Originally committed as revision 13001 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 13000 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12999 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12998 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12997 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12996 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12995 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12994 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12993 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12992 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12991 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12990 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12989 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12988 to svn://svn.ffmpeg.org/ffmpeg/trunk

So if referenced blocks is 16x8, 8x16 or 8x8 partitions, skip block
will have them too.

Originally committed as revision 12987 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12986 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12985 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12984 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12983 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12981 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12980 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12979 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12978 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12976 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12975 to svn://svn.ffmpeg.org/ffmpeg/trunk

Also implemented routine for calculating sum(a[i] * a[i-offset]>>shift)

Originally committed as revision 12974 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12973 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12969 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12968 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12967 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12966 to svn://svn.ffmpeg.org/ffmpeg/trunk

fix adpcm ima qt decoding, channel at init is 0, correct stereo out since samples += avctx->channels

Originally committed as revision 12965 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12964 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12963 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12955 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 12953 to svn://svn.ffmpeg.org/ffmpeg/trunk

and then reads the channel_map stuff

Originally committed as revision 12944 to svn://svn.ffmpeg.org/ffmpeg/trunk

Patch by Thorsten Jordan tjordanATmacrosystem de

Originally committed as revision 12936 to svn://svn.ffmpeg.org/ffmpeg/trunk

num_coeffs is just one byte.
It would make no sense for num_coeffs to be as high as 2^32.

Originally committed as revision 12931 to svn://svn.ffmpeg.org/ffmpeg/trunk

Remove the big comment and add comments about the skipped fields.

Originally committed as revision 12930 to svn://svn.ffmpeg.org/ffmpeg/trunk