval
     2/ the formula was a bit too aggressive (removing 20% of bitrate
        per frame until distance to next iframe was 1)
    
    I posted a RFC to try to settle a decision on what behavior this
    setting should have. We have still have no clear answer so i prefer
    just fixing the misleading name right now and wait for a common
    position about its behavior later.
    
    Libraries are *binary* compatible, but *source code* compatibility
    is broken (rename rc_2pass2_t->min_key_interval to kfthreshold).
    This is probably the last API change.
    
    NB: fixes a type problem during scaling parameter computing which
        was causing insane pb_iboost_tax_ratio values.

    modified files:
     src/plugins/plugin_2pass2.c src/xvid.h vfw/src/codec.c
     vfw/src/config.c vfw/src/config.h


2003-12-05 00:20:28 GMT                                  patch-119

    Summary:
      ivop decision tuning.
    Revision:
      xvidcore--devapi4--1.0--patch-119

    ivop decision tuning from sysKin.

    modified files:
     src/motion/vop_type_decision.c


2003-12-03 18:55:29 GMT                                  patch-118

    Summary:
      VOSH header always written.
    Revision:
      xvidcore--devapi4--1.0--patch-118

    * profile is set to sane default value in BitstreamWriteVolHeaders
    * VOSH is now always written (note that the ending code is never
      written)
    * doubled variable removed from bvop estimation file.
    
    bitstream version set to 23
    Version 22 was used in CVS by a fix from michael for VOSH, it was just
    lacking the sane default value setting when profile is 0x00 from user.
    0x00 is a reserved profile ID in the spec so it wasn't making much sense
    to write 0x00.

    modified files:
     src/bitstream/bitstream.c src/motion/estimation_bvop.c
     src/xvid.h


2003-12-03 15:29:30 GMT                                  patch-117

    Summary:
      VFW GUI fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-117

    min key was misleading because it was legacy code from dev-api-3.

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/resource.rc


2003-12-03 15:22:25 GMT                                  patch-116

    Summary:
      ME fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-116

    From syskin:
    - small typo in chroma sad reset
    - code tweaking
      + adv diamond search used intead of mainsearch
      + and some other stuff
    - thresholds tuned.

    modified files:
     src/motion/estimation.h src/motion/estimation_bvop.c
     src/motion/estimation_pvop.c


2003-12-01 11:17:20 GMT                                  patch-115

    Summary:
      Small changes and interlacing bugfix.
    Revision:
      xvidcore--devapi4--1.0--patch-115

    * Interlacing bugfix, code got swapped in a very old patch
      (back in March)
    * xvid_encraw forces VOSH writing.
    * VFW GUI fixes.

    modified files:
     examples/Makefile examples/xvid_encraw.c
     src/utils/mbtransquant.c vfw/src/config.c vfw/src/resource.rc


2003-11-30 15:47:41 GMT                                  patch-114

    Summary:
      Thread safe MPEG4 quantization functions + xvid_bench update
    Revision:
      xvidcore--devapi4--1.0--patch-114

    * Thread safe MPEG4 quantization functions.
      Cleaned up version of patch provided by Michael
       - fixed compiling problems on gcc
       - added const qualifiers every where it was possible to help C compiler
         optimization.
       - added the mpeg_quant_matrices param to all ASM function prototype
         in comments (even if it's not used, that shows we do it deliberatly)
       - forces m[intra][0][0] = 8, otherwise XviD could write invalid streams.
    * Added real CRC computing in xvid_bench.c

    modified files:
     TODO examples/xvid_bench.c src/bitstream/bitstream.c
     src/decoder.c src/decoder.h src/encoder.c src/encoder.h
     src/motion/estimation.h src/motion/estimation_pvop.c
     src/motion/estimation_rd_based.c src/quant/quant.h
     src/quant/quant_h263.c src/quant/quant_matrix.c
     src/quant/quant_matrix.h src/quant/quant_mpeg.c
     src/quant/x86_asm/quantize_h263_3dne.asm
     src/quant/x86_asm/quantize_h263_mmx.asm
     src/quant/x86_asm/quantize_mpeg_mmx.asm
     src/quant/x86_asm/quantize_mpeg_xmm.asm
     src/utils/mbtransquant.c src/xvid.c


2003-11-29 18:10:25 GMT                                  patch-113

    Summary:
      Fixed csp asm rules for real?
    Revision:
      xvidcore--devapi4--1.0--patch-113

    Fixed csp asm rules for real?
    

    modified files:
     build/win32/libxvidcore.dsp


2003-11-29 17:58:09 GMT                                  patch-112

    Summary:
      TODO/ChangeLog updated
    Revision:
      xvidcore--devapi4--1.0--patch-112

    TODO/ChangeLog updated
    

    modified files:
     ChangeLog TODO


#########################################################################
# 1.0.0 beta1 (Bitstream Version 21)
#########################################################################

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:
      8x8 16bit Block SSE optimization.
    Revision:
      xvidcore--devapi4--1.0--patch-95

    MMXed the calculation of SSE for 8x8 16bit blocks. This helps quite
    a lot VHQ=4 mode.
    
    My tests show with trellis:chroma_me:
     - ~20% speed improvement for vhq=4.
     - at least 5% when using vhq=1.
    
    Of course this speedup vanishes if more CPU intensive features are used.
    CruNcher who used gmc/qpel, noticed "only" a ~5% speed improvement.
    
    NB: i'm of course talking about overall speed improvement. Such a small
        patch for such a big improvement :-)

    modified files:
     src/motion/estimation_rd_based.c src/motion/sad.c
     src/motion/sad.h src/motion/x86_asm/sad_mmx.asm src/xvid.c


2003-11-13 22:34:33 GMT                                  patch-94

    Summary:
      Various small bug fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-94

    * encoder.c: GMC code fix in encoder.c. Now gmcval is initialized correctly
      when using GME.
    * xvid_decraw.c: Fix elementary stream output.
    * plugin_2pass2.c: Small parsing bug in stats reading in 2pass2.
    * decoder.c: Read resync markers in bframes.

    modified files:
     examples/xvid_decraw.c src/decoder.c src/encoder.c
     src/plugins/plugin_2pass2.c


2003-11-11 16:24:05 GMT                                  patch-93

    Summary:
      VFW update for overflow control
    Revision:
      xvidcore--devapi4--1.0--patch-93

    From Koepi.
    * Added widget and code for overflow control strength.
    * Removed widgets for payback options and kfthresholds.
    * Activated frame stats in DebugOutputView all the time.
    
    From me:
    * Activated static motion detection in cartoon mode.

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/resource.h
     vfw/src/resource.rc


2003-11-09 20:47:47 GMT                                  patch-92

    Summary:
      New two pass code.
    Revision:
      xvidcore--devapi4--1.0--patch-92

    New two pass code. I may say it's just a fixed version, though it looks
    more like a "take all the ideas and write it again" version. It performs
    better with all natural sequences i have and a bit worse with anime.
    
    Including it now, allow me improving the code during the beta releases.

    modified files:
     src/encoder.c src/plugins/plugin_2pass1.c
     src/plugins/plugin_2pass2.c src/xvid.h vfw/src/codec.c
     vfw/src/config.c vfw/src/config.h


2003-11-09 17:07:16 GMT                                  patch-91

    Summary:
      Fixes for bframe compensation (used in psnr tests).
    Revision:
      xvidcore--devapi4--1.0--patch-91

    * transfer_8to16_sub2_(c|mmx|xmm|3dne) write back the compensated
      result to current frame pointer.
    * transfer_8to16_sub2_mmx uses proper rounding (a+b+1)/2. The +1
      operation was missing.
    * Blocks skipped in bframes must be compensated for psnr computing.

    modified files:
     src/encoder.c src/motion/estimation_bvop.c
     src/utils/mem_transfer.c
     src/utils/x86_asm/mem_transfer_3dne.asm
     src/utils/x86_asm/mem_transfer_mmx.asm


2003-11-05 16:05:44 GMT                                  patch-90

    Summary:
      Speed improvement not wasting setedges and interpolate calls.
    Revision:
      xvidcore--devapi4--1.0--patch-90

    Patch from syskin.
    
    * This patch avoids calling setedges and interpolate for uneeded cases:
       - setedges is only called once per frame.
       - interpolate is called only when the previous rounding
         was different from the one needed.
    * Interpolation has been optimized a bit for qpel case, we do the
      hv pass down to top to use the cache more efficiently (hope so).

    modified files:
     src/encoder.c src/encoder.h src/image/image.c


2003-11-03 19:51:12 GMT                                  patch-89

    Summary:
      SSE2 dev16 fix + xvid_bench DCT block alignments.
    Revision:
      xvidcore--devapi4--1.0--patch-89

    * Small error fixed by Skal in his dev16 code (missing pshufd).
    * Blocks used by DCT tests are now aligned with DECLARE_ALIGNED_MATRIX
      this avoids the well know segfaults when using SSE2 instructions that
      suppose data alignment.

    modified files:
     examples/xvid_bench.c src/motion/x86_asm/sad_sse2.asm


2003-11-03 15:42:23 GMT                                  patch-88

    Summary:
      Align .rodata section for non coff objects
    Revision:
      xvidcore--devapi4--1.0--patch-88

    Align .rodata section for non coff objects
    

    modified files:
     src/bitstream/x86_asm/cbp_mmx.asm
     src/bitstream/x86_asm/cbp_sse2.asm
     src/dct/x86_asm/fdct_mmx_ffmpeg.asm
     src/dct/x86_asm/fdct_mmx_skal.asm
     src/dct/x86_asm/fdct_sse2_skal.asm
     src/dct/x86_asm/idct_3dne.asm src/dct/x86_asm/idct_mmx.asm
     src/dct/x86_asm/idct_sse2_dmitry.asm
     src/dct/x86_asm/simple_idct_mmx.asm
     src/image/x86_asm/colorspace_rgb_mmx.asm
     src/image/x86_asm/colorspace_yuyv_mmx.asm
     src/image/x86_asm/interpolate8x8_3dn.asm
     src/image/x86_asm/interpolate8x8_3dne.asm
     src/image/x86_asm/interpolate8x8_mmx.asm
     src/image/x86_asm/interpolate8x8_xmm.asm
     src/image/x86_asm/qpel_mmx.asm
     src/image/x86_asm/reduced_mmx.asm
     src/motion/x86_asm/sad_3dn.asm src/motion/x86_asm/sad_3dne.asm
     src/motion/x86_asm/sad_mmx.asm src/motion/x86_asm/sad_sse2.asm
     src/motion/x86_asm/sad_xmm.asm
     src/quant/x86_asm/quantize_h263_3dne.asm
     src/quant/x86_asm/quantize_h263_mmx.asm
     src/quant/x86_asm/quantize_mpeg_mmx.asm
     src/quant/x86_asm/quantize_mpeg_xmm.asm
     src/utils/x86_asm/cpuid.asm
     src/utils/x86_asm/interlacing_mmx.asm
     src/utils/x86_asm/mem_transfer_3dne.asm


2003-11-02 23:01:43 GMT                                  patch-87

    Summary:
      SSE2 update
    Revision:
      xvidcore--devapi4--1.0--patch-87

    * Added Dmitry SSE2 iDCT code back.
    * Plugged Dmitry iDCT as default for SSE2
    * Fixed a bug in xvid_bench that was making it would test some CPU
      instruction set w/o host CPU support. xvidcore init was simply
      discarding irrelevant cpu flags.

    new files:
     src/dct/x86_asm/.arch-ids/idct_sse2_dmitry.asm.id
     src/dct/x86_asm/idct_sse2_dmitry.asm

    modified files:
     build/generic/sources.inc build/win32/libxvidcore.dsp
     examples/xvid_bench.c src/dct/fdct.h src/dct/idct.h src/xvid.c


2003-10-31 14:53:26 GMT                                  patch-86

    Summary:
      Better handling of old windres versions + GNU make dependency.
    Revision:
      xvidcore--devapi4--1.0--patch-86

    Old versions of GNU windres (<2.14) don't have the same short options.
    But long options remain the same so it's better to use long option names
    to have ful compatibility with older versions.
    
    The Makefile appears to be dependent on GNU make because shell expansion
    for retrieving the path of the Makefile is wrong when using `` even with
    a single expansion assignment :=. It keeps being expanded when used.

    modified files:
     doc/INSTALL vfw/bin/Makefile


2003-10-29 11:31:28 GMT                                  patch-85

    Summary:
      Added sse2 f/iDCT code from skal
    Revision:
      xvidcore--devapi4--1.0--patch-85

    * Added sse2 f/iDCT code from skal
    * Added hooking in xvid.c

    new files:
     src/dct/x86_asm/.arch-ids/fdct_sse2_skal.asm.id
     src/dct/x86_asm/fdct_sse2_skal.asm

    modified files:
     build/generic/sources.inc build/win32/libxvidcore.dsp
     src/xvid.c


2003-10-29 00:19:10 GMT                                  patch-84

    Summary:
      Fix the static motion detection
    Revision:
      xvidcore--devapi4--1.0--patch-84

    Fix the static motion detection
    

    modified files:
     src/motion/estimation_pvop.c


2003-10-28 23:39:46 GMT                                  patch-83

    Summary:
      Added cartoon option handling.
    Revision:
      xvidcore--devapi4--1.0--patch-83

    Added cartoon widgets + handling code.
    
    NB: static motion detection is disabled because of crashes on P4
        cpus.

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/resource.h vfw/src/resource.rc


2003-10-28 17:44:09 GMT                                  patch-82

    Summary:
      ASM cleanups;
    Revision:
      xvidcore--devapi4--1.0--patch-82

    * Applied same style to all asm files
    * Replaced current sad sse2 operators with skal's ones
    * Removed old and unused colorspace asm files

    removed files:
     src/image/x86_asm/.arch-ids/rgb_to_yv12_mmx.asm.id
     src/image/x86_asm/.arch-ids/yuv_to_yv12_mmx.asm.id
     src/image/x86_asm/.arch-ids/yuyv_to_yv12_mmx.asm.id
     src/image/x86_asm/.arch-ids/yv12_to_rgb24_mmx.asm.id
     src/image/x86_asm/.arch-ids/yv12_to_rgb32_mmx.asm.id
     src/image/x86_asm/.arch-ids/yv12_to_yuyv_mmx.asm.id
     src/image/x86_asm/rgb_to_yv12_mmx.asm
     src/image/x86_asm/yuv_to_yv12_mmx.asm
     src/image/x86_asm/yuyv_to_yv12_mmx.asm
     src/image/x86_asm/yv12_to_rgb24_mmx.asm
     src/image/x86_asm/yv12_to_rgb32_mmx.asm
     src/image/x86_asm/yv12_to_yuyv_mmx.asm

    modified files:
     build/generic/sources.inc src/bitstream/x86_asm/cbp_3dne.asm
     src/bitstream/x86_asm/cbp_mmx.asm
     src/bitstream/x86_asm/cbp_sse2.asm
     src/dct/x86_asm/fdct_mmx_ffmpeg.asm
     src/dct/x86_asm/fdct_mmx_skal.asm
     src/dct/x86_asm/idct_3dne.asm src/dct/x86_asm/idct_mmx.asm
     src/dct/x86_asm/simple_idct_mmx.asm
     src/image/x86_asm/colorspace_mmx.inc
     src/image/x86_asm/colorspace_rgb_mmx.asm
     src/image/x86_asm/colorspace_yuv_mmx.asm
     src/image/x86_asm/colorspace_yuyv_mmx.asm
     src/image/x86_asm/interpolate8x8_3dn.asm
     src/image/x86_asm/interpolate8x8_3dne.asm
     src/image/x86_asm/interpolate8x8_mmx.asm
     src/image/x86_asm/interpolate8x8_xmm.asm
     src/image/x86_asm/qpel_mmx.asm
     src/image/x86_asm/reduced_mmx.asm
     src/motion/x86_asm/sad_3dn.asm src/motion/x86_asm/sad_3dne.asm
     src/motion/x86_asm/sad_mmx.asm src/motion/x86_asm/sad_sse2.asm
     src/motion/x86_asm/sad_xmm.asm
     src/quant/x86_asm/quantize_h263_3dne.asm
     src/quant/x86_asm/quantize_h263_mmx.asm
     src/quant/x86_asm/quantize_mpeg_mmx.asm
     src/quant/x86_asm/quantize_mpeg_xmm.asm
     src/utils/x86_asm/cpuid.asm
     src/utils/x86_asm/interlacing_mmx.asm
     src/utils/x86_asm/mem_transfer_3dne.asm
     src/utils/x86_asm/mem_transfer_mmx.asm


2003-10-27 01:13:47 GMT                                  patch-81

    Summary:
      d_mv_bits speedup from sysKin
    Revision:
      xvidcore--devapi4--1.0--patch-81

    d_mv_bits speedup from sysKin
    

    modified files:
     src/motion/motion_inlines.h


2003-10-27 00:55:51 GMT                                  patch-80

    Summary:
      fDCT changes, new asm CodingStyle applied to dct dir
    Revision:
      xvidcore--devapi4--1.0--patch-80

    * Ported the ffmpeg fDCT functions (mmx and xmm).
    * Modified the skal's versions a bit to allow rolling loops.
    * Activated Skal's fDCTs (unrolled versions) for mmx _and_ xmm
      (old code was ignoring xmm versions)
    * Removed the SSE2 versions (they'll be back later)
    * .data -> .rodata
    * Applied announced asm CodingStyle to the dct dir
      (I'll have to add a section with the said CodingStyle)

    modified files:
     build/generic/sources.inc build/win32/libxvidcore.dsp
     src/dct/fdct.h src/dct/idct.h
     src/dct/x86_asm/fdct_mmx_ffmpeg.asm
     src/dct/x86_asm/fdct_mmx_skal.asm
     src/dct/x86_asm/idct_3dne.asm src/dct/x86_asm/idct_mmx.asm
     src/dct/x86_asm/simple_idct_mmx.asm src/xvid.c

    renamed files:
     src/dct/x86_asm/.arch-ids/fdct_mmx.asm.id
       ==> src/dct/x86_asm/.arch-ids/fdct_mmx_ffmpeg.asm.id
     src/dct/x86_asm/.arch-ids/fdct_xmm.asm.id
       ==> src/dct/x86_asm/.arch-ids/fdct_mmx_skal.asm.id
     src/dct/x86_asm/fdct_mmx.asm
       ==> src/dct/x86_asm/fdct_mmx_ffmpeg.asm
     src/dct/x86_asm/fdct_xmm.asm
       ==> src/dct/x86_asm/fdct_mmx_skal.asm


2003-10-25 13:48:42 GMT                                  patch-79

    Summary:
      BQuant->PQuant fix.
    Revision:
      xvidcore--devapi4--1.0--patch-79

    When using closed_gop, a BFrame before an IFrame is turned into a PFrame.
    Thus is original quant has to be computed back, a rounding was causing these
    frames to be orig_quant-1. As a consequence we had very big frames before
    the IFrame, loosing many bits for nearly no visual benefit.

    modified files:
     src/encoder.c


2003-10-25 10:26:48 GMT                                  patch-78

    Summary:
      Added closed gop option to xvid_encraw
    Revision:
      xvidcore--devapi4--1.0--patch-78

    Added closed gop option to xvid_encraw
    

    modified files:
     examples/xvid_encraw.c


2003-10-24 17:39:53 GMT                                  patch-77

    Summary:
      RD fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-77

    Inter RD optimization relied on buggy functions to predict bitsize.

    modified files:
     src/motion/estimation_pvop.c src/motion/motion_inlines.h


2003-10-22 15:48:01 GMT                                  patch-76

    Summary:
      Small INSTALL update due to previous build patches.
    Revision:
      xvidcore--devapi4--1.0--patch-76

    Small INSTALL update due to previous build patches.

    modified files:
     doc/INSTALL


2003-10-21 21:27:46 GMT                                  patch-75

    Summary:
      Removed unused next_block vars.
    Revision:
      xvidcore--devapi4--1.0--patch-75

    Removed unused next_block vars.

    modified files:
     src/decoder.c


2003-10-21 21:24:15 GMT                                  patch-74

    Summary:
      VFW build changes.
    Revision:
      xvidcore--devapi4--1.0--patch-74

    The build system has been modified to look like the core lib one
    minus the configure system.

    modified files:
     vfw/bin/Makefile vfw/bin/sources.inc vfw/src/config.c
     vfw/vfw.dsp

    renamed files:
     vfw/bin/.arch-ids/Makefile.cygwin.id
       ==> vfw/bin/.arch-ids/Makefile.id
     vfw/bin/.arch-ids/Makefile.inc.id
       ==> vfw/bin/.arch-ids/sources.inc.id
     vfw/bin/Makefile.cygwin
       ==> vfw/bin/Makefile
     vfw/bin/Makefile.inc
       ==> vfw/bin/sources.inc
     vfw/src/.arch-ids/config.rc.id
       ==> vfw/src/.arch-ids/resource.rc.id
     vfw/src/config.rc
       ==> vfw/src/resource.rc


2003-10-21 17:00:09 GMT                                  patch-73

    Summary:
      Decoder cleanups and speedup
    Revision:
      xvidcore--devapi4--1.0--patch-73

    cleanups, speedups from sysKin

    modified files:
     src/decoder.c


2003-10-21 16:22:15 GMT                                  patch-72

    Summary:
      Build fixes for newer nasm versions.
    Revision:
      xvidcore--devapi4--1.0--patch-72

    nasm does not take care of adding trailing slashes to include paths.
    A patch to upstream authors has been refused because "the backslash()
    feature has been abandoned to get back to old nasm behavior"
    
    Their choice is kinda stupid as nasm is now open to user mistakes... :\
    So we fix that on ou side.

    modified files:
     build/generic/configure.in build/win32/libxvidcore.dsp


2003-10-17 15:13:12 GMT                                  patch-71

    Summary:
      Updated docs.
    Revision:
      xvidcore--devapi4--1.0--patch-71

    The doc of devapi4 is mostly out dated, i t is much better not to
    keep it in the repository at the moment. We'll add new docs later.
    
    Added a INSTALL doc that explains the build/install process for
    supported platforms. It's a first try, things may be added later.

    new files:
     doc/.arch-ids/INSTALL.id doc/INSTALL

    removed files:
     doc/.arch-ids/API.dox.id doc/.arch-ids/Makefile.id
     doc/.arch-ids/foot.inc.in.id doc/.arch-ids/header.tex.in.id
     doc/.arch-ids/xvid-decoding.txt.id
     doc/.arch-ids/xvid-encoder.txt.id doc/API.dox doc/Makefile
     doc/foot.inc.in doc/header.tex.in doc/xvid-decoding.txt
     doc/xvid-encoder.txt

    modified files:
     CodingStyle README doc/README

    renamed files:
     .arch-ids/README.txt.id
       ==> .arch-ids/README.id
     .arch-ids/changelog.txt.id
       ==> .arch-ids/ChangeLog.id
     README.txt
       ==> README
     changelog.txt
       ==> ChangeLog


2003-10-15 13:53:11 GMT                                  patch-70

    Summary:
      Better cross compilation handling.
    Revision:
      xvidcore--devapi4--1.0--patch-70

    With this patch it is now possible to cross compile xvid
    quite easily for win32 platform on a build linux host.
    
    Recipe for debian system:
     $ apt-get install mingw32
     (or create your own cross compiler/binutils suite and install mingw32
      header files -- sorry i don't have a recipe for this, this is let as
      an exercice for the reader)
     $ cd ${xvidcore}
     $ cd build/generic
     $ ./bootstrap.sh
     $ ./configure --host=i586-mingw32msvc
      (all occurences of i586-mingw32msvc may be replaced with the right
       prefix you've choosen for your cross compiler and cross binutils)
     $ make
     $ cd ../../vfw/bin
     $ make -f Makefile.cygwin \
      CC=i586-mingw32msvc-gcc WINDRES=i586-mingw32msvc-windres
    
    Enjoy your win32 xvid.dll build by free software, on a free OS, for a devil
    OS target.

    modified files:
     build/generic/configure.in vfw/bin/Makefile.cygwin


2003-10-14 15:17:28 GMT                                  patch-69

    Summary:
      Fixed Qpel+Interpolation decoding. Cleaned up mb->mode usage.
    Revision:
      xvidcore--devapi4--1.0--patch-69

    * Fixed interpolate mode + qpel decoding.
    * MB->mb_type completely replaced by MB->mode

    modified files:
     src/decoder.c


2003-10-12 21:57:24 GMT                                  patch-68

    Summary:
      ac/dc prediction for intra RD search.
    Revision:
      xvidcore--devapi4--1.0--patch-68

    From syskin, added real ac/dc prediction for INTRA's bitcount.

    modified files:
     src/motion/estimation_rd_based.c


2003-10-09 18:15:50 GMT                                  patch-67

    Summary:
      Pigrated asm code to new quant API.
    Revision:
      xvidcore--devapi4--1.0--patch-67

    Many changes that are mostly cosmetic in the asm files.
     * indent
     * added xor eax, eax in quant_(h263|mpeg)_intra_.* functions
       (just to make sure the returned value isn't random)
     * added xor eax, eax in dequant_(h263|mpeg)_.* functions
       (just to make sure the returned value isn't random)
     * synced cpuid.asm XVID_CPU_feature constants with the one
       defined in the C code (xvid.h)
     * enabled all cpu tests in xvid_bench.c

    modified files:
     examples/xvid_bench.c src/quant/quant_h263.c
     src/quant/quant_mpeg.c
     src/quant/x86_asm/quantize_h263_3dne.asm
     src/quant/x86_asm/quantize_h263_mmx.asm
     src/quant/x86_asm/quantize_mpeg_mmx.asm
     src/quant/x86_asm/quantize_mpeg_xmm.asm
     src/utils/x86_asm/cpuid.asm src/xvid.h


2003-10-08 21:05:47 GMT                                  patch-66

    Summary:
      Updated xvid_bench for quant API changes
    Revision:
      xvidcore--devapi4--1.0--patch-66

    Updated xvid_bench for quant API changes
    

    modified files:
     examples/xvid_bench.c


2003-10-07 13:03:51 GMT                                  patch-65

    Summary:
      Quant functions API changes (first step)
    Revision:
      xvidcore--devapi4--1.0--patch-65

    In the road to instance safe mpeg quantization, a small cleanup
    to the quant API was needed. It consists in changing the way we
    name the functions quant_{mpeg|h263}_{inter|intra}_{arch} and in
    a move to a more unified API (even intra functions return the sum
    of coefficients, it can be used as a complexity measure at a later
    time).
    
    This patch touch lot of files, but all changes are trivial.
    
    NB: we should check the IA64 asm validity, i changed things but
        i can't test them.

    new files:
     src/quant/.arch-ids/quant.h.id src/quant/quant.h

    removed files:
     src/quant/.arch-ids/quant_h263.h.id
     src/quant/.arch-ids/quant_mpeg4.h.id src/quant/quant_h263.h
     src/quant/quant_mpeg4.h

    modified files:
     build/generic/sources.inc build/win32/libxvidcore.dsp
     src/decoder.c src/encoder.c src/image/qpel.c
     src/motion/estimation_rd_based.c
     src/quant/ia64_asm/quant_h263_ia64.s src/quant/quant_h263.c
     src/quant/quant_matrix.c src/quant/quant_matrix.h
     src/quant/quant_mpeg.c
     src/quant/x86_asm/quantize_h263_3dne.asm
     src/quant/x86_asm/quantize_h263_mmx.asm
     src/quant/x86_asm/quantize_mpeg_mmx.asm
     src/quant/x86_asm/quantize_mpeg_xmm.asm
     src/utils/mbtransquant.c src/xvid.c src/xvid.h

    renamed files:
     src/quant/.arch-ids/quant_mpeg4.c.id
       ==> src/quant/.arch-ids/quant_mpeg.c.id
     src/quant/quant_mpeg4.c
       ==> src/quant/quant_mpeg.c
     src/quant/x86_asm/.arch-ids/quantize4_mmx.asm.id
       ==> src/quant/x86_asm/.arch-ids/quantize_mpeg_mmx.asm.id
     src/quant/x86_asm/.arch-ids/quantize4_xmm.asm.id
       ==> src/quant/x86_asm/.arch-ids/quantize_mpeg_xmm.asm.id
     src/quant/x86_asm/.arch-ids/quantize_3dne.asm.id
       ==> src/quant/x86_asm/.arch-ids/quantize_h263_3dne.asm.id
     src/quant/x86_asm/.arch-ids/quantize_mmx.asm.id
       ==> src/quant/x86_asm/.arch-ids/quantize_h263_mmx.asm.id
     src/quant/x86_asm/quantize4_mmx.asm
       ==> src/quant/x86_asm/quantize_mpeg_mmx.asm
     src/quant/x86_asm/quantize4_xmm.asm
       ==> src/quant/x86_asm/quantize_mpeg_xmm.asm
     src/quant/x86_asm/quantize_3dne.asm
       ==> src/quant/x86_asm/quantize_h263_3dne.asm
     src/quant/x86_asm/quantize_mmx.asm
       ==> src/quant/x86_asm/quantize_h263_mmx.asm


2003-10-05 00:15:15 GMT                                  patch-64

    Summary:
      Updated ChangeLog
    Revision:
      xvidcore--devapi4--1.0--patch-64

    Updated ChangeLog
    

    modified files:
     changelog.txt


2003-10-04 16:04:30 GMT                                  patch-63

    Summary:
      Removed legacy 2pass code from vfw
    Revision:
      xvidcore--devapi4--1.0--patch-63

    Removed legacy 2pass code from vfw
    

    removed files:
     vfw/src/.arch-ids/2pass.c.id vfw/src/.arch-ids/2pass.h.id
     vfw/src/2pass.c vfw/src/2pass.h


2003-10-04 00:41:38 GMT                                  patch-62

    Summary:
      Working VFW mingw/cygwin build system.
    Revision:
      xvidcore--devapi4--1.0--patch-62

    This patch fixes the VFW building process. Now it should work out
    of the box using these steps:
     # cd ${xvidcore}
     # cd build/generic
     # ./bootstrap.sh <-- only needed for CVS checkouts.
     # ./configure
     # make
     # cd ../../vfw/bin
     # make -f Makefile.cygwin
    
    Then install as usual clicking on the inf file or "make install"
    in the vfw/bin dir.

    modified files:
     vfw/bin/Makefile.cygwin vfw/src/config.rc vfw/src/debug.h
     vfw/src/driverproc.c


2003-10-03 17:00:53 GMT                                  patch-61

    Summary:
      Fixes for alternate scan and interlacing support.
    Revision:
      xvidcore--devapi4--1.0--patch-61

    Fixes from CVS (by sysKin) for:
     - added alternate scan support with VHQ
     - fixed interlacing support in s/b-frames. May fix a potential problem
       as field_pred struct field seemed not to be initialized anywhere.
       As it's not supported yet, write an hardcoded 0 bit.
    
    Fixes from me for the fixes from syKin:
     - scan_table effectively used in MBCodingBVOP
     - Block_CalcBits(Intra) fixes to data->scan_table (implies prototype
       change and code modification every where the functions were used)
    
    I also increased BS version as it might result in different bitstreams
    It's now at version 20.

    modified files:
     src/bitstream/mbcoding.c src/bitstream/mbcoding.h
     src/decoder.c src/encoder.c src/motion/estimation.h
     src/motion/estimation_rd_based.c src/xvid.h


2003-10-03 15:41:37 GMT                                  patch-60

    Summary:
      Removed BIGLUT support.
    Revision:
      xvidcore--devapi4--1.0--patch-60

    Remobed legacy code for BIGLUT support. It was unused and RD based
    Motion Estimation was not even compatible with this type of VLC coding.

    modified files:
     build/generic/configure.in src/bitstream/mbcoding.c
     src/bitstream/mbcoding.h src/prediction/mbprediction.c


2003-10-03 13:25:17 GMT                                  patch-59

    Summary:
      Bugfix for PFrames+ Ext Search.
    Revision:
      xvidcore--devapi4--1.0--patch-59

    In Qpel mode, the code was doing a diamond search for wrong predictors.
    This resulted in poor performance as the diamond search was sitting there
    for sometime.

    modified files:
     src/motion/estimation.h src/motion/estimation_pvop.c


2003-10-02 16:50:51 GMT                                  patch-58

    Summary:
      Added VFW makefile for cygwin/minsys
    Revision:
      xvidcore--devapi4--1.0--patch-58

    Added VFW makefile for cygwin/minsys. I can't test it so it is probably not
    right out of the box. Waiting for feedback in order to fix it.

    new files:
     vfw/bin/.arch-ids/Makefile.cygwin.id
     vfw/bin/.arch-ids/Makefile.inc.id vfw/bin/Makefile.cygwin
     vfw/bin/Makefile.inc vfw/src/w32api/.arch-ids/=id
     vfw/src/w32api/.arch-ids/vfw.h.id vfw/src/w32api/vfw.h

    new directories:
     vfw/src/w32api vfw/src/w32api/.arch-ids


2003-10-02 13:35:15 GMT                                  patch-57

    Summary:
      Cleaned up the lumimasking code.
    Revision:
      xvidcore--devapi4--1.0--patch-57

    The lumimasking code was not very plugin oriented as it has been ported
    from old XviD versions. This patch cleans up the code and integrates it
    better with plugin design. No changes done in teh functionnal code.

    modified files:
     src/plugins/plugin_lumimasking.c


2003-10-01 23:07:07 GMT                                  patch-56

    Summary:
      Cleaned up trailing space chars.
    Revision:
      xvidcore--devapi4--1.0--patch-56

    The kind of patch we would love to avoid as they make merging
    a nightmare while they're kind of useless patches.
    
    Applied sed 's/[ \t]*$//' to all c/h files.

    modified files:
     src/bitstream/bitstream.c src/bitstream/bitstream.h
     src/bitstream/cbp.c src/bitstream/mbcoding.c src/dct/idct.c
     src/dct/simple_idct.c src/decoder.c src/decoder.h
     src/encoder.c src/encoder.h src/global.h
     src/image/colorspace.c src/image/colorspace.h src/image/font.c
     src/image/image.c src/image/image.h src/image/interpolate8x8.c
     src/image/interpolate8x8.h src/image/qpel.c src/image/qpel.h
     src/image/reduced.c src/motion/estimation_rd_based.c
     src/motion/gmc.c src/motion/gmc.h src/motion/motion.h
     src/motion/sad.c src/plugins/plugin_2pass1.c
     src/plugins/plugin_2pass2.c src/plugins/plugin_fixed.c
     src/plugins/plugin_psnr.c src/plugins/plugin_single.c
     src/portab.h src/prediction/mbprediction.c
     src/utils/mbtransquant.c src/utils/timer.c src/xvid.c
     src/xvid.h


2003-09-30 18:10:18 GMT                                  patch-55

    Summary:
      Code cleanups.
    Revision:
      xvidcore--devapi4--1.0--patch-55

    It's been a while since the last ISOC89 conformance cleanup.
    
    Using the following switches help a lot :-)
     -Wall -Wsign-compare -Wredundant-decls -Wunreachable-code -Wnested-externs \
     -ansi
    
    Result: 0 warning/0 error

    modified files:
     src/bitstream/vlc_codes.h src/global.h src/image/qpel.h
     src/motion/estimation.h src/motion/estimation_bvop.c
     src/motion/estimation_gmc.c src/motion/estimation_pvop.c
     src/motion/estimation_rd_based.c src/motion/gmc.c
     src/motion/gmc.h src/motion/motion_inlines.h
     src/motion/vop_type_decision.c


2003-09-29 00:31:32 GMT                                  patch-54

    Summary:
      Memory leakage fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-54

    The pEnc->queue was allocated but not freed when bframes == 0.
    And queue images were not freed as well.

    modified files:
     examples/Makefile src/encoder.c src/image/image.c
     src/utils/mem_align.c


2003-09-28 16:45:02 GMT                                  patch-53

    Summary:
      Fixes the unitialized mcsel bit in RD based ME
    Revision:
      xvidcore--devapi4--1.0--patch-53

    Fixes the unitialized mcsel bit in RD based ME
    

    modified files:
     src/motion/estimation_rd_based.c


2003-09-28 01:00:06 GMT                                  patch-52

    Summary:
      Fix the XviD constant version initialization
    Revision:
      xvidcore--devapi4--1.0--patch-52

    Fix the XviD constant version initialization
    

    modified files:
     src/xvid.h


2003-09-28 00:47:05 GMT                                  patch-51

    Summary:
      Fix to the build system (the come back).
    Revision:
      xvidcore--devapi4--1.0--patch-51

    Ok this one fixes the way we build the targets. The VPATH thingy really
    works with the library targets. My understanding of the VPATH mechanism
    was wrong. Now it should be ok (i promess).

    modified files:
     build/generic/Makefile


2003-09-27 11:45:18 GMT                                  patch-50

    Summary:
      Small fix to previous patch.
    Revision:
      xvidcore--devapi4--1.0--patch-50

    A pair ofdouble quotes prevented "make" to sort out the VPATH
    dependences in "=build". This resulted in compiling all the
    sources, all the time, whether a file changed or not.

    modified files:
     build/generic/Makefile


2003-09-26 22:39:44 GMT                                  patch-49

    Summary:
      Updated the build files for *nix.
    Revision:
      xvidcore--devapi4--1.0--patch-49

    - Fixed MacOSX build (w/o module option). The subversion was not right
      it was just minor version though it has to be major.minor.
    - Fixed bootstrap.sh for MacOSX environment, it now looks for
      glibtoolize if libtoolize is not found.
    - The unified Makefile now builds XviD out of source tree in the directory
      =build. It's cleaner, and clashes much less w/ arch/tla source linting.
    - Tuned the tagging regexps so:
      + autoconf files are ignored (considered backup => not erased, not copied
        and not commited/imported)
      + =build is also considered backup.

    modified files:
     build/generic/Makefile build/generic/bootstrap.sh
     build/generic/configure.in build/generic/platform.inc.in
     build/generic/sources.inc examples/Makefile
     {arch}/=tagging-method


2003-09-24 01:38:03 GMT                                  patch-48

    Summary:
      Bug fix to decoder (mcsel/acpred bits swapped)
    Revision:
      xvidcore--devapi4--1.0--patch-48

    As reported here:
    http://www.xvid.org/modules.php?op=modload&name=phpBB2&file=viewtopic&t=1513&highlight=
    
    <quote>
    in the spec, 'there is 'mcsel' before 'ac_pred_flag'.
    however, if you see the code, it's changed.
    </quote>
    
    We were doing the opposite, ac_pred before mcsel.

    modified files:
     src/decoder.c


2003-09-11 17:19:35 GMT                                  patch-47

    Summary:
      Small fix to GMC+QuarterPel -- BS version bumped to 19
    Revision:
      xvidcore--devapi4--1.0--patch-47

    During the split up, a line has been disabled that prevented good
    quartelpel+GMC. Fixed.
    
