Replace mvd>2 + mvd>32 by MIN((mvd+28)*17>>9, 2)
authorMichael Niedermayer <michaelni@gmx.at>
Wed, 24 Feb 2010 16:16:08 +0000 (16:16 +0000)
committerMichael Niedermayer <michaelni@gmx.at>
Wed, 24 Feb 2010 16:16:08 +0000 (16:16 +0000)
commit6d0155c79c5a30ecea5be1edac47c1f412a7a124
treea3b2b1ae9be6b1ae0f72d10a02d3c8867479073b
parentbc77fce6ba17cb7c87f7c19751ba22ceba713aca
Replace mvd>2 + mvd>32 by MIN((mvd+28)*17>>9, 2)
same speed as far as i can meassure but it might have fewer branches on some
archs.
Idea from x264 / jason

Originally committed as revision 22027 to svn://svn.ffmpeg.org/ffmpeg/trunk
libavcodec/h264_cabac.c