On 6/5/11 1:20 PM, Jason Garrett-Glaser wrote:
NOTE: THIS IS A WORK-IN-PROGRESS PATCH

Tested with:
8x8dct (inc 8x8dct+CAVLC)
CABAC/CAVLC
Loopfilter
B-frames
Weighted bipred
Chroma QP offsets
Edge emulation

Not tested with:
Explicit weighted pred
MBAFF (JM crashes on this one)
PAFF (JM appears to be bugged on this one)
---
  libavcodec/dsputil.h           |    2 +-
  libavcodec/dsputil_template.c  |    6 +-
  libavcodec/h264.c              |  753 ++++++++++++++++++++++++-------------
  libavcodec/h264.h              |  185 +++++----
  libavcodec/h264_cabac.c        |  808 +++++++++++++++++++++++++++++++++++-----
  libavcodec/h264_cavlc.c        |  196 ++++++----
  libavcodec/h264_loopfilter.c   |   70 +++--
  libavcodec/h264_ps.c           |   14 +-
  libavcodec/h264dsp.h           |    8 +-
  libavcodec/h264idct_template.c |   44 ++-
  libavcodec/mpegvideo.c         |   25 +-
  libavcodec/snow.c              |    6 +-
  libavcodec/x86/dsputil_mmx.c   |    4 +-
  13 files changed, 1525 insertions(+), 596 deletions(-)


You might try to split some of the small things like making it not assume w and h being the same. Beside that the new code seems sort of ok. I will need to re-read it at least another time but after the first pass all I could find isn't exactly related to your changes.

We discussed on irc about the next improvements. Surely would be good having this patch land and then cleanup this code so we don't have that many magic numbers around and obscure log messages.

lu
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to