    This fix + patch-43 are a good reason to bump up the bitstream
    version to 19.

    modified files:
     src/motion/gmc.h src/xvid.h


2003-09-11 17:11:28 GMT                                  patch-46

    Summary:
      Build process fix for MacOSX+module option
    Revision:
      xvidcore--devapi4--1.0--patch-46

    This patch fixes the install rule of the MacOSX module style library.
    It adds a PRE_SHARED_LIB == SHARED_LIB for all platforms except
    MacOSX that uses it a different way, and exploits the fact
    $(SPECIFIC_LDFLAGS) is the last var of the build line to insert a
    command for post linking the right .so file.

    modified files:
     build/generic/Makefile build/generic/configure.in
     build/generic/platform.inc.in


2003-09-11 13:56:40 GMT                                  patch-45

    Summary:
      Tree cleanup and build files updated.
    Revision:
      xvidcore--devapi4--1.0--patch-45

    * Win32 files switched to Unix format.
      Ok, on IRC, we sorted out what was b0rking the project files each
      time I commit them in DOS format. In fact the Unix CVS does upload
      DOS format files if i commit them, and then Win32 CVS users get
      double \r\n files. So the best is to use unix format in CVS. This
      way, the checked out versions are right for Win32 users.
    
      NB: this imply a unix2dos conversion when doing a release from a
          Unix box. I think it's better than current policy, but automatic
          tarballs/zip of the tree should take care of that.
    
    * odivx and xvid_stat examples removed. Associated project files have
      been removed as well.
    
    * The divx4 compatibility layer has been removed. Associated options
      in the configure script removed. libxvidcore.def has no more need
      to be an autoconf generated file.
    
    * rawdec removed. It has never been used and will never be.
    
    * Added major api appending to the macosx module build.

    removed files:
     build/win32/.arch-ids/odivx_enc_dec.dsp.id
     build/win32/.arch-ids/xvid_stat.dsp.id
     build/win32/odivx_enc_dec.dsp build/win32/xvid_stat.dsp
     examples/.arch-ids/odivx_enc_dec.c.id
     examples/.arch-ids/xvid_stat.c.id examples/odivx_enc_dec.c
     examples/xvid_stat.c rawdec/.arch-ids/=id
     rawdec/.arch-ids/rawdec.c.id rawdec/.arch-ids/rawdec.dsp.id
     rawdec/rawdec.c rawdec/rawdec.dsp src/.arch-ids/divx4.c.id
     src/.arch-ids/divx4.h.id src/divx4.c src/divx4.h

    modified files:
     build/generic/Makefile build/generic/configure.in
     build/generic/libxvidcore.def build/win32/libxvidcore.dsp
     build/win32/xvid_decraw.dsp build/win32/xvid_encraw.dsp
     build/win32/xvidcore.dsw

    renamed files:
     build/generic/.arch-ids/libxvidcore.def.in.id
       ==> build/generic/.arch-ids/libxvidcore.def.id
     build/generic/libxvidcore.def.in
       ==> build/generic/libxvidcore.def

    removed directories:
     rawdec rawdec/.arch-ids


2003-09-11 12:59:19 GMT                                  patch-44

    Summary:
      Replaced malloc.h header file with stdlib.h
    Revision:
      xvidcore--devapi4--1.0--patch-44

    Replaced malloc.h header file with stdlib.h
    

    modified files:
     src/plugins/plugin_lumimasking.c


2003-09-10 22:33:04 GMT                                  patch-43

    Summary:
      Fixed a problem for 'power of 2' framerates
    Revision:
      xvidcore--devapi4--1.0--patch-43

    Fixed a problem for 'power of 2' framerates
    

    modified files:
     src/bitstream/bitstream.c


2003-09-10 21:57:12 GMT                                  patch-42

    Summary:
      Motion Estimation module splitting.
    Revision:
      xvidcore--devapi4--1.0--patch-42

    The motion estimation module was the biggest file of the source
    tree. After some previous attempts, sysKin decided to split it
    up again. This time he's done it right.
    
    This split up just changes the organization of ME functions
    inside different files. This should help a bit in keeping the
    motion estimation being manageable by a normal human ;-)
    
    Here is the splitting logic quoted from sysKin's email to
    xvid-devel:
     - estimation.h: header files #included in all ME modules:
                     + deftypes,
                     + macros
                     + constants
                     NB: no code.
     - estimation_bvop.c: motion estimation for b-vops.
                          everything in it :)
     - estimation_common.c: functions shared among all ME
                            modules:
                            + diamonds
                            + subpel
                            + refinement
                            + picture
                            + manipulation
                            + tables
                            + ...  etc.
     - estimation_gmc.c: gruel's GME code
     - estimation_pvop.c: ME for p-vops. Also SAD-based mode
                          decision
     - estimation_rd_based.c: everything R-D-based: mode
                              decison (including _Fast) and ME.
     - gmc.c, gmc.h: no change.

    new files:
     src/motion/.arch-ids/estimation.h.id
     src/motion/.arch-ids/estimation_bvop.c.id
     src/motion/.arch-ids/estimation_common.c.id
     src/motion/.arch-ids/estimation_gmc.c.id
     src/motion/.arch-ids/estimation_pvop.c.id
     src/motion/.arch-ids/estimation_rd_based.c.id
     src/motion/.arch-ids/motion_inlines.h.id
     src/motion/.arch-ids/vop_type_decision.c.id
     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/motion_inlines.h src/motion/vop_type_decision.c

    removed files:
     src/motion/.arch-ids/motion_est.c.id
     src/motion/.arch-ids/motion_est.h.id
     src/motion/.arch-ids/smp_motion_est.c.id
     src/motion/.arch-ids/smp_motion_est.h.id
     src/motion/motion_est.c src/motion/motion_est.h
     src/motion/smp_motion_est.c src/motion/smp_motion_est.h

    modified files:
     build/generic/sources.inc build/win32/libxvidcore.dsp
     src/bitstream/mbcoding.h src/motion/gmc.c src/motion/gmc.h
     src/motion/motion.h src/motion/motion_comp.c src/motion/sad.c
     src/motion/sad.h src/prediction/mbprediction.c
     src/prediction/mbprediction.h src/utils/mbfunctions.h


2003-09-10 00:40:44 GMT                                  patch-41

    Summary:
      Cleanups and fix to (trellis+thresholding) logic
    Revision:
      xvidcore--devapi4--1.0--patch-41

    Cleanups to some functions (loop unrolling, call to functions
    through function array pointers)...
    
    Fix to the trellis+thresholding logic. It was comparing the
    return value of trellis with a threshold but the trellis function
    returns the last non zero coeff index... this was basically
    comparing apples with oranges... funny but wrong. Trellis now
    returns the sum of absolute coeffs, so the comparison is logical.
    Btw, as discussed on the devel ML, this is probably uneeded as
    trellis does an RD optimized coeff distribution.

    modified files:
     src/utils/mbtransquant.c


2003-09-09 13:13:58 GMT                                  patch-40

    Summary:
      Missing ressource for dshow frontend
    Revision:
      xvidcore--devapi4--1.0--patch-40

    Missing ressource for dshow frontend
    

    new files:
     dshow/src/.arch-ids/XviD_logo.bmp.id dshow/src/XviD_logo.bmp


2003-09-08 11:02:10 GMT                                  patch-39

    Summary:
      Small fixes for fast mode decision
    Revision:
      xvidcore--devapi4--1.0--patch-39

    Small fixes for fast mode decision
    

    modified files:
     src/motion/motion_est.c src/xvid.h


2003-09-05 23:45:48 GMT                                  patch-38

    Summary:
      New RD mode decision and subpel refinement.
    Revision:
      xvidcore--devapi4--1.0--patch-38

    New stuff from michael. It deals with mode decision and subpel
    refinement. Integration of these new flags are not settled.
    
    Wait and see. Further testing is needed.

    modified files:
     src/motion/motion_est.c src/motion/motion_est.h src/xvid.h


2003-08-29 13:56:30 GMT                                  patch-37

    Summary:
      Still more ME tuning
    Revision:
      xvidcore--devapi4--1.0--patch-37

    Still more ME tuning
    

    modified files:
     src/motion/motion_est.c


2003-08-28 12:43:22 GMT                                  patch-36

    Summary:
      Removed expanded the cvs Id field
    Revision:
      xvidcore--devapi4--1.0--patch-36

    Removed expanded the cvs Id field
    

    modified files:
     src/image/x86_asm/qpel_mmx.asm


2003-08-28 12:39:44 GMT                                  patch-35

    Summary:
      More motion est cleanup and bugfixes.
    Revision:
      xvidcore--devapi4--1.0--patch-35

    Still more bugfixes, cleanups and improvements to the Motion Est
    by sysKin

    modified files:
     src/motion/motion_est.c src/motion/motion_est.h


2003-08-26 13:57:39 GMT                                  patch-34

    Summary:
      Added final bits of Aspect Ratio flag.
    Revision:
      xvidcore--devapi4--1.0--patch-34

    Peter did think of the AR flag since the very beginning of devapi4.
    He just forgot to code the final bits to effectively write it to
    the bitstream. This patch adds these missing final bits so XviD now
    reads and writes AR flags.

    modified files:
     ./src/bitstream/bitstream.c ./src/encoder.c ./src/encoder.h
     ./src/xvid.h


2003-08-25 16:41:09 GMT                                  patch-33

    Summary:
      Small motion estimation cleanup.
    Revision:
      xvidcore--devapi4--1.0--patch-33

    Cleanups from sysKin.

    modified files:
     ./src/motion/motion_est.c ./src/motion/motion_est.h


2003-08-25 14:59:28 GMT                                  patch-32

    Summary:
      Frame padding bug.
    Revision:
      xvidcore--devapi4--1.0--patch-32

    We were edging the image repeating pixels from the image directly,
    but the standard says we must repeat from a 16 pixel boundary.
    
    See Chapter 7.6.4 of the standard.

    modified files:
     ./src/image/image.c


2003-08-23 15:07:44 GMT                                  patch-31

    Summary:
      New Qpel code.
    Revision:
      xvidcore--devapi4--1.0--patch-31

    Isibaar commited a new piece of QPel code that seems to be optimized
    for ia32(mmx) architectures. I had to clean it up a bit to make it
    respectful of architecture separations. This code is disabled for
    non ia32 arch, a comment mention it's only faster on ia32... i wonder
    if it's true, some tests have to be done on sourceforge compile farm
    in order to confirm that.
    
    Compared to the bared CVS commit, this patch includes:
     - fixes the unix build.
     - better architecture separation.
     - CodingStyle respected.

    new files:
     ./src/image/.arch-ids/qpel.c.id
     ./src/image/.arch-ids/qpel.h.id ./src/image/qpel.c
     ./src/image/qpel.h
     ./src/image/x86_asm/.arch-ids/qpel_mmx.asm.id
     ./src/image/x86_asm/qpel_mmx.asm

    modified files:
     ./build/generic/sources.inc ./build/win32/libxvidcore.dsp
     ./src/encoder.c ./src/motion/motion_comp.c ./src/xvid.c


2003-08-22 13:20:36 GMT                                  patch-30

    Summary:
      sad32v does really what it's expected (ie 32x32 SAD :-)
    Revision:
      xvidcore--devapi4--1.0--patch-30

    sad32v does really what it's expected (ie 32x32 SAD :-)
    

    modified files:
     ./src/motion/sad.c


2003-08-18 19:00:49 GMT                                  patch-29

    Summary:
      64bit fix.
    Revision:
      xvidcore--devapi4--1.0--patch-29

    The interpolation code was unsafe on 64bit platforms, the offset was
    badly sized, resulting in segfaults.

    modified files:
     ./src/image/interpolate8x8.h


2003-08-17 14:08:48 GMT                                  patch-28

    Summary:
      Greyscale mode fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-28

    We were missing some greyscale tests in the encoder loop... noticeably
    in the PVOP function when coding an intra block, and in BVOP function
    when coding all types.
    
    I added the cbp trick in the cases discussed above.

    modified files:
     ./src/encoder.c


2003-08-13 11:47:33 GMT                                  patch-27

    Summary:
      Forgotten bit for IA64 separation
    Revision:
      xvidcore--devapi4--1.0--patch-27

    Forgotten bit for IA64 separation
    

    modified files:
     ./src/image/interpolate8x8.h


2003-08-11 15:42:30 GMT                                  patch-26

    Summary:
      Some qpel changes (sync with Isibaar)
    Revision:
      xvidcore--devapi4--1.0--patch-26

    Some qpel changes (sync with Isibaar)
    

    modified files:
     ./examples/xvid_encraw.c


2003-08-11 15:30:04 GMT                                  patch-25

    Summary:
      Better architecture separation.
    Revision:
      xvidcore--devapi4--1.0--patch-25

    Architecture depending functions were declared whatever arch you
    were compiling. This patch fixes that.
    
    I also removed the simple_idct hack in decoder.c as it was simply
    not used. Better not have ugly code in there.
    
    xvid_bench, should now compile and run on all archs. However I did
    not put the cpu definitions for each arch, i just separated
    ARCH_IS_IA32 so even ARCH_IS_GENERIC can compile

    modified files:
     ./examples/xvid_bench.c ./src/bitstream/cbp.h ./src/dct/fdct.h
     ./src/dct/idct.h ./src/decoder.c ./src/image/colorspace.h
     ./src/image/interpolate8x8.h ./src/image/reduced.h
     ./src/motion/sad.h ./src/quant/quant_h263.h
     ./src/quant/quant_mpeg4.h ./src/utils/emms.h


2003-08-09 20:47:42 GMT                                  patch-24

    Summary:
      Updated changelog
    Revision:
      xvidcore--devapi4--1.0--patch-24

    Updated changelog
    

    modified files:
     ./changelog.txt


2003-08-09 20:31:17 GMT                                  patch-23

    Summary:
      Workaround to a GMC bug due to a MS compiler bug.
    Revision:
      xvidcore--devapi4--1.0--patch-23

    Christoph did use a trick to speed up code that resulted in badly
    optimized code (teh compiler was missing a cast)

    modified files:
     ./src/motion/gmc.c


2003-08-09 17:09:00 GMT                                  patch-22

    Summary:
      Fixes to xvid_decraw
    Revision:
      xvidcore--devapi4--1.0--patch-22

    xvid_decraw has always been used on not so high bitrate sequences
    and not so big sequences neither.
    
    I've been doing lot of tests on the Matrix 2 trailer
    (1000x540 25fps ~5Mbits/s) and xvid_decraw was not able to handle that
    because of bugs in the buffer filling algorithm.
    
    This patch fixes the buffer filling and catches up with christoph changes
    in CVS (wrong help message and option parsing).

    modified files:
     ./examples/xvid_decraw.c


2003-08-08 21:31:59 GMT                                  patch-21

    Summary:
      Added QPel and GMC options.
    Revision:
      xvidcore--devapi4--1.0--patch-21

    XviD has so many options that we forget to propose them all on the
    CLI, here are two more: GMC and Qpel.

    modified files:
     ./examples/xvid_encraw.c


2003-08-07 19:26:28 GMT                                  patch-20

    Summary:
      SVOP handling in status window
    Revision:
      xvidcore--devapi4--1.0--patch-20

    SVOP handling in status window
    

    modified files:
     ./vfw/src/status.c


2003-08-07 19:25:03 GMT                                  patch-19

    Summary:
      Warning cleanups by chl
    Revision:
      xvidcore--devapi4--1.0--patch-19

    Warning cleanups by chl
    

    modified files:
     ./src/encoder.c ./src/encoder.h ./src/image/interpolate8x8.c
     ./src/motion/motion_est.c ./src/xvid.h


2003-08-06 21:13:35 GMT                                  patch-18

    Summary:
      Fix to GMC sprite trajectory code
    Revision:
      xvidcore--devapi4--1.0--patch-18

    Fix to GMC sprite trajectory code
    

    modified files:
     ./src/bitstream/mbcoding.c ./src/xvid.h


2003-08-06 10:57:25 GMT                                  patch-17

    Summary:
      Fixes a bug in BVOP block skipping thresholding
    Revision:
      xvidcore--devapi4--1.0--patch-17

    Fixes a bug in BVOP block skipping thresholding
    

    modified files:
     ./src/motion/motion_est.c


2003-08-03 14:57:32 GMT                                  patch-16

    Summary:
      Functions renaming + motion fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-16

    BITS flags have been renamed to RD (Rate Distorsion) flags... however
    function names were still xxxBitsxxx.
    
    Improved frame type decision
    
    Fix for DQUANTS plugins, their quant was never checked against valid
    [1..31] range.

    modified files:
     ./src/encoder.c ./src/motion/motion_est.c
     ./src/motion/motion_est.h


2003-08-02 15:00:49 GMT                                  patch-15

    Summary:
      API cleanup.
    Revision:
      xvidcore--devapi4--1.0--patch-15

    Since we started devapi3 and then devapi4, feature names did not change
    because it was just convenient to keep them to minimize the change impact.
    But most of the flags were now not even suggesting what they do. So this
    patch cleans the API.
    
