changelog

TMS320C6713Xvid视频压缩算法源代码.rar

# 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:
    
     * [e-mail]--2004-1/xvidcore--head--0.0--patch-70
       Out of bounds MVs clipping
    
     * [e-mail]--2004-1/xvidcore--head--0.0--patch-71
       Decoder optimization (fixing regression)
    

    modified files:
     src/decoder.c

    new patches:
     [e-mail]--2004-1/xvidcore--head--0.0--patch-70
     [e-mail]--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: