Originally committed as revision 2414 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2413 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2412 to svn://svn.ffmpeg.org/ffmpeg/trunk

move q_*_matrix out of MpegEncContext (40k ->23k) dct_quantize() is even slightly faster now, dont ask my why ...

Originally committed as revision 2411 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 11214 to svn://svn.mplayerhq.hu/mplayer/trunk/postproc

Originally committed as revision 2410 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2409 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2408 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2407 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2406 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2405 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2404 to svn://svn.ffmpeg.org/ffmpeg/trunk

sample_aspect_ratio
aspect ratio in JPEG JFIF is SAR not DAR !
removed nonsense SAR guessing code
various related cleanups
bugs?

Originally committed as revision 2403 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2402 to svn://svn.ffmpeg.org/ffmpeg/trunk

fix mpeg2 aspect_ratio for the rare case that active display area != AVCodecContext.width/height
decode sequence display extension & picture display extension

Originally committed as revision 2401 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2400 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2399 to svn://svn.ffmpeg.org/ffmpeg/trunk

common, especially since Apple iTunes treats files with the two
extensions differently. Without the patch, FFmpeg won't know that m4a
is for MPEG-4 files
patch by (Dan Christiansen <danchr at daimi dot au dot dk>)

Originally committed as revision 2398 to svn://svn.ffmpeg.org/ffmpeg/trunk

have it passed to GCC)
- Moving the usage in configure to the top of it, so you don't have to
wait needlessly
- Silence compiler output during configuration
- Add support for configuring with IBM XLC

patch by (Dan Christiansen <danchr at daimi dot au dot dk>)

Originally committed as revision 2397 to svn://svn.ffmpeg.org/ffmpeg/trunk

    * fixing YUV4MPEG format.

    * fixing a bug in DV codec where coded_frame was not set.

Originally committed as revision 2396 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2395 to svn://svn.ffmpeg.org/ffmpeg/trunk

     was not set.

Originally committed as revision 2394 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2393 to svn://svn.ffmpeg.org/ffmpeg/trunk

picture buffer cleanup, this fixes a assertion failure (if assertion checking was manualy enabled for utils.c) and a memleak at least

Originally committed as revision 2392 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2391 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2390 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2389 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2388 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2387 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 11124 to svn://svn.mplayerhq.hu/mplayer/trunk/postproc

Originally committed as revision 11123 to svn://svn.mplayerhq.hu/mplayer/trunk/postproc

Originally committed as revision 2386 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2385 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2384 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2383 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2382 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2381 to svn://svn.ffmpeg.org/ffmpeg/trunk

    * fixing an unitilaized field AVImageInfo::interleaved.

      As a matter of fact we have two problems here: first
      AVImageInfo::interleaved is never initialzed in img.c.
      The second problem is that even if we want interleaved
      PNGs there's a bug in png.c that prevents us from generating
      them.

      The later problem is for somebody familiar with PNG to
      handle.

      And yes, always setting AVImageInfo::interleaved to 0 isn't
      such a good idea, but it still beats garbage values.
orginal commit by roman shaposhnik

Originally committed as revision 2380 to svn://svn.ffmpeg.org/ffmpeg/trunk

  * backing out the 0-sized packets patch. We have to devise a more
    sensible approach.
orginal commit by roman shaposhnik

Originally committed as revision 2379 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 2378 to svn://svn.ffmpeg.org/ffmpeg/trunk