    This patch also change the way we describe flags, it's more compact and
    shows better flags are bit sets that must not overlap. This change fixes
    a plugin flag overlapping problem as well.

    modified files:
     ./examples/xvid_encraw.c ./src/decoder.c ./src/encoder.c
     ./src/encoder.h ./src/motion/motion_est.c
     ./src/motion/motion_est.h ./src/xvid.c ./src/xvid.h
     ./vfw/src/codec.c ./vfw/src/config.c


2003-07-29 22:25:12 GMT                                  patch-14

    Summary:
      Fixed bogus memory accesses
    Revision:
      xvidcore--devapi4--1.0--patch-14

    Fixed bogus memory accesses
    

    modified files:
     ./src/encoder.c ./src/plugins/plugin_2pass1.c


2003-07-28 12:22:33 GMT                                  patch-13

    Summary:
      Bitstream version increased to 16
    Revision:
      xvidcore--devapi4--1.0--patch-13

    Bitstream version increased to 16
    

    modified files:
     ./src/xvid.h


2003-07-25 12:01:51 GMT                                  patch-12

    Summary:
      Added gmc files to teh windows project file
    Revision:
      xvidcore--devapi4--1.0--patch-12

    Added gmc files to teh windows project file
    

    modified files:
     ./build/win32/libxvidcore.dsp ./vfw/src/codec.c


2003-07-25 12:00:31 GMT                                  patch-11

    Summary:
      Added cartoon mode from Isibaar
    Revision:
      xvidcore--devapi4--1.0--patch-11

    Added cartoon mode from Isibaar
    

    modified files:
     ./src/motion/motion_est.c ./src/plugins/plugin_single.c
     ./src/utils/mbtransquant.c ./src/xvid.h


2003-07-25 10:30:41 GMT                                  patch-10

    Summary:
      Bitstream syntax comments.
    Revision:
      xvidcore--devapi4--1.0--patch-10

    This patch does not change the bitstream but adds some comments that
    can help in order to understand (lack of) calls to BitstreamPadAlways.

    modified files:
     ./src/bitstream/bitstream.c ./src/encoder.c


2003-07-22 16:34:25 GMT                                  patch-9

    Summary:
      Fixes Bistream errors in VOL (+ forced stuffing)
    Revision:
      xvidcore--devapi4--1.0--patch-9

    After a detailed bugreport at:
    http://www.xvid.org/modules.php?op=modload&name=phpBB2&file=viewtopic&t=1387&highlight=
    
    I discovered that:
    
    1/ we did not write video_signal_type, but we were padding to the next
    byte, that's why we had video_signal_type=0 and then only 1s until the
    next byte boundary. This explains the 11 next_start_code();
    
    2/ video_object_type_indication = Reserved is right on my machine,
    please check again, but i doubt there is a bug there, we use 3 hard
    wired values and none of them is zero.
    
    3/ 01 : next_start_code() *** Was wrong in 24.02.2003; is correct in
    dev-api-4!!! *** was a bug in fact... when we write user data, we pad to
    the next byte boundary (if needed) like the standard says... by
    chancepadding was almost always done, thus the next_start_code() was
    respected.
    
    4/ The extra stuffing bits were caused by a forced padding between our
    VOL function writer and VOP header function writer.

    modified files:
     ./src/bitstream/bitstream.c ./src/encoder.c


2003-07-16 22:57:44 GMT                                  patch-8

    Summary:
      Fixed quant4_intra_xmm and quant_intra_3dne bug for DC<0.
    Revision:
      xvidcore--devapi4--1.0--patch-8

    These two functions were suffering the same error that consists in
    emulating idiv with some an inversed divisor array and an imul
    instruction followed by a right shift...
    
    That was always decreasing the ressult by 1 for negative DC values.
    A not so bad solution is simply to use a cmov instruction and choose
    the right value according to the DC value. As these function were
    for PIII and Athlon, we are sure we can use the cmov instruction.
    
    PS: the fix is somewhere in cosmetic changes... sorry but the code was
    too ugly to fix it like it was.

    modified files:
     ./src/quant/x86_asm/quantize4_xmm.asm
     ./src/quant/x86_asm/quantize_3dne.asm


2003-07-16 12:58:21 GMT                                  patch-7

    Summary:
      Fixed the build system (error caused by patch-5)
    Revision:
      xvidcore--devapi4--1.0--patch-7

    IA64 cahnges were wrong in the build system, they make all platforms
    try to compile a directory... That patch should fix the IA64 target
    build and get back to previous behavior for other architectures.

    modified files:
     ./build/generic/configure.in ./build/generic/platform.inc.in
     ./build/generic/sources.inc


2003-07-13 12:16:55 GMT                                  patch-6

    Summary:
      Updates for GME and some cleanups.
    Revision:
      xvidcore--devapi4--1.0--patch-6

    This is a all in one patch from syskin:
    * mcsel decision moved to ModeDecision() function. That makes motion
      loop completely aware of macroblock mode and vectors (amv in that case).
      A simple copy&paste was needed to make the the decision R-D based, to be
      compatible with MODEDECISION_BITS.
    
    * many bugs fixed. Most of them very small,  the only big one was that
      BITS was misunderstanding a flag and was thinking that mpeg quant is
      used when h263 quant is used and vice versa :( Also, correct cbp with
      inter4v mode makes mode decision better. Two speedups - for BITS (no
      more dequantization when sum == 0) and for ChromaME (chroma sad not
      computed if total sad too big before that). Some GMC compiler warnings
      removed. Probably more, I don't remember ;) I haven't touched P/B/I
      decision for once.
    
    * compiler warnings removed, mostly "const mismatch" in get_amv() <-- or
      what was his name.
    
    * two functions made 2x smaller, shorter and faster.

    modified files:
     ./src/encoder.c ./src/global.h ./src/image/interpolate8x8.h
     ./src/motion/gmc.c ./src/motion/gmc.h
     ./src/motion/motion_est.c ./src/motion/motion_est.h
     ./vfw/src/codec.c


2003-07-10 17:35:59 GMT                                  patch-5

    Summary:
      IA64 updates.
    Revision:
      xvidcore--devapi4--1.0--patch-5

    Changes from Stephan Krause <s_kraste at ira.uka.de>
    
    Small updates so ia64 is supposed to work. Further testing is needed
    because tests have only been done with xvid_encraw.

    modified files:
     ./build/generic/platform.inc.in ./examples/xvid_encraw.c
     ./src/motion/motion_est.c ./src/xvid.c


2003-07-10 17:27:01 GMT                                  patch-4

    Summary:
      Removed remaining expanded $ lines from the arch repo
    Revision:
      xvidcore--devapi4--1.0--patch-4

    Removed remaining expanded $ lines from the arch repo
    

    modified files:
     ./CodingStyle ./build/generic/bootstrap.sh
     ./doc/xvid-encoder.txt ./src/bitstream/ppc_asm/cbp_altivec.s
     ./src/bitstream/ppc_asm/cbp_ppc.s
     ./src/dct/x86_asm/fdct_xmm.asm
     ./src/image/x86_asm/colorspace_yuv_mmx.asm
     ./src/image/x86_asm/reduced_mmx.asm
     ./src/image/x86_asm/yuv_to_yv12_mmx.asm
     ./src/image/x86_asm/yv12_to_rgb24_mmx.asm
     ./src/image/x86_asm/yv12_to_rgb32_mmx.asm
     ./src/motion/ppc_asm/sad_altivec.c ./todo.txt


2003-07-02 23:20:39 GMT                                  patch-3

    Summary:
      Reset the IFrame counter when an iframe is encoded
    Revision:
      xvidcore--devapi4--1.0--patch-3

    Reset the IFrame counter when an iframe is encoded
    

    modified files:
     ./src/encoder.c


2003-06-29 21:58:24 GMT                                  patch-2

    Summary:
      Added 3 warp point GMC.
    Revision:
      xvidcore--devapi4--1.0--patch-2

    Added 3 warp point GMC (first cvs commit + bitstream warp writing fix
    from cvs)

    new files:
     ./src/motion/.arch-ids/gmc.c.id
     ./src/motion/.arch-ids/gmc.h.id ./src/motion/gmc.c
     ./src/motion/gmc.h

    modified files:
     ./build/generic/sources.inc ./src/bitstream/bitstream.c
     ./src/decoder.c ./src/decoder.h ./src/encoder.c
     ./src/encoder.h ./src/global.h ./src/motion/motion.h
     ./src/motion/motion_comp.c ./src/motion/motion_est.c
     ./src/motion/motion_est.h ./src/utils/mbfunctions.h
     ./src/xvid.h


2003-06-29 21:35:01 GMT                                  patch-1

    Summary:
      Updated changelog
    Revision:
      xvidcore--devapi4--1.0--patch-1

    Updated changelog
    

    modified files:
     ./changelog.txt


2003-06-27 17:01:46 GMT                                  base-0

    Summary:
      tag of ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-80
    Revision:
      xvidcore--devapi4--1.0--base-0

    (automatically generated log message)

    new patches:
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--base-0
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-1
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-2
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-3
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-4
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-5
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-6
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-7
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-8
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-9
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-10
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-11
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-12
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-13
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-14
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-15
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-16
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-17
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-18
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-19
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-20
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-21
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-22
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-23
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-24
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-25
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-26
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-27
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-28
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-29
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-30
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-31
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-32
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-33
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-34
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-35
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-36
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-37
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-38
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-39
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-40
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-41
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-42
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-43
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-44
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-45
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-46
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-47
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-48
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-49
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-50
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-51
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-52
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-53
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-54
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-55
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-56
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-57
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-58
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-59
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-60
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-61
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-62
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-63
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-64
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-65
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-66
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-67
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-68
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-69
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-70
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-71
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-72
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-73
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-74
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-75
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-76
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-77
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-78
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-79
     ed.gomez@free.fr--main/xvidcore--devapi4--1.0--patch-80
     ed.gomez@free.fr--main/xvidcore--stable--0.9--base-0
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-1
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-2
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-3
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-4
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-5
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-6
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-7
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-8
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-9
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-10
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-11
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-12
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-13
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-14
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-15
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-16
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-17
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-18
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-19
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-20
     ed.gomez@free.fr--main/xvidcore--stable--0.9--version-0
     ed.gomez@free.fr--main/xvidcore--stable--1.0--base-0
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-1
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-2
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-3
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-4
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-5
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-6


2003-06-27 13:42:52 GMT                                  patch-80

    Summary:
      Still fixes and improvements to motion estimation
    Revision:
      xvidcore--devapi4--1.0--patch-80

    Still fixes and improvements to motion estimation.

    modified files:
     src/motion/motion_est.c src/motion/motion_est.h


2003-06-27 13:35:20 GMT                                  patch-79

    Summary:
      Added compile time PNM reading
    Revision:
      xvidcore--devapi4--1.0--patch-79

    Added compile time PNM reading. It can be useful to test RGB<->YV12
    conversions inside XviD.

    modified files:
     examples/xvid_encraw.c


2003-06-24 12:19:01 GMT                                  patch-78

    Summary:
      Fixes to the RD ME
    Revision:
      xvidcore--devapi4--1.0--patch-78

    Fixes to the RD ME.

    modified files:
     src/motion/motion_est.c src/motion/motion_est.h


2003-06-14 09:14:11 GMT                                  patch-77

    Summary:
      Zone update.
    Revision:
      xvidcore--devapi4--1.0--patch-77

    Removed zone warning boxes (they are counter productive)
    Added zone-based force key frame option.

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/config.rc vfw/src/resource.h


2003-06-14 09:06:37 GMT                                  patch-76

    Summary:
      Fixes a bug where type was not respected in a BEFORE plugin.
    Revision:
      xvidcore--devapi4--1.0--patch-76

    Fixes a bug where type was not respected in a BEFORE plugin.
    
    The plugin framework was not copying what was passed to the
    call_plugins function.

    modified files:
     src/encoder.c


2003-06-12 23:03:38 GMT                                  patch-75

    Summary:
      Fixed the old "yellow line on left" with rgb output.
    Revision:
      xvidcore--devapi4--1.0--patch-75

    Fixed the old "yellow line on left" with rgb output.

    modified files:
     src/image/x86_asm/colorspace_rgb_mmx.asm


2003-06-12 23:02:10 GMT                                  patch-74

    Summary:
      Removed log2bin ia32 optimization.
    Revision:
      xvidcore--devapi4--1.0--patch-74

    Removed log2bin ia32 optimization.

    modified files:
     src/bitstream/bitstream.c


2003-06-12 22:55:10 GMT                                  patch-73

    Summary:
      Fixed some small things in encoder.
    Revision:
      xvidcore--devapi4--1.0--patch-73

    Removed definitively the Hint stuff.
    
    Fixed some XXX thingies nad some cleanup.

    modified files:
     src/encoder.c


2003-06-12 22:51:55 GMT                                  patch-72

    Summary:
      Back to Walken's Idct
    Revision:
      xvidcore--devapi4--1.0--patch-72

    The simple_idct idea was not so good as is. Waiting for a better
    solution from michael.

    modified files:
     src/xvid.c src/xvid.h


2003-06-10 22:45:57 GMT                                  patch-71

    Summary:
      VFW front end update (New live quant histogram window)
    Revision:
      xvidcore--devapi4--1.0--patch-71

    Update to the VFW frontend. It includes a new window that shows
    live quantizer histogram during encoding session.

    new files:
     vfw/src/.arch-ids/status.c.id vfw/src/.arch-ids/status.h.id
     vfw/src/status.c vfw/src/status.h

    modified files:
     vfw/src/codec.c vfw/src/codec.h vfw/src/config.c
     vfw/src/config.h vfw/src/config.rc vfw/src/driverproc.c
     vfw/src/resource.h vfw/vfw.dsp


2003-06-10 20:58:09 GMT                                  patch-70

    Summary:
      Fixed the win32 project file because of patch-64
    Revision:
      xvidcore--devapi4--1.0--patch-70

    Removed adapt_quant.[ch] files from the libxvidcore win32 project
    file.

    modified files:
     build/win32/libxvidcore.dsp


2003-06-10 20:53:31 GMT                                  patch-69

    Summary:
      Added direct target frame size support + cosmetic.
    Revision:
      xvidcore--devapi4--1.0--patch-69

    If the target bitrate is < 0, it is now interpreted as a target
    size in kbytes.
    
    I did also some cosmetic work to remove all space indents ^_^.

    modified files:
     src/plugins/plugin_2pass2.c


2003-06-10 09:13:40 GMT                                  patch-68

    Summary:
      xvid_bench updates and corresponding Makefile changes.
    Revision:
      xvidcore--devapi4--1.0--patch-68

    As mentionned on the devel mailing list, xvid_bench did not even
    compile anymore.
    
    This patch updates xvid_bench to the new API. xvid_bench is now
    compiled with other examples by the Makefile, this makes mandatory
    to include ../build/generic/platform.inc to have the ARCH_IS_xxxx
    constants. Dunno if it has an impact on Win32 project files.

    modified files:
     examples/Makefile examples/xvid_bench.c


2003-06-10 09:05:14 GMT                                  patch-67

    Summary:
      Probably a small copy/paste error
    Revision:
      xvidcore--devapi4--1.0--patch-67

    XVID_CSP_BGR was advertised as being a 32bit packed format -> 24bit
    is the right pixel size

    modified files:
     src/xvid.h


2003-06-09 19:39:47 GMT                                  patch-66

    Summary:
      Activated simple_idct_mmx.
    Revision:
      xvidcore--devapi4--1.0--patch-66

    This patch activates simple_idct_mmx use. However it tries to
    make sure old streams (< version 10) are decoded using the mmx
    Walten's version.
    
    A noticeable bitstream version change, it is now numbered 11.
    The number 10 is used on the cvs_head version for the same code
    change.

    modified files:
     src/bitstream/bitstream.c src/dct/simple_idct.c
     src/dct/x86_asm/simple_idct_mmx.asm src/decoder.c
     src/decoder.h src/xvid.c src/xvid.h


2003-06-09 19:15:18 GMT                                  patch-65

    Summary:
      Remaining include of adapt_quant.h
    Revision:
      xvidcore--devapi4--1.0--patch-65

    encoder.c was still including adapt_quant.h. Removed.

    modified files:
     src/encoder.c


2003-06-09 17:49:44 GMT                                  patch-64

    Summary:
      Moved code from adapt_quant.c to the lumimasking plugin.
    Revision:
      xvidcore--devapi4--1.0--patch-64

    The lumimasking plugin was using functions from outside. As I understand
    what plugins are, they should not rely on code outside their module as
    much as it is possible to achieve.
    
    Here it was clear, the plugin could be made standalone.
    
    PS: it seems lumimasking is a no-op plugin, it's probably a bug in the
        plugin framework. No time to track this.

    removed files:
     src/quant/.arch-ids/adapt_quant.c.id
     src/quant/.arch-ids/adapt_quant.h.id src/quant/adapt_quant.c
     src/quant/adapt_quant.h

    modified files:
     build/generic/sources.inc src/plugins/plugin_lumimasking.c


2003-06-09 13:45:29 GMT                                  patch-63

    Summary:
      Legal GNU GPL Headers and copyright holders.
    Revision:
      xvidcore--devapi4--1.0--patch-63

    Added Legal GNU GPL headers and copyright holders as defined in
    XviD 0.9.x.
    
