This is an av_cold function, and we don't need to duplicate variables just to
save a few characters.

Originally committed as revision 25979 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25978 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25977 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25976 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25975 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25974 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25973 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25972 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25971 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25970 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25969 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25968 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25967 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25966 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25965 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25964 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25963 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25962 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25961 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25960 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25959 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25958 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25957 to svn://svn.ffmpeg.org/ffmpeg/trunk

Turn it into 2 macros, and use av_clip_int16() and lrintf().
This matches the int16 to float sample conversion in audioconvert.c.
The regression test output is different due to lrintf() rounding.

Originally committed as revision 25956 to svn://svn.ffmpeg.org/ffmpeg/trunk

Also, use LFG from libavutil instead of random().

Originally committed as revision 25955 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25954 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25953 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25952 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25951 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25950 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25949 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25948 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25947 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25946 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25945 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25944 to svn://svn.ffmpeg.org/ffmpeg/trunk

allows playback of files where multiple qdm2 packets have been merged.

Originally committed as revision 25943 to svn://svn.ffmpeg.org/ffmpeg/trunk

Patch by Nolan L nol888 <=> gmail >=< com.

See thread:
Subject: [FFmpeg-devel] [PATCH] Port gradfun to libavfilter (GCI)
Date: Mon, 29 Nov 2010 07:18:14 -0500

Originally committed as revision 25942 to svn://svn.ffmpeg.org/ffmpeg/trunk

Patch by James Zern, jzern google

Originally committed as revision 25941 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 25940 to svn://svn.ffmpeg.org/ffmpeg/trunk