    There are still some wrong copyright (atm noted 'Anonymous') and i
    missed probably some old headers that contain the GNU GPL pattern my
    script searched for.

    modified files:
     dshow/src/CAbout.cpp dshow/src/CAbout.h
     dshow/src/CXvidDecoder.cpp dshow/src/CXvidDecoder.h
     dshow/src/IXvidDecoder.h examples/xvid_bench.c
     examples/xvid_decraw.c examples/xvid_encraw.c
     examples/xvid_stat.c rawdec/rawdec.c src/bitstream/bitstream.c
     src/bitstream/bitstream.h src/bitstream/cbp.c
     src/bitstream/cbp.h src/bitstream/mbcoding.c
     src/bitstream/mbcoding.h src/bitstream/vlc_codes.h
     src/bitstream/zigzag.h src/dct/fdct.c src/dct/fdct.h
     src/dct/idct.c src/dct/idct.h src/dct/simple_idct.c
     src/decoder.c src/decoder.h src/divx4.c src/divx4.h
     src/encoder.c src/encoder.h src/global.h
     src/image/colorspace.c src/image/colorspace.h src/image/font.c
     src/image/font.h src/image/image.c src/image/image.h
     src/image/interpolate8x8.c src/image/interpolate8x8.h
     src/image/reduced.c src/image/reduced.h src/motion/motion.h
     src/motion/motion_comp.c src/motion/motion_est.c
     src/motion/motion_est.h src/motion/ppc_asm/sad_altivec.c
     src/motion/sad.c src/motion/sad.h src/motion/smp_motion_est.c
     src/motion/smp_motion_est.h src/plugins/plugin_2pass1.c
     src/plugins/plugin_2pass2.c src/plugins/plugin_dump.c
     src/plugins/plugin_fixed.c src/plugins/plugin_lumimasking.c
     src/plugins/plugin_psnr.c src/plugins/plugin_single.c
     src/portab.h src/prediction/mbprediction.h
     src/quant/adapt_quant.c src/quant/adapt_quant.h
     src/quant/quant_h263.h src/quant/quant_matrix.c
     src/quant/quant_matrix.h src/quant/quant_mpeg4.c
     src/quant/quant_mpeg4.h src/utils/emms.c src/utils/emms.h
     src/utils/mbfunctions.h src/utils/mbtransquant.c
     src/utils/mem_align.c src/utils/mem_align.h
     src/utils/mem_transfer.c src/utils/mem_transfer.h
     src/utils/timer.c src/utils/timer.h src/xvid.c src/xvid.h
     vfw/src/2pass.h vfw/src/codec.h vfw/src/config.h
     vfw/src/debug.h vfw/src/resource.h vfw/src/vfwext.h


2003-06-09 01:13:50 GMT                                  patch-62

    Summary:
      ANSI C comments.
    Revision:
      xvidcore--devapi4--1.0--patch-62

    Turned all // ISO C99 comments into ISO C89 (aka ANSI C) coment style.
    
    Now XviD compiles fine with gcc 3.x -std=iso89 option. This should help
    those people who want to get XviD working on DSPs or any other exotic
    hardware. This type of exotic hardware is usually shipped with a very
    spartiate ANSI C compiler.
    
    NB: Big patch that breaks all kind of cherry picking merges.

    modified files:
     examples/odivx_enc_dec.c examples/xvid_bench.c
     src/bitstream/bitstream.c src/bitstream/mbcoding.c
     src/bitstream/mbcoding.h src/bitstream/zigzag.h src/dct/fdct.c
     src/dct/idct.c src/dct/simple_idct.c src/decoder.c
     src/decoder.h src/encoder.c src/encoder.h src/global.h
     src/image/colorspace.c src/image/font.c src/image/image.c
     src/image/interpolate8x8.c src/image/interpolate8x8.h
     src/image/reduced.c src/motion/motion.h
     src/motion/motion_comp.c src/motion/motion_est.c
     src/motion/motion_est.h src/motion/sad.c
     src/plugins/plugin_2pass2.c src/prediction/mbprediction.c
     src/prediction/mbprediction.h src/quant/adapt_quant.c
     src/quant/adapt_quant.h src/quant/quant_h263.c
     src/quant/quant_h263.h src/quant/quant_mpeg4.c
     src/quant/quant_mpeg4.h src/utils/mbtransquant.c
     src/utils/mem_transfer.c src/utils/timer.c src/xvid.c
     vfw/src/2pass.c vfw/src/codec.c vfw/src/codec.h
     vfw/src/config.c vfw/src/config.h vfw/src/driverproc.c


2003-06-04 18:19:56 GMT                                  patch-61

    Summary:
      Removed AltCC from VFW frontend
    Revision:
      xvidcore--devapi4--1.0--patch-61

    A previous patch removed AltCC from the 2pass plugin. Thus we
    remove the frontend panels for AltCC and corresponding code.

    modified files:
     src/xvid.h vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/config.rc


2003-05-29 14:47:28 GMT                                  patch-60

    Summary:
      Lot of two pass updates.
    Revision:
      xvidcore--devapi4--1.0--patch-60

    * Removed Alt curve treatment
    
    * After reading VFW code, i found out that it was using the blocks stats
      fields  to  retrieve  the number  of  MBs  in  a  frame. So  all  this
      min_size[]  was  not meant  to  discover  a  min_size for  each  frame
      according to  its intra  MBs but rather  an hardcoded minimum  for all
      frames as advertised in earlier  cvs revisions.  It would be easier if
      original code was commented :-(
    
    * Some comment changes
    * bquant_error and pquant_error have been replaced by an array
      quant_error[3][32] indexed by frame type and quantizer value.
    * Moved some initialization stuff
    
    * I read VFW and noticed that min_length was supposed to be:
      min{hard coded length, min{observed lengths}}
    
    * Force frame type during the second pass.
    
    * Simplified equations.
     
      Scaling was needed because of the non linear formulas used in AltCC but
      now we can directly use avg_length[s->type-1] instead of "first prescaling
      bframes to pframes lengths then use pframe stats and at last prescaling back
      frame length to bframe lengths"
     
      See my new XXX: question about the overflow.

    modified files:
     src/plugins/plugin_2pass2.c


2003-05-25 10:01:55 GMT                                  patch-59

    Summary:
      Function reordering, fix minimum "hardcoded" frame sizes in internal_sacle().
    Revision:
      xvidcore--devapi4--1.0--patch-59

    Fixed a bug where hardcoded miminum frame lengths were computed only for
    first frame (IFrame) and was applied for al frames. I just moved the formulas
    into the frame loop.
    
    Lot of cosmetic work, function reodrering etc etc so the plugin function
    come first, and then we have sub function and helper functions. Some fixes
    in my previous comments.

    modified files:
     src/plugins/plugin_2pass2.c


2003-05-22 23:11:21 GMT                                  patch-58

    Summary:
      Added the container_frame_overhead field to the 2pass2 RC structure.
    Revision:
      xvidcore--devapi4--1.0--patch-58

    In my previous patches, i disabled container format overhead compensation
    because xvidcore can be used for other things than AVI. However this
    compensation is usefull, so it's back with its own structure field that
    specifies how much bytes the container uses for a frame (average value).
    
    We can now do some direct ogm, matroska encodings without loosing a single
    byte... :-)

    modified files:
     src/plugins/plugin_2pass2.c src/xvid.h vfw/src/codec.c


2003-05-22 22:22:47 GMT                                  patch-57

    Summary:
      Fixed an overflow bug in target filesize computation.
    Revision:
      xvidcore--devapi4--1.0--patch-57

    rc->target was an uint64_t data to avoid overflow when dealing
    with long movies and/or high bitrates. The problem is that its
    initialization was using int32 data, thus this was resulting
    in an overflow in its initial computation. Quite silly, but this
    bug drived me crazy during 4 hours...

    modified files:
     src/plugins/plugin_2pass2.c


2003-05-22 18:53:19 GMT                                  patch-56

    Summary:
      Added the mrproper Makefile target.
    Revision:
      xvidcore--devapi4--1.0--patch-56

    Added the mrproper Makefile target that deletes even bootstrapped
    files. mrproper name comes from the linux kernel makefile, i was
    out of inspiration.

    modified files:
     build/generic/Makefile


2003-05-22 17:30:15 GMT                                  patch-55

    Summary:
      Fix a nasty bug due to a typo mistake.
    Revision:
      xvidcore--devapi4--1.0--patch-55

    We were comparing frame length with a wrong min_size[index]
    that was out of bounds (in internal_scale).

    modified files:
     src/plugins/plugin_2pass2.c


2003-05-22 17:24:19 GMT                                  patch-54

    Summary:
      Removed automatic \n in DPRINTF calls.
    Revision:
      xvidcore--devapi4--1.0--patch-54

    Removed automatic \n in DPRINTF calls.

    modified files:
     src/bitstream/bitstream.c src/bitstream/mbcoding.c
     src/decoder.c src/encoder.c src/image/image.c
     src/plugins/plugin_2pass2.c src/portab.h
     src/prediction/mbprediction.c


2003-05-22 17:03:38 GMT                                  patch-53

    Summary:
      Cleaned up a bit, added comments.
    Revision:
      xvidcore--devapi4--1.0--patch-53

    I Cleaned up the plugin_before function. I added some comments at the
    same time, so now it should be more easy to understand the meaning of
    all these if/else thingies :-)

    modified files:
     src/plugins/plugin_2pass2.c


2003-05-18 12:12:49 GMT                                  patch-52

    Summary:
      Update of xvid_encraw (vop_debug, debug, max key frame)
    Revision:
      xvidcore--devapi4--1.0--patch-52

    Added a -vop_debug option. This makes xvidcore to print out frame
    information directly into the encoded frame.
    
    Changed the meaning of the -debug option. It activates now the
    internal xvidcore debug output.
    
    Added a -max_key_interval.

    modified files:
     examples/xvid_encraw.c


2003-05-18 12:01:31 GMT                                  patch-51

    Summary:
      Missing RateControl removal from Win32 visual project.
    Revision:
      xvidcore--devapi4--1.0--patch-51

    RateControl removal was missing in the visual c project.

    modified files:
     build/win32/libxvidcore.dsp


2003-05-18 00:08:46 GMT                                  patch-50

    Summary:
      Removed legacy RateControl module.
    Revision:
      xvidcore--devapi4--1.0--patch-50

    Removed all code related to the old RateControl module.

    removed files:
     src/utils/.arch-ids/ratecontrol.h.id
     src/utils/.arch-ids/ratecontrol.c.id src/utils/ratecontrol.h
     src/utils/ratecontrol.c

    modified files:
     build/generic/sources.inc src/encoder.h


2003-05-17 23:54:55 GMT                                  patch-49

    Summary:
      VFW Update.
    Revision:
      xvidcore--devapi4--1.0--patch-49

    Added support for the debug option. The registry key debug has been
    changed to vop_debug. The reg key debug is now used for the codec
    debugging output.
    
    Some work on zones and mispellings.

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/config.rc vfw/src/resource.h


2003-05-17 23:50:38 GMT                                  patch-48

    Summary:
      2pass plugin updates for zone support.
    Revision:
      xvidcore--devapi4--1.0--patch-48

    A bit more work on zones support in the 2Pass2 plugin.
    
    Simple cleanup in the 2Pass1 plugin.

    modified files:
     src/plugins/plugin_2pass1.c src/plugins/plugin_2pass2.c


2003-05-17 21:07:43 GMT                                  patch-47

    Summary:
      Debug is now controled through xvid_global + INIT
    Revision:
      xvidcore--devapi4--1.0--patch-47

    Debug is now controlled through a global variable. It can be set thanx
    to the API using the new xvid_gbl_init_t.debug field.
    
    All DPRINTF constants have been turned into XVID_DEBUG_xxxx. They have
    been moved to xvid.h.

    modified files:
     src/bitstream/bitstream.c src/bitstream/mbcoding.c
     src/decoder.c src/encoder.c src/image/image.c
     src/plugins/plugin_2pass2.c src/plugins/plugin_single.c
     src/portab.h src/prediction/mbprediction.c
     src/utils/ratecontrol.c src/xvid.c src/xvid.h


2003-05-17 20:32:59 GMT                                  patch-46

    Summary:
      Fix for the patch-44.
    Revision:
      xvidcore--devapi4--1.0--patch-46

    The fix from pete was breaking lot of other stuff, or at least it was
    outlining it.
    
    Now it works reliably.

    modified files:
     src/encoder.c


2003-05-15 17:31:04 GMT                                  patch-45

    Summary:
      Removed XVID_VOP_DYNAMIC_BFRAMES flag.
    Revision:
      xvidcore--devapi4--1.0--patch-45

    The encoder loop bugfix removed this flag, so it's now being removed
    from xvid_encraw.

    modified files:
     examples/xvid_encraw.c


2003-05-15 17:24:55 GMT                                  patch-44

    Summary:
      Fix to the encoder loop (was not respecting dynamic decision).
    Revision:
      xvidcore--devapi4--1.0--patch-44

    The long awaited fix to the encoder loop that was not respecting the
    dynamic decision performed by the MEAnlyse function.

    modified files:
     src/encoder.c src/xvid.h


2003-05-14 23:27:59 GMT                                  patch-43

    Summary:
      Added module building for MacOSX.
    Revision:
      xvidcore--devapi4--1.0--patch-43

    Added the --enable-macosx_module option to the configure script.
    It allows module building on that platform as it differenciates
    loadable modules (a la dlopen) and dynamic libs that are simply
    linked at compile time.
    
    This was needed for transcode.
    
    Patch contributed by Tilmann Bitterberg <transcode at tibit.org>

    modified files:
     build/generic/configure.in


2003-05-14 20:21:30 GMT                                  patch-42

    Summary:
      Merged RD ME from cvs_head.
    Revision:
      xvidcore--devapi4--1.0--patch-42

    Syskin has changed a bit the ME algorithm, so now it does a kind of
    RD optimization of Vector search.

    modified files:
     src/motion/motion_est.c src/motion/motion_est.h


2003-05-14 18:40:40 GMT                                  patch-41

    Summary:
      Merged syskin ME changes.
    Revision:
      xvidcore--devapi4--1.0--patch-41

    Merged last syskin ME changes. Matches motion_est.c:1.69 and motion_est.h:1.7
    minus unneeded code plus some changes due to new API.

    modified files:
     src/motion/motion_est.c src/motion/motion_est.h


2003-05-14 17:28:52 GMT                                  patch-40

    Summary:
      Small update to xvid_encraw.
    Revision:
      xvidcore--devapi4--1.0--patch-40

    I added an help message to mention the fact we can repeat the zone options.

    modified files:
     examples/xvid_encraw.c


2003-05-14 14:19:12 GMT                                  patch-39

    Summary:
      VFW Update (zone support, profile support)
    Revision:
      xvidcore--devapi4--1.0--patch-39

    VFW Update (zone support, profile support)

    new files:
     vfw/src/.arch-ids/vfwext.h.id vfw/src/.arch-ids/debug.h.id
     vfw/src/vfwext.h vfw/src/debug.h

    modified files:
     vfw/src/codec.c vfw/src/config.c vfw/src/config.h
     vfw/src/config.rc vfw/src/driverproc.c vfw/src/driverproc.def
     vfw/src/resource.h vfw/vfw.dsp


2003-05-14 14:02:05 GMT                                  patch-38

    Summary:
      Add support for single RC and zones to xvid_encraw.
    Revision:
      xvidcore--devapi4--1.0--patch-38

    Adds support for single RC and zones to xvid_encraw.

    modified files:
     examples/xvid_encraw.c


2003-05-14 13:58:56 GMT                                  patch-37

    Summary:
      Fixes for Win32 build of libxvidcore.
    Revision:
      xvidcore--devapi4--1.0--patch-37

    A previous patch left the Win32 build process incomplete and not up
    to date.

    modified files:
     build/generic/libxvidcore.def.in build/win32/libxvidcore.dsp


2003-05-13 00:10:12 GMT                                  patch-36

    Summary:
      Small fixes.
    Revision:
      xvidcore--devapi4--1.0--patch-36

    data->quant fix.
    
    Fixed some coding bugs in trellis code. Used __inline and not inline.

    modified files:
     src/encoder.c src/utils/mbtransquant.c


2003-05-13 00:05:03 GMT                                  patch-35

    Summary:
      CBR plugin is renamed Single pass. Fixed Quant plugin is disabled.
    Revision:
      xvidcore--devapi4--1.0--patch-35

    With the zones feature, the CBR plugin could be used for all type
    of one pass RC. The better thing to do would be to include fixed
    quant to this new single pass plugin.
    
    Btw, a (clean) solution has not been found yet. I am obliged to
    disable the fixed quant plugin. This breaks xvid_encraw :-(

    modified files:
     build/generic/sources.inc src/plugins/plugin_single.c
     src/xvid.h

    renamed files:
     src/plugins/.arch-ids/plugin_cbr.c.id
       ==> src/plugins/.arch-ids/plugin_single.c.id
     src/plugins/plugin_cbr.c
       ==> src/plugins/plugin_single.c


2003-05-12 23:49:14 GMT                                  patch-34

    Summary:
      Removed quant limits per RC plugin, moved to global settings.
    Revision:
      xvidcore--devapi4--1.0--patch-34

    The I/P/B Frames' min/max quantizers have moved from RC plugins' interface
    to the general encoding interface.
    
    The CBR plugin has been updated for zones and the quent limits move.

    modified files:
     src/encoder.c src/encoder.h src/plugins/plugin_2pass2.c
     src/plugins/plugin_cbr.c src/xvid.h


2003-05-12 23:25:54 GMT                                  patch-33

    Summary:
      Added encoding zones
    Revision:
      xvidcore--devapi4--1.0--patch-33

    Added encoding zones in 2pass plugins. The idea behind "zones" is
    to define frame ranges for which we change the plugin's behavior.

    modified files:
     src/encoder.c src/encoder.h src/plugins/plugin_2pass1.c
     src/xvid.h


2003-05-12 23:10:17 GMT                                  patch-32

    Summary:
      Added the profile setting.
    Revision:
      xvidcore--devapi4--1.0--patch-32

    Added the profile setting to user API.

    modified files:
     src/bitstream/bitstream.c src/bitstream/bitstream.h
     src/encoder.c src/encoder.h src/xvid.h


2003-05-11 23:59:01 GMT                                  patch-31

    Summary:
      Changed quality presets.
    Revision:
      xvidcore--devapi4--1.0--patch-31

    The presets have been changed so now we should have better PSNR with
    higher quality presets in all cases.
    
    I changed a bit the way we treat quality overflow or overflow, now i
    just clip the value to allowed range.

    modified files:
     examples/xvid_encraw.c


2003-05-11 20:47:55 GMT                                  patch-30

    Summary:
      Some cleanups in the trellis code.
    Revision:
      xvidcore--devapi4--1.0--patch-30

    Some cleanup work on trellis code. Should compile file on Visual C++ now.

    modified files:
     src/utils/mbtransquant.c


2003-05-10 23:53:28 GMT                                  patch-29

    Summary:
      New trellis code
    Revision:
      xvidcore--devapi4--1.0--patch-29

    New trellis code from skal. It should be reworked a bit so it integrates
    better into XviD code.

    modified files:
     src/bitstream/mbcoding.c src/bitstream/mbcoding.h
     src/utils/mbtransquant.c


2003-05-10 23:43:11 GMT                                  patch-28

    Summary:
      Intra frame decision.
    Revision:
      xvidcore--devapi4--1.0--patch-28

    These are syskin's words:  hopefully, intra frame at every scene
    change (we really need it to cut things).
    

    modified files:
     src/motion/motion_est.c


2003-05-05 21:50:25 GMT                                  patch-27

    Summary:
      Fixed double last calculation in trellis quantization.
    Revision:
      xvidcore--devapi4--1.0--patch-27

    chl changelog message:  Removed double calculation of "last" => +0.02dB

    modified files:
     src/bitstream/mbcoding.c


2003-05-05 21:46:29 GMT                                  patch-26

    Summary:
      Added config.status to the distclean target.
    Revision:
      xvidcore--devapi4--1.0--patch-26

    The target distclean is supposed to clean all files so the remaining
    ones are those supposed to be found in a distribution tarball...
    config.status is not one of them, so let's add this to the distclean
    target.

    modified files:
     build/generic/Makefile


2003-05-05 21:39:47 GMT                                  patch-25

    Summary:
      configure.in tuning (API number and lib sonames).
    Revision:
      xvidcore--devapi4--1.0--patch-25

    I fixed a typo which prevented SPECIFIC_CFLAGS to be properly set by
    the configure script.
    
    I also bumped the API version number as API 3.0 is current cvs_head
    and this branch is the next major API version
    
    While trying to build my own debian package out of xvidcore, i ran into
    trouble with the soname not respecting some basic rules that prevented
    having different library revisions running alongside (with different major
    APIs). This has been fixed adding the major API number to the library
    SONAME.

    modified files:
     build/generic/configure.in


2003-04-27 23:22:30 GMT                                  patch-24

    Summary:
      Cleaned CBR plugin a bit, adds structure for a better initial quant.
    Revision:
      xvidcore--devapi4--1.0--patch-24

    Just a clean up turning default values to preprocessor constants. I added
    a get_initial_quant for trying to retrieve support in a near future, a good
    quantizer according to the desired target bitrate. This will be done thanks
    to a simple LUT where we'll have lut[quant] = average_bitrate;. This seems
    stupid but it'll be better than starting with an hardcoded value.

    modified files:
     src/plugins/plugin_cbr.c


2003-04-27 23:18:20 GMT                                  patch-23

    Summary:
      b-frames look good in still motion, after all.
    Revision:
      xvidcore--devapi4--1.0--patch-23

    b-frames look good in still motion, after all.

    modified files:
     src/motion/motion_est.c


2003-04-27 23:14:39 GMT                                  patch-22

    Summary:
      Add initial trellis quantization to inter+h263 frames.
    Revision:
      xvidcore--devapi4--1.0--patch-22

    This is the initial support of trellis quantization for inter frames
    + h263 quantization method.
    
    Complete support is on the way.

    modified files:
     examples/xvid_encraw.c src/bitstream/mbcoding.c
     src/bitstream/mbcoding.h src/utils/mbtransquant.c src/xvid.h


2003-04-27 22:59:47 GMT                                  patch-21

    Summary:
      Fixes 2 memory leaks.
    Revision:
      xvidcore--devapi4--1.0--patch-21

    After a valgrind pass I fixed these 2 leaks. We have still to fix
    an MEAnalysis on unitialized data.

    modified files:
     src/encoder.c src/utils/mem_align.c


2003-04-27 22:50:27 GMT                                  patch-20

    Summary:
      Adds Avg PSNR output to xvid_encraw.
    Revision:
      xvidcore--devapi4--1.0--patch-20

    Adds Avg PSNR output to xvid_encraw.

    modified files:
     examples/xvid_encraw.c


2003-04-27 22:40:45 GMT                                  patch-19

    Summary:
      Fixes the vfw Visual Project.
    Revision:
      xvidcore--devapi4--1.0--patch-19

    A missing file has been removed from the project file.

    modified files:
     vfw/vfw.dsp


2003-04-14 20:07:47 GMT                                  patch-18

    Summary:
      Fixes plugin initialization in xvid_encraw.
    Revision:
      xvidcore--devapi4--1.0--patch-18

    We were initializing plugins' versions before a memset...
    <homer>Doh...</homer>

    modified files:
     examples/xvid_encraw.c


2003-04-14 15:28:57 GMT                                  patch-17

    Summary:
      Fixed function prototypes <-> definitions mismatching.
    Revision:
      xvidcore--devapi4--1.0--patch-17

    Fixed function prototypes <-> definitions mismatching.

    modified files:
     src/utils/mbfunctions.h src/utils/mbtransquant.c


2003-04-14 15:23:15 GMT                                  patch-16

    Summary:
      VFW frontend update
    Revision:
      xvidcore--devapi4--1.0--patch-16

    The VFW frontend has been updated.

    modified files:
     vfw/src/codec.c vfw/src/codec.h vfw/src/config.c
     vfw/src/config.h vfw/src/config.rc vfw/src/driverproc.c
     vfw/src/resource.h vfw/vfw.dsp


2003-04-14 15:05:20 GMT                                  patch-15

    Summary:
      Fixed BITS decision for low quants, reworked p/b/i decision.
    Revision:
      xvidcore--devapi4--1.0--patch-15

    sysKin's log message: improved vhq (does not decrease psnr anymore - at least
                          for low quants) and tweaked p/b/i decision again.
    
    I Added a fix to this CVS commit to avoid ALU Exception (division by zero). It
    has been commited to cvs_head as well by sysKin.

    modified files:
     src/motion/motion_est.c


2003-04-10 13:01:07 GMT                                  patch-14

    Summary:
      Removed all ABS() macros.
    Revision:
      xvidcore--devapi4--1.0--patch-14

    All  ABS   macros  have   been  replace  with   their  stdlib.h/math.h
    equivalent. This gives a 33%  overall speedup for the plain C encoder,
    while the  ia32 one seems to  suffer a small speed  loss. However this
    speed loss is  very small and it  seems it depends on the  CPU type as
    the abs/fabs usage  is impacting badly on sad  functions but impacting
    well on interpolate functions ... weird inst'it ?

    modified files:
     src/bitstream/mbcoding.c src/global.h src/image/image.c
     src/motion/motion_comp.c src/motion/motion_est.c
     src/motion/sad.c src/prediction/mbprediction.c
     src/utils/mbtransquant.c src/xvid.c


2003-04-09 18:44:24 GMT                                  patch-13

    Summary:
      Added GNU profiling option to the configure script.
    Revision:
      xvidcore--devapi4--1.0--patch-13

    In  order to allow  easy profiling  using GNU  tools (gprof,  gcov), I
    added  the  --enable-gnuprofile  to  the configure.in  template.  This
    modifies the SPECIFIC_LDFLAGS and SPECIFIC_CFLAGS for library building
    so they include all needed options for profiling and test coverage.
    
    /!\  When  compiling your  own  program,  don't  forget to  use  these
         options:
           -pg -fprofile-arcs -ftest-coverage
    
         When linking your program, you  MUST use the -pg option too, else
         your binary will not use/create profiling informations.

    modified files:
     build/generic/configure.in


2003-04-09 16:09:33 GMT                                  patch-12

    Summary:
      Build fix from release-0_9_1-fixes@cvs.xvid.org
    Revision:
      xvidcore--devapi4--1.0--patch-12

    MacOSX build process was wrong on the linking stage as it was ignoring
    the equivalent of the linux soname thingy.

    modified files:
     build/generic/configure.in


2003-04-09 13:44:06 GMT                                  patch-11

    Summary:
      Syncing arch tree with xvid.org cvs.
    Revision:
      xvidcore--devapi4--1.0--patch-11

    Synced with all the work done  in the xvid.org cvs repository. I could
    not  maintain a  complete list  of all  items but  here is  a  kind of
    digest.
    
     + Merged build files fixes from the release-0_9_1-fixes branch.
     + Synced all motion estimation changes from the cvs_head branch.
     + Added rate control plugins.
     + Added lumimasking plugin.
     + Synced optimizations from cvs_head in interpolate and cbp
       functions.
     + xvid_encraw improvements.
     + new mbtransquant set of functions.
     + Fixed bframe SSE calculation.

    new files:
     src/plugins/.arch-ids/plugin_2pass1.c.id
     src/plugins/.arch-ids/plugin_2pass2.c.id
     src/plugins/.arch-ids/plugin_cbr.c.id
     src/plugins/.arch-ids/plugin_fixed.c.id
     src/plugins/.arch-ids/plugin_lumimasking.c.id
     vfw/src/.arch-ids/XviD_logo.bmp.id src/plugins/plugin_2pass1.c
     src/plugins/plugin_2pass2.c src/plugins/plugin_cbr.c
     src/plugins/plugin_fixed.c src/plugins/plugin_lumimasking.c
     vfw/src/XviD_logo.bmp

    modified files:
     build/generic/Makefile build/generic/bootstrap.sh
     build/generic/configure.in build/generic/libxvidcore.def.in
     build/generic/platform.inc.in build/generic/sources.inc
     build/win32/libxvidcore.dsp examples/xvid_decraw.c
     examples/xvid_encraw.c examples/xvid_stat.c
     src/bitstream/bitstream.c src/bitstream/bitstream.h
     src/bitstream/cbp.c src/bitstream/mbcoding.c
     src/bitstream/vlc_codes.h src/decoder.c src/divx4.c
     src/encoder.c src/encoder.h src/global.h src/image/image.c
     src/image/interpolate8x8.c src/motion/motion.h
     src/motion/motion_comp.c src/motion/motion_est.c
     src/motion/motion_est.h src/motion/sad.c
     src/plugins/plugin_dump.c src/plugins/plugin_psnr.c
     src/portab.h src/prediction/mbprediction.c
     src/utils/mbfunctions.h src/utils/mbtransquant.c src/xvid.c
     src/xvid.h vfw/src/2pass.c vfw/src/codec.c


2003-03-16 00:21:32 GMT                                  patch-10

    Summary:
      Added suxen plugin system (Synced with CVS)
    Revision:
      xvidcore--devapi4--1.0--patch-10

    Sync with the CVS and thus adds the plugin framework.

    new files:
     src/plugins/.arch-ids/=id
     src/plugins/.arch-ids/plugin_dump.c.id
     src/plugins/.arch-ids/plugin_psnr.c.id
     src/plugins/plugin_dump.c src/plugins/plugin_psnr.c

    modified files:
     build/win32/libxvidcore.dsp examples/Makefile
     examples/xvid_decraw.c examples/xvid_encraw.c
     src/bitstream/bitstream.c src/encoder.c src/encoder.h
     src/portab.h src/utils/mbtransquant.c src/xvid.h

    new directories:
     src/plugins/.arch-ids src/plugins


2003-03-11 23:37:06 GMT                                  patch-9

    Summary:
      Changed xvid_decraw option handling for -d/-m.
    Revision:
      xvidcore--devapi4--1.0--patch-9

    -d and -m options were boolean so option values were not needed.

    modified files:
     examples/xvid_decraw.c


2003-03-11 23:30:16 GMT                                  patch-8

    Summary:
      Fixed frame counting in xvid_encraw.
    Revision:
      xvidcore--devapi4--1.0--patch-8

    We were branching before incrementing  the frame counter when core was
    buffering frames.  This was resulting  in wrong frame  counting during
    the buffering phase.

    modified files:
     examples/xvid_encraw.c


2003-03-11 23:07:01 GMT                                  patch-7

    Summary:
      Ported xvid_decraw to new API.
    Revision:
      xvidcore--devapi4--1.0--patch-7

    xvid_decraw  has been  ported to  new  API. It  basically works  fine,
    however i'm not completly staisfied. If I do a step by step run then i
    can see  that second frame  is reported as  a VOL decoding  though the
    first IFrame has been consumed. This makes xvid_decraw does not report
    correctly frames' length.
    
    Except  that, xvid_decraw  works well  enough  to activate  it in  the
    makefile.

    modified files:
     examples/Makefile examples/xvid_decraw.c


2003-03-11 20:19:44 GMT                                  patch-6

    Summary:
      Fix an important API comment in main header.
    Revision:
      xvidcore--devapi4--1.0--patch-6

    This patch fixes a comment in xvid.h which was simply wrong and
    could lead to uneeded code.

    modified files:
     src/xvid.h


2003-03-11 00:36:34 GMT                                  patch-5

    Summary:
      PSNR is now an option. More consistent -m/-s option handling.
    Revision:
      xvidcore--devapi4--1.0--patch-5

    I turned PSNR stats into an option (-s). I fixed the handling of the
    -m option that required only a bool.
    
    I could say in french "ma stupidité dans toute sa grandeur".
    
    -m presence is enough to significate "save _m_peg stream", we don't
    need the boolean value.

    modified files:
     examples/xvid_encraw.c


2003-03-10 00:36:15 GMT                                  patch-4

    Summary:
      Adds extended stats support even for bframes in xvidcore.
    Revision:
      xvidcore--devapi4--1.0--patch-4

    This patch enables core extended stats support even for bframes. It
    modifies the way the MBTransQuantBVOP function does its work. It used
    to not dequant, idct the MB because bframes are never used as reference
    frames. However if we want to compute stats, then we must perform these
    inverse transformations.

    modified files:
     src/encoder.c src/utils/mbfunctions.h src/utils/mbtransquant.c


2003-03-09 16:42:27 GMT                                  patch-3

    Summary:
      Adds extended stats support.
    Revision:
      xvidcore--devapi4--1.0--patch-3

    This patch enables core extended stats support. It seems that xvidcore
    does not compute sse for BFrames, I have tried to adds this by an ugly
    hack  but  it did  not  work  as expected,  i  suppose  core does  not
    decompress bframes as they are  not used as reference frames (unlike P
    and I frames).
    
    If we  succeed in enabling sse  calculation in core  for bframes, then
    xvid_stat will  not be needed anymore.  This will save  lot of trouble
    with frame matching in PSNR computation when bframes are enabled.

    modified files:
     examples/xvid_encraw.c


2003-03-09 00:23:52 GMT                                  patch-2

    Summary:
      Updated xvid_encraw for new API.
    Revision:
      xvidcore--devapi4--1.0--patch-2

    This patch updates the xvid_encraw example to support the
    new API. As it's the first patch for API 4 support, I
    disabled all other examples.
    
    BUG: first frame type is Unknown, I suppose I'm missing
         a subtility of the new API.

    modified files:
     examples/Makefile examples/xvid_encraw.c src/encoder.c


2003-03-06 22:08:43 GMT                                  patch-1

    Summary:
      Synced with dev-api-4 XviD branch.
    Revision:
      xvidcore--devapi4--1.0--patch-1

    Synced with dev-api-4 XviD branch. My branching was done at a later
    point than CVS. This resulted in version skew, now this branch is
    synced with CVS.

    new files:
     vfw/.arch-ids/=id vfw/bin/.arch-ids/=id vfw/src/.arch-ids/=id
     vfw/.arch-ids/vfw.dsp.id vfw/bin/.arch-ids/xvid.inf.id
     vfw/src/.arch-ids/2pass.c.id vfw/src/.arch-ids/2pass.h.id
     vfw/src/.arch-ids/codec.c.id vfw/src/.arch-ids/codec.h.id
     vfw/src/.arch-ids/config.c.id vfw/src/.arch-ids/config.h.id
     vfw/src/.arch-ids/config.rc.id
     vfw/src/.arch-ids/driverproc.c.id
     vfw/src/.arch-ids/driverproc.def.id
     vfw/src/.arch-ids/resource.h.id rawdec/.arch-ids/rawdec.c.id
     rawdec/.arch-ids/rawdec.dsp.id rawdec/.arch-ids/=id
     dshow/.arch-ids/=id dshow/.arch-ids/authors.txt.id
     dshow/.arch-ids/dshow.dsp.id dshow/src/.arch-ids/=id
     dshow/src/.arch-ids/CAbout.cpp.id
     dshow/src/.arch-ids/CAbout.h.id
     dshow/src/.arch-ids/CXvidDecoder.cpp.id
     dshow/src/.arch-ids/CXvidDecoder.h.id
     dshow/src/.arch-ids/IXvidDecoder.h.id
     dshow/src/.arch-ids/resource.h.id
     dshow/src/.arch-ids/xvid.ax.def.id
     dshow/src/.arch-ids/xvid.ax.rc.id vfw/vfw.dsp vfw/bin/xvid.inf
     vfw/src/2pass.c vfw/src/2pass.h vfw/src/codec.c
     vfw/src/codec.h vfw/src/config.c vfw/src/config.h
     vfw/src/config.rc vfw/src/driverproc.c vfw/src/driverproc.def
     vfw/src/resource.h rawdec/rawdec.c rawdec/rawdec.dsp
     dshow/authors.txt dshow/dshow.dsp dshow/src/CAbout.cpp
     dshow/src/CAbout.h dshow/src/CXvidDecoder.cpp
     dshow/src/CXvidDecoder.h dshow/src/IXvidDecoder.h
     dshow/src/resource.h dshow/src/xvid.ax.def
     dshow/src/xvid.ax.rc

    modified files:
     build/generic/Makefile build/generic/bootstrap.sh
     build/generic/configure.in build/generic/platform.inc.in
     build/win32/libxvidcore.dsp examples/Makefile
     examples/xvid_bench.c examples/xvid_decraw.c
     examples/xvid_encraw.c examples/xvid_stat.c
     src/bitstream/bitstream.c src/bitstream/bitstream.h
     src/bitstream/mbcoding.c src/bitstream/vlc_codes.h
     src/decoder.c src/decoder.h src/encoder.c src/encoder.h
     src/global.h src/image/colorspace.c src/image/colorspace.h
     src/image/image.c src/image/image.h src/image/interpolate8x8.c
     src/motion/motion.h src/motion/motion_comp.c
     src/motion/motion_est.c src/motion/motion_est.h
     src/motion/smp_motion_est.c src/motion/smp_motion_est.h
     src/portab.h src/prediction/mbprediction.c
     src/utils/mbtransquant.c src/xvid.c src/xvid.h todo.txt

    new directories:
     dshow/.arch-ids dshow/src/.arch-ids rawdec/.arch-ids
     vfw/.arch-ids vfw/bin/.arch-ids vfw/src/.arch-ids vfw vfw/bin
     vfw/src rawdec dshow dshow/src


2003-03-06 21:27:16 GMT                                  base-0

    Summary:
      tag of ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-6
    Revision:
      xvidcore--devapi4--1.0--base-0

    (automatically generated log message)
    

    new patches:
     ed.gomez@free.fr--main/xvidcore--stable--0.9--base-0
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-1
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-2
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-3
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-4
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-5
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-6
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-7
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-8
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-9
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-10
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-11
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-12
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-13
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-14
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-15
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-16
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-17
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-18
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-19
     ed.gomez@free.fr--main/xvidcore--stable--0.9--patch-20
     ed.gomez@free.fr--main/xvidcore--stable--0.9--version-0
     ed.gomez@free.fr--main/xvidcore--stable--1.0--base-0
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-1
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-2
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-3
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-4
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-5
     ed.gomez@free.fr--main/xvidcore--stable--1.0--patch-6

2003-02-15 18:40:33 GMT                                  patch-3

    Summary:
      Added latest Pete's chroma optimization code.
    Revision:
      xvidcore--stable--1.0--patch-3

    During merge, i forgot that small piece of code.

    modified files:
     src/encoder.c src/xvid.h


2003-02-15 14:48:12 GMT                                  patch-2

    Summary:
      Fixed compilation and moved back to plain GPL.
    Revision:
      xvidcore--stable--1.0--patch-2

    portab.h was missing the DPRINTF_RC flag.
    I merged the watcom C portab.h part.
    
    Back to plain GPL as it is in dev-api-3 branch.

    modified files:
     LICENSE src/portab.h


2003-02-15 14:05:17 GMT                                  patch-1

    Summary:
      Updated tree to dev-api-3 branch.
    Revision:
      xvidcore--stable--1.0--patch-1

    This patch consists basically of merging things with dev-api-3
    CVS branch.
    
     - All asm, C, h files have been copied from this CVS branch.
     - Fixed generic sources.inc to take care of changes.
     - Fixed architecture stuff in sources (ARCH_IS_...).
     - Updated examples.
     - Updated libxvidcore.dsp.
     - Fixed libxvidcore.dsp for ARCH_IS_... constants.
     - Changed .so and .a naming convention. It is now postfixed
       with the API version to avoid API incompatibilities with 2.1
       which had had a postfix.
     - Fixed configure version 0.9.1 <-> 1.0.0 cvs-snaphot
    
    This is the first attempt. I don't even know if it compiles well.

    new files:
     src/bitstream/x86_asm/.arch-ids/cbp_3dne.asm.id
     src/dct/.arch-ids/simple_idct.c.id
     src/dct/x86_asm/.arch-ids/fdct_xmm.asm.id
     src/dct/x86_asm/.arch-ids/idct_3dne.asm.id
     src/dct/x86_asm/.arch-ids/simple_idct_mmx.asm.id
     src/image/.arch-ids/font.c.id src/image/.arch-ids/font.h.id
     src/image/.arch-ids/reduced.c.id
     src/image/.arch-ids/reduced.h.id
     src/image/x86_asm/.arch-ids/colorspace_mmx.inc.id
     src/image/x86_asm/.arch-ids/colorspace_rgb_mmx.asm.id
     src/image/x86_asm/.arch-ids/colorspace_yuv_mmx.asm.id
     src/image/x86_asm/.arch-ids/colorspace_yuyv_mmx.asm.id
     src/image/x86_asm/.arch-ids/interpolate8x8_3dne.asm.id
     src/image/x86_asm/.arch-ids/reduced_mmx.asm.id
     src/motion/.arch-ids/motion_est.h.id
     src/motion/.arch-ids/smp_motion_est.c.id
     src/motion/.arch-ids/smp_motion_est.h.id
     src/motion/x86_asm/.arch-ids/sad_3dne.asm.id
     src/quant/x86_asm/.arch-ids/quantize4_xmm.asm.id
     src/quant/x86_asm/.arch-ids/quantize_3dne.asm.id
     src/utils/x86_asm/.arch-ids/interlacing_mmx.asm.id
     src/utils/x86_asm/.arch-ids/mem_transfer_3dne.asm.id
     src/bitstream/x86_asm/cbp_3dne.asm src/dct/simple_idct.c
     src/dct/x86_asm/fdct_xmm.asm src/dct/x86_asm/idct_3dne.asm
     src/dct/x86_asm/simple_idct_mmx.asm src/image/font.c
     src/image/font.h src/image/reduced.c src/image/reduced.h
     src/image/x86_asm/colorspace_mmx.inc
     src/image/x86_asm/colorspace_rgb_mmx.asm
     src/image/x86_asm/colorspace_yuv_mmx.asm
     src/image/x86_asm/colorspace_yuyv_mmx.asm
     src/image/x86_asm/interpolate8x8_3dne.asm
     src/image/x86_asm/reduced_mmx.asm src/motion/motion_est.h
     src/motion/smp_motion_est.c src/motion/smp_motion_est.h
     src/motion/x86_asm/sad_3dne.asm
     src/quant/x86_asm/quantize4_xmm.asm
     src/quant/x86_asm/quantize_3dne.asm
     src/utils/x86_asm/interlacing_mmx.asm
     src/utils/x86_asm/mem_transfer_3dne.asm

    modified files:
     build/generic/Makefile build/generic/configure.in
     build/generic/sources.inc build/win32/libxvidcore.dsp
     examples/odivx_enc_dec.c examples/xvid_bench.c
     examples/xvid_decraw.c examples/xvid_encraw.c
     examples/xvid_stat.c src/bitstream/bitstream.c
     src/bitstream/bitstream.h src/bitstream/cbp.c
     src/bitstream/cbp.h src/bitstream/mbcoding.c
     src/bitstream/mbcoding.h src/bitstream/ppc_asm/cbp_altivec.s
     src/bitstream/ppc_asm/cbp_ppc.s src/bitstream/vlc_codes.h
     src/bitstream/x86_asm/cbp_mmx.asm
     src/bitstream/x86_asm/cbp_sse2.asm src/bitstream/zigzag.h
     src/dct/fdct.c src/dct/fdct.h src/dct/ia64_asm/fdct_ia64.s
     src/dct/idct.c src/dct/idct.h src/dct/ppc_asm/fdct_altivec.s
     src/dct/ppc_asm/idct_altivec.s src/dct/x86_asm/fdct_mmx.asm
     src/dct/x86_asm/idct_mmx.asm src/decoder.c src/decoder.h
     src/divx4.c src/divx4.h src/encoder.c src/encoder.h
     src/global.h src/image/colorspace.c src/image/colorspace.h
     src/image/image.c src/image/image.h src/image/interpolate8x8.c
     src/image/interpolate8x8.h
     src/image/x86_asm/interpolate8x8_3dn.asm
     src/image/x86_asm/interpolate8x8_mmx.asm
     src/image/x86_asm/interpolate8x8_xmm.asm
     src/image/x86_asm/rgb_to_yv12_mmx.asm
     src/image/x86_asm/yuv_to_yv12_mmx.asm
     src/image/x86_asm/yuyv_to_yv12_mmx.asm
     src/image/x86_asm/yv12_to_rgb24_mmx.asm
     src/image/x86_asm/yv12_to_rgb32_mmx.asm
     src/image/x86_asm/yv12_to_yuyv_mmx.asm
     src/motion/ia64_asm/sad_ia64.s src/motion/motion.h
     src/motion/motion_comp.c src/motion/motion_est.c
     src/motion/ppc_asm/sad_altivec.c src/motion/sad.c
     src/motion/sad.h src/motion/x86_asm/sad_3dn.asm
     src/motion/x86_asm/sad_mmx.asm src/motion/x86_asm/sad_sse2.asm
     src/motion/x86_asm/sad_xmm.asm src/portab.h
     src/prediction/mbprediction.c src/prediction/mbprediction.h
     src/quant/adapt_quant.c src/quant/adapt_quant.h
     src/quant/quant_h263.c src/quant/quant_h263.h
     src/quant/quant_matrix.c src/quant/quant_matrix.h
     src/quant/quant_mpeg4.c src/quant/quant_mpeg4.h
     src/quant/x86_asm/quantize4_mmx.asm
     src/quant/x86_asm/quantize_mmx.asm src/utils/emms.c
     src/utils/emms.h src/utils/ia64_asm/mem_transfer_ia64.s
     src/utils/mbfunctions.h src/utils/mbtransquant.c
     src/utils/mem_align.c src/utils/mem_align.h
     src/utils/mem_transfer.c src/utils/mem_transfer.h
     src/utils/ratecontrol.c src/utils/ratecontrol.h
     src/utils/timer.c src/utils/timer.h
     src/utils/x86_asm/cpuid.asm
     src/utils/x86_asm/mem_transfer_mmx.asm src/xvid.c src/xvid.h


2003-02-14 23:01:44 GMT                                  base-0

    Summary:
      Upcoming 1.0 version continuation
    Revision:
      xvidcore--stable--1.0--base-0

    Continuation of the xvidcore--stable--0.9 version.

    new directories:
     {arch}/xvidcore/xvidcore--stable/xvidcore--stable--1.0
     {arch}/xvidcore/xvidcore--stable/xvidcore--stable--1.0/ed.gomez@free.fr--main
     {arch}/xvidcore/xvidcore--stable/xvidcore--stable--1.0/ed.gomez@free.fr--main/patch-log


2003-02-11 21:03:19 GMT                                  patch-20

    Summary:
      Removed BFrame outdated bframe/qpel decoding.
    Revision:
      xvidcore--stable--0.9--patch-20

    Michael noticed there were still pieces of bframe and qpel decoding.
    He proposed to remove it or upgrade it... Feeling too lazy to merge
    all differences (the too much divergent) file from dev-api-3.

    modified files:
     changelog.txt src/decoder.c src/image/interpolate8x8.c
     src/image/interpolate8x8.h


2003-02-11 18:40:48 GMT                                  patch-19

    Summary:
      Fixed libxvidcore.def, revamped Makefile output, fixed ia64 build, added ranlib detection.
    Revision:
      xvidcore--stable--0.9--patch-19

    Ok this patch does lot of things.
    
    First, it fixes libxvidcore.def for win32 targets. This file is now
    generated at configure time. This way we make sure no symobols are
    exported without being compiled in.
    
    Second, I revamped Makefile so its output is more eye candy.
    
    Third change, I removed the ia64 dct file from SRC_IA64 variable,
    this was interfering with the DCT_IA64_SOURCES variable.
    
    Fourth change, ranlib is detected at configure time and used in the
    Makeile through the RANLIB variable.

    modified files:
     build/generic/Makefile build/generic/configure.in
     build/generic/libxvidcore.def.in build/generic/platform.inc.in
     build/generic/sources.inc

    renamed files:
     build/generic/.arch-ids/libxvidcore.def.id
       ==> build/generic/.arch-ids/libxvidcore.def.in.id
     build/generic/libxvidcore.def
       ==> build/generic/libxvidcore.def.in



2003-02-10 23:31:01 GMT                                  patch-18

    Summary:
      Fixed xvid_encraw help message.
    Revision:
      xvidcore--stable--0.9--patch-18

    Fixed xvid_encraw help message.

    modified files:
     examples/xvid_encraw.c


2003-02-10 23:06:32 GMT                                  patch-17

    Summary:
      Added IA64 DCT source choice according to the compiler basename.
    Revision:
      xvidcore--stable--0.9--patch-17

    The IA64 dct file must be choosen according to the compiler.
    I chose to look for a basename based on the *ecc* regexp, all
    other compiler will be treated as being the GNU C compiler.
    
    Hope this is enough.

    modified files:
     build/generic/Makefile build/generic/configure.in
     build/generic/platform.inc.in


2003-02-10 13:49:25 GMT                                  patch-16

    Summary:
      Changed linking option on PPC platforms (-flat_namespace)
    Revision:
      xvidcore--stable--0.9--patch-16

    Guillaume sent me this fix for PPC platforms.

    modified files:
     build/generic/configure.in


2003-02-09 23:15:18 GMT                                  patch-15

    Summary:
      Added the configure bootstrap script.
    Revision:
      xvidcore--stable--0.9--patch-15

    Added the configure bootstrap script.

    new files:
     build/generic/.arch-ids/bootstrap.sh.id
     build/generic/bootstrap.sh


2003-02-09 23:06:51 GMT                                  patch-14

    Summary:
      The PPC port is now disabled because it is outdated.
    Revision:
      xvidcore--stable--0.9--patch-14

    The PPC port is now disabled because it is outdated.

    modified files:
     build/generic/configure.in


2003-02-09 23:01:30 GMT                                  patch-13

    Summary:
      More "unknown compiler" friendly portab.h file.
    Revision:
      xvidcore--stable--0.9--patch-13

    Cristoph pointed out that portab.h was a problem when used with
    unknown compilers. This patch tries to fix that.

    modified files:
     src/portab.h


2003-02-09 00:49:32 GMT                                  patch-12

    Summary:
      DivX4 compatibility layer has been turned into an option (default:disable).
    Revision:
      xvidcore--stable--0.9--patch-12

    The divx4 compatibility API has been turned into an option. This
    has been a long wanted thing by mplayer's guys, so here it is.
    As we say in french "mieux vaut tard que jamais".

    modified files:
     build/generic/Makefile build/generic/configure.in
     build/generic/platform.inc.in build/generic/sources.inc


2003-02-08 23:29:55 GMT                                  patch-11

    Summary:
      Fixed WIN32/_MSC_VER confusion and updated MSVC libxvidcore.dsp project file.
    Revision:
      xvidcore--stable--0.9--patch-11

    This patch set fixes all WIN32/_MSC_VER conditional compilation
    in examples and in the Illegal Instruction detection for SSE2
    support in xvid.c.
    
    libxvidcore.dsp file as been updated with the right defines for
    x86 support (ARCH_IS_IA32, ARCH_IS_32BIT, ARCH_IS_LITTLE_ENDIAN)
    
    Hope Win32 is now completely ready.

    modified files:
     build/win32/libxvidcore.dsp examples/xvid_bench.c
     examples/xvid_decraw.c examples/xvid_encraw.c
     examples/xvid_stat.c src/xvid.c


2003-02-08 14:55:19 GMT                                  patch-10

    Summary:
      Fixed MacOSX build.
    Revision:
      xvidcore--stable--0.9--patch-10

    Two fixes for MacOSX. It adds a missing option to gcc to allow the
    linking stage on this platform (-fno-common). It fixes Altivec test
    which was outputting result on the console.

    modified files:
     build/generic/configure.in


2003-02-08 12:49:17 GMT                                  patch-9

    Summary:
      Added Altivec detection (Guillaume Morin)
    Revision:
      xvidcore--stable--0.9--patch-9

    Added Altivec detection test in configure.in.

    modified files:
     build/generic/configure.in


2003-02-08 12:25:46 GMT                                  patch-8

    Summary:
      Fixed portab.h for _DEBUG target and GCC/ICC compilers.
    Revision:
      xvidcore--stable--0.9--patch-8

    During the Unix build system change, i had to turn the DRPINTF macro
    into a real function because teh MacOSX compilers do not support the
    variadic macros as defined in ISO C99 standard. During this change,
    i forgot to adapt the macro code and the #include <stdarg.h> needed
    for variadic functions.

    modified files:
     src/portab.h


2003-02-08 11:45:00 GMT                                  patch-7

    Summary:
      Changed the way I add strings into variables.
    Revision:
      xvidcore--stable--0.9--patch-7

    I changed the way I add strings to variables (CFLAGS and so on). Now i use
    var="$var string2" instead of var=$var" string2".
    
    Fixed a typo reported by Pete for the cygwin part.

    modified files:
     build/generic/configure.in


2003-02-07 23:16:57 GMT                                  patch-6

    Summary:
      Fixed the "ar" "s" option for some platforms.
    Revision:
      xvidcore--stable--0.9--patch-6

    The "s" option of the "ar" program is not standard accross all
    platforms. I had at least problems on OpenBSD and an old Solaris
    version.

    modified files:
     build/generic/Makefile


2003-02-07 22:19:37 GMT                                  patch-5

    Summary:
      Fixed a BSD checking in ansm output format.
    Revision:
      xvidcore--stable--0.9--patch-5

    Fixed a BSD checking in ansm output format.

    modified files:
     build/generic/configure.in


2003-02-07 21:18:14 GMT                                  patch-4

    Summary:
      Fixed options and added the --disable-assembly option
    Revision:
      xvidcore--stable--0.9--patch-4

    Options have been fixed because they were not taking care of the
    enable_feature variable.
    
    The --disable-assembly options has been added. This is a good way
    to compile XviD on nearly all platforms without having to deal with
    the assembly code -- useful on PPC platform at the moment where gcc
    seems to use a different kind of assembly syntax.

    modified files:
     build/generic/configure.in


2003-02-06 21:49:16 GMT                                  patch-3

    Summary:
      Fixes for the new build system in sources.
    Revision:
      xvidcore--stable--0.9--patch-3

    This patch fixes source files according to the new defines used
    by the reworked build system.

    modified files:
     src/bitstream/bitstream.h src/divx4.h src/portab.h
     src/utils/emms.h src/xvid.c


2003-02-06 21:22:55 GMT                                  patch-2

    Summary:
      Changed build system for Unix OSes
    Revision:
      xvidcore--stable--0.9--patch-2

    Changed the build system for Unix systems. It is now built upon
    an autoconf script that automatically configures the sources.
    
    The Makefile is portable accross various platforms and "make"
    programs. It is at least working on these platforms for now:
     - Debian GNU/Linux - StrongARM - Alphave67 (alpha 64bit) - ia32
       UltraSparcIII
     - Solaris - UltraSparcI - Sparc 32bit on old sun stations ( i don't
       remember the exact name)
     - FreeBSD 4.7 - ia32
     - RedHat 7.3 - ia32
     - Gentoo 1.4 - ia32
     - the Irix box according to christoph tests
     - ia64 - Unknown OS?
    
    The unix unified makefile supports:
     - gmake
     - pmake
    
    ToDo things to finish this new build system:
     - Manage the ecc/gcc source choice for ia64
     - Someone to test the makefile on Cygwin and/or mingw+minsys
     - Update MSVC projects (replace 2 or 3 defines)
     - See why MacOSX is complaining about duplicated symbols, it seems the
       mach ABI does not alow namespace collisions even between C modules.
       And add altivec detection in configure.in

    new files:
     build/generic/.arch-ids/configure.in.id
     build/generic/.arch-ids/Makefile.id
     build/generic/.arch-ids/platform.inc.in.id
     build/generic/.arch-ids/sources.inc.id
     build/generic/configure.in build/generic/Makefile
     build/generic/platform.inc.in build/generic/sources.inc

    removed files:
     build/generic/.arch-ids/Makefile.beos.id
     build/generic/.arch-ids/Makefile.cygwin.id
     build/generic/.arch-ids/Makefile.dj.id
     build/generic/.arch-ids/Makefile.freebsd.id
     build/generic/.arch-ids/Makefile.generic.id
     build/generic/.arch-ids/Makefile.ia64.id
     build/generic/.arch-ids/Makefile.inc.id
     build/generic/.arch-ids/Makefile.irix64.id
     build/generic/.arch-ids/Makefile.linuxppc.id
     build/generic/.arch-ids/Makefile.linuxppc_altivec.id
     build/generic/.arch-ids/Makefile.linuxx86.id
     build/generic/.arch-ids/Makefile.sparc.id
     build/generic/Makefile.beos build/generic/Makefile.cygwin
     build/generic/Makefile.dj build/generic/Makefile.freebsd
     build/generic/Makefile.generic build/generic/Makefile.ia64
     build/generic/Makefile.inc build/generic/Makefile.irix64
     build/generic/Makefile.linuxppc
     build/generic/Makefile.linuxppc_altivec
     build/generic/Makefile.linuxx86 build/generic/Makefile.sparc


2003-02-06 21:11:17 GMT                                  patch-1

    Summary:
      Updated to current stable CVS_HEAD
    Revision:
      xvidcore--stable--0.9--patch-1

    Updated files to current stable CVS_HEAD versions.

    new files:
     build/win32/.arch-ids/odivx_enc_dec.dsp.id
     build/win32/.arch-ids/xvidcore.dsw.id
     build/win32/.arch-ids/xvid_bench.dsp.id
     build/win32/odivx_enc_dec.dsp build/win32/xvidcore.dsw
     build/win32/xvid_bench.dsp

    modified files:
     authors.txt build/generic/Makefile.beos
     build/generic/Makefile.generic build/generic/Makefile.linuxx86
     doc/Makefile examples/Makefile examples/odivx_enc_dec.c
     examples/xvid_bench.c examples/xvid_encraw.c
     examples/xvid_stat.c src/bitstream/bitstream.c
     src/bitstream/bitstream.h src/bitstream/cbp.c
     src/bitstream/mbcoding.c src/bitstream/vlc_codes.h
     src/bitstream/zigzag.h src/dct/fdct.c src/dct/idct.c
     src/dct/idct.h src/decoder.c src/decoder.h src/divx4.h
     src/encoder.c src/encoder.h src/global.h
     src/image/colorspace.c src/image/image.c
     src/image/interpolate8x8.c src/image/interpolate8x8.h
     src/motion/motion.h src/motion/motion_comp.c
     src/motion/motion_est.c src/motion/sad.c src/portab.h
     src/prediction/mbprediction.c src/prediction/mbprediction.h
     src/quant/adapt_quant.c src/quant/adapt_quant.h
     src/quant/quant_h263.c src/quant/quant_mpeg4.c
     src/utils/emms.h src/utils/mbfunctions.h
     src/utils/mbtransquant.c src/utils/mem_align.c
     src/utils/mem_transfer.c src/utils/timer.c src/utils/timer.h
     src/xvid.h todo.txt


2003-02-06 20:59:19 GMT                                  base-0

    Summary:
      Imported xvidcore 0.9.0 into arch repository
    Revision:
      xvidcore--stable--0.9--base-0

    Imported xvidcore 0.9.0 into arch repository. I hope I forgot nothing.

    new files:
     ./.arch-ids/CodingStyle.id ./.arch-ids/LICENSE.id
     ./.arch-ids/README.txt.id ./.arch-ids/authors.txt.id
     ./.arch-ids/changelog.txt.id ./.arch-ids/todo.txt.id
     ./CodingStyle ./LICENSE ./README.txt ./authors.txt
     ./build/.arch-ids/=id ./build/generic/.arch-ids/=id
     ./build/generic/.arch-ids/Makefile.beos.id
     ./build/generic/.arch-ids/Makefile.cygwin.id
     ./build/generic/.arch-ids/Makefile.dj.id
     ./build/generic/.arch-ids/Makefile.freebsd.id
     ./build/generic/.arch-ids/Makefile.generic.id
     ./build/generic/.arch-ids/Makefile.ia64.id
     ./build/generic/.arch-ids/Makefile.inc.id
     ./build/generic/.arch-ids/Makefile.irix64.id
     ./build/generic/.arch-ids/Makefile.linuxppc.id
     ./build/generic/.arch-ids/Makefile.linuxppc_altivec.id
     ./build/generic/.arch-ids/Makefile.linuxx86.id
     ./build/generic/.arch-ids/Makefile.sparc.id
     ./build/generic/.arch-ids/libxvidcore.def.id
     ./build/generic/Makefile.beos ./build/generic/Makefile.cygwin
     ./build/generic/Makefile.dj ./build/generic/Makefile.freebsd
     ./build/generic/Makefile.generic ./build/generic/Makefile.ia64
     ./build/generic/Makefile.inc ./build/generic/Makefile.irix64
     ./build/generic/Makefile.linuxppc
     ./build/generic/Makefile.linuxppc_altivec
     ./build/generic/Makefile.linuxx86
     ./build/generic/Makefile.sparc ./build/generic/libxvidcore.def
     ./build/win32/.arch-ids/=id
     ./build/win32/.arch-ids/libxvidcore.dsp.id
     ./build/win32/.arch-ids/xvid_decraw.dsp.id
     ./build/win32/.arch-ids/xvid_encraw.dsp.id
     ./build/win32/.arch-ids/xvid_stat.dsp.id
     ./build/win32/libxvidcore.dsp ./build/win32/xvid_decraw.dsp
     ./build/win32/xvid_encraw.dsp ./build/win32/xvid_stat.dsp
     ./changelog.txt ./doc/.arch-ids/=id ./doc/.arch-ids/API.dox.id
     ./doc/.arch-ids/Makefile.id ./doc/.arch-ids/README.id
     ./doc/.arch-ids/foot.inc.in.id
     ./doc/.arch-ids/header.tex.in.id
     ./doc/.arch-ids/xvid-decoding.txt.id
     ./doc/.arch-ids/xvid-encoder.txt.id ./doc/API.dox
     ./doc/Makefile ./doc/README ./doc/foot.inc.in
     ./doc/header.tex.in ./doc/xvid-decoding.txt
     ./doc/xvid-encoder.txt ./examples/.arch-ids/=id
     ./examples/.arch-ids/Makefile.id
     ./examples/.arch-ids/README.id
     ./examples/.arch-ids/cactus.pgm.bz2.id
     ./examples/.arch-ids/odivx_enc_dec.c.id
     ./examples/.arch-ids/xvid_bench.c.id
     ./examples/.arch-ids/xvid_decraw.c.id
     ./examples/.arch-ids/xvid_encraw.c.id
     ./examples/.arch-ids/xvid_stat.c.id ./examples/Makefile
     ./examples/README ./examples/cactus.pgm.bz2
     ./examples/odivx_enc_dec.c ./examples/xvid_bench.c
     ./examples/xvid_decraw.c ./examples/xvid_encraw.c
     ./examples/xvid_stat.c ./src/.arch-ids/=id
     ./src/.arch-ids/decoder.c.id ./src/.arch-ids/decoder.h.id
     ./src/.arch-ids/divx4.c.id ./src/.arch-ids/divx4.h.id
     ./src/.arch-ids/encoder.c.id ./src/.arch-ids/encoder.h.id
     ./src/.arch-ids/global.h.id ./src/.arch-ids/portab.h.id
     ./src/.arch-ids/xvid.c.id ./src/.arch-ids/xvid.h.id
     ./src/bitstream/.arch-ids/=id
     ./src/bitstream/.arch-ids/bitstream.c.id
     ./src/bitstream/.arch-ids/bitstream.h.id
     ./src/bitstream/.arch-ids/cbp.c.id
     ./src/bitstream/.arch-ids/cbp.h.id
     ./src/bitstream/.arch-ids/mbcoding.c.id
     ./src/bitstream/.arch-ids/mbcoding.h.id
     ./src/bitstream/.arch-ids/vlc_codes.h.id
     ./src/bitstream/.arch-ids/zigzag.h.id
     ./src/bitstream/bitstream.c ./src/bitstream/bitstream.h
     ./src/bitstream/cbp.c ./src/bitstream/cbp.h
     ./src/bitstream/mbcoding.c ./src/bitstream/mbcoding.h
     ./src/bitstream/ppc_asm/.arch-ids/=id
     ./src/bitstream/ppc_asm/.arch-ids/cbp_altivec.s.id
     ./src/bitstream/ppc_asm/.arch-ids/cbp_ppc.s.id
     ./src/bitstream/ppc_asm/cbp_altivec.s
     ./src/bitstream/ppc_asm/cbp_ppc.s ./src/bitstream/vlc_codes.h
     ./src/bitstream/x86_asm/.arch-ids/=id
     ./src/bitstream/x86_asm/.arch-ids/cbp_mmx.asm.id
     ./src/bitstream/x86_asm/.arch-ids/cbp_sse2.asm.id
     ./src/bitstream/x86_asm/cbp_mmx.asm
     ./src/bitstream/x86_asm/cbp_sse2.asm ./src/bitstream/zigzag.h
     ./src/dct/.arch-ids/=id ./src/dct/.arch-ids/README.IJG.id
     ./src/dct/.arch-ids/fdct.c.id ./src/dct/.arch-ids/fdct.h.id
     ./src/dct/.arch-ids/idct.c.id ./src/dct/.arch-ids/idct.h.id
     ./src/dct/README.IJG ./src/dct/fdct.c ./src/dct/fdct.h
     ./src/dct/ia64_asm/.arch-ids/=id
     ./src/dct/ia64_asm/.arch-ids/fdct_ia64.s.id
     ./src/dct/ia64_asm/.arch-ids/genidct.py.id
     ./src/dct/ia64_asm/.arch-ids/idct_fini.s.id
     ./src/dct/ia64_asm/.arch-ids/idct_ia64_ecc.s.id
     ./src/dct/ia64_asm/.arch-ids/idct_ia64_gcc.s.id
     ./src/dct/ia64_asm/.arch-ids/idct_init.s.id
     ./src/dct/ia64_asm/fdct_ia64.s ./src/dct/ia64_asm/genidct.py
     ./src/dct/ia64_asm/idct_fini.s
     ./src/dct/ia64_asm/idct_ia64_ecc.s
     ./src/dct/ia64_asm/idct_ia64_gcc.s
     ./src/dct/ia64_asm/idct_init.s ./src/dct/idct.c
     ./src/dct/idct.h ./src/dct/ppc_asm/.arch-ids/=id
     ./src/dct/ppc_asm/.arch-ids/fdct_altivec.s.id
     ./src/dct/ppc_asm/.arch-ids/idct_altivec.s.id
     ./src/dct/ppc_asm/fdct_altivec.s
     ./src/dct/ppc_asm/idct_altivec.s
     ./src/dct/x86_asm/.arch-ids/=id
     ./src/dct/x86_asm/.arch-ids/fdct_mmx.asm.id
     ./src/dct/x86_asm/.arch-ids/idct_mmx.asm.id
     ./src/dct/x86_asm/fdct_mmx.asm ./src/dct/x86_asm/idct_mmx.asm
     ./src/decoder.c ./src/decoder.h ./src/divx4.c ./src/divx4.h
     ./src/encoder.c ./src/encoder.h ./src/global.h
     ./src/image/.arch-ids/=id
     ./src/image/.arch-ids/colorspace.c.id
     ./src/image/.arch-ids/colorspace.h.id
     ./src/image/.arch-ids/image.c.id
     ./src/image/.arch-ids/image.h.id
     ./src/image/.arch-ids/interpolate8x8.c.id
     ./src/image/.arch-ids/interpolate8x8.h.id
     ./src/image/colorspace.c ./src/image/colorspace.h
     ./src/image/ia64_asm/.arch-ids/=id
     ./src/image/ia64_asm/.arch-ids/README.id
     ./src/image/ia64_asm/.arch-ids/interpolate8x8_ia64.s.id
     ./src/image/ia64_asm/.arch-ids/interpolate8x8_ia64_exact.s.id
     ./src/image/ia64_asm/README
     ./src/image/ia64_asm/interpolate8x8_ia64.s
     ./src/image/ia64_asm/interpolate8x8_ia64_exact.s
     ./src/image/image.c ./src/image/image.h
     ./src/image/interpolate8x8.c ./src/image/interpolate8x8.h
     ./src/image/x86_asm/.arch-ids/=id
     ./src/image/x86_asm/.arch-ids/interpolate8x8_3dn.asm.id
     ./src/image/x86_asm/.arch-ids/interpolate8x8_mmx.asm.id
     ./src/image/x86_asm/.arch-ids/interpolate8x8_xmm.asm.id
     ./src/image/x86_asm/.arch-ids/rgb_to_yv12_mmx.asm.id
     ./src/image/x86_asm/.arch-ids/yuv_to_yv12_mmx.asm.id
     ./src/image/x86_asm/.arch-ids/yuyv_to_yv12_mmx.asm.id
     ./src/image/x86_asm/.arch-ids/yv12_to_rgb24_mmx.asm.id
     ./src/image/x86_asm/.arch-ids/yv12_to_rgb32_mmx.asm.id
     ./src/image/x86_asm/.arch-ids/yv12_to_yuyv_mmx.asm.id
     ./src/image/x86_asm/interpolate8x8_3dn.asm
     ./src/image/x86_asm/interpolate8x8_mmx.asm
     ./src/image/x86_asm/interpolate8x8_xmm.asm
     ./src/image/x86_asm/rgb_to_yv12_mmx.asm
     ./src/image/x86_asm/yuv_to_yv12_mmx.asm
     ./src/image/x86_asm/yuyv_to_yv12_mmx.asm
     ./src/image/x86_asm/yv12_to_rgb24_mmx.asm
     ./src/image/x86_asm/yv12_to_rgb32_mmx.asm
     ./src/image/x86_asm/yv12_to_yuyv_mmx.asm
     ./src/motion/.arch-ids/=id ./src/motion/.arch-ids/motion.h.id
     ./src/motion/.arch-ids/motion_comp.c.id
     ./src/motion/.arch-ids/motion_est.c.id
     ./src/motion/.arch-ids/sad.c.id
     ./src/motion/.arch-ids/sad.h.id
     ./src/motion/ia64_asm/.arch-ids/=id
     ./src/motion/ia64_asm/.arch-ids/calc_delta_1.s.id
     ./src/motion/ia64_asm/.arch-ids/calc_delta_2.s.id
     ./src/motion/ia64_asm/.arch-ids/calc_delta_3.s.id
     ./src/motion/ia64_asm/.arch-ids/halfpel8_refine_ia64.s.id
     ./src/motion/ia64_asm/.arch-ids/sad_ia64.s.id
     ./src/motion/ia64_asm/calc_delta_1.s
     ./src/motion/ia64_asm/calc_delta_2.s
     ./src/motion/ia64_asm/calc_delta_3.s
     ./src/motion/ia64_asm/halfpel8_refine_ia64.s
     ./src/motion/ia64_asm/sad_ia64.s ./src/motion/motion.h
     ./src/motion/motion_comp.c ./src/motion/motion_est.c
     ./src/motion/ppc_asm/.arch-ids/=id
     ./src/motion/ppc_asm/.arch-ids/README.id
     ./src/motion/ppc_asm/.arch-ids/sad_altivec.c.id
     ./src/motion/ppc_asm/.arch-ids/sad_altivec.s.id
     ./src/motion/ppc_asm/README ./src/motion/ppc_asm/sad_altivec.c
     ./src/motion/ppc_asm/sad_altivec.s ./src/motion/sad.c
     ./src/motion/sad.h ./src/motion/x86_asm/.arch-ids/=id
     ./src/motion/x86_asm/.arch-ids/sad_3dn.asm.id
     ./src/motion/x86_asm/.arch-ids/sad_mmx.asm.id
     ./src/motion/x86_asm/.arch-ids/sad_sse2.asm.id
     ./src/motion/x86_asm/.arch-ids/sad_xmm.asm.id
     ./src/motion/x86_asm/sad_3dn.asm
     ./src/motion/x86_asm/sad_mmx.asm
     ./src/motion/x86_asm/sad_sse2.asm
     ./src/motion/x86_asm/sad_xmm.asm ./src/portab.h
     ./src/prediction/.arch-ids/=id
     ./src/prediction/.arch-ids/mbprediction.c.id
     ./src/prediction/.arch-ids/mbprediction.h.id
     ./src/prediction/mbprediction.c
     ./src/prediction/mbprediction.h ./src/quant/.arch-ids/=id
     ./src/quant/.arch-ids/adapt_quant.c.id
     ./src/quant/.arch-ids/adapt_quant.h.id
     ./src/quant/.arch-ids/quant_h263.c.id
     ./src/quant/.arch-ids/quant_h263.h.id
     ./src/quant/.arch-ids/quant_matrix.c.id
     ./src/quant/.arch-ids/quant_matrix.h.id
     ./src/quant/.arch-ids/quant_mpeg4.c.id
     ./src/quant/.arch-ids/quant_mpeg4.h.id
     ./src/quant/adapt_quant.c ./src/quant/adapt_quant.h
     ./src/quant/ia64_asm/.arch-ids/=id
     ./src/quant/ia64_asm/.arch-ids/quant_h263_ia64.s.id
     ./src/quant/ia64_asm/quant_h263_ia64.s
     ./src/quant/quant_h263.c ./src/quant/quant_h263.h
     ./src/quant/quant_matrix.c ./src/quant/quant_matrix.h
     ./src/quant/quant_mpeg4.c ./src/quant/quant_mpeg4.h
     ./src/quant/x86_asm/.arch-ids/=id
     ./src/quant/x86_asm/.arch-ids/quantize4_mmx.asm.id
     ./src/quant/x86_asm/.arch-ids/quantize_mmx.asm.id
     ./src/quant/x86_asm/quantize4_mmx.asm
     ./src/quant/x86_asm/quantize_mmx.asm ./src/utils/.arch-ids/=id
     ./src/utils/.arch-ids/emms.c.id
     ./src/utils/.arch-ids/emms.h.id
     ./src/utils/.arch-ids/mbfunctions.h.id
     ./src/utils/.arch-ids/mbtransquant.c.id
     ./src/utils/.arch-ids/mem_align.c.id
     ./src/utils/.arch-ids/mem_align.h.id
     ./src/utils/.arch-ids/mem_transfer.c.id
     ./src/utils/.arch-ids/mem_transfer.h.id
     ./src/utils/.arch-ids/ratecontrol.c.id
     ./src/utils/.arch-ids/ratecontrol.h.id
     ./src/utils/.arch-ids/timer.c.id
     ./src/utils/.arch-ids/timer.h.id ./src/utils/emms.c
     ./src/utils/emms.h ./src/utils/ia64_asm/.arch-ids/=id
     ./src/utils/ia64_asm/.arch-ids/mem_transfer_ia64.s.id
     ./src/utils/ia64_asm/mem_transfer_ia64.s
     ./src/utils/mbfunctions.h ./src/utils/mbtransquant.c
     ./src/utils/mem_align.c ./src/utils/mem_align.h
     ./src/utils/mem_transfer.c ./src/utils/mem_transfer.h
     ./src/utils/ratecontrol.c ./src/utils/ratecontrol.h
     ./src/utils/timer.c ./src/utils/timer.h
     ./src/utils/x86_asm/.arch-ids/=id
     ./src/utils/x86_asm/.arch-ids/cpuid.asm.id
     ./src/utils/x86_asm/.arch-ids/mem_transfer_mmx.asm.id
     ./src/utils/x86_asm/cpuid.asm
     ./src/utils/x86_asm/mem_transfer_mmx.asm ./src/xvid.c
     ./src/xvid.h ./todo.txt