[MPlayer-DOCS] CVS: main/DOCS/man/en mplayer.1,1.538,1.539

[Diego Biurrun]

Michael Niedermayer writes:
 > On Sunday 08 February 2004 19:24, Diego Biurrun CVS wrote:
 > > Modified Files:
 > > 	mplayer.1
 > > Log Message:
 > > bit_exact lavc option, patch by "Steven M. Schultz" <sms at 2BSD.COM>
 > >
 > > --- mplayer.1	6 Feb 2004 20:31:57 -0000	1.538
 > > +++ mplayer.1	8 Feb 2004 18:24:24 -0000	1.539
 > > @@ -4064,6 +4064,12 @@
 > >  .B atag=<value>
 > >  Use the specified Windows audio format tag (e.g.\& atag=0x55).
 > >  .TP
 > > +.B bit_exact
 > > +Use only bit exact algorithms (except (i)dct).
 > > +This also suppresses the user_data header in MPEG-4 streams
 > > +(that section causes some players such as Apple's
 > > +Quicktime player to declare the stream corrupt).
 > great
 > 
 > smart user reads this, thinks its good to be bit exact & it even helps 
 > playback with apples player, so he uses it for encoding, and cant decode it 
 > later as the soo much hated userdata contains the version number of the 
 > encoder and without it theres no way to automatically work around past 
 > encoder bugs during playback,

Huh?  Does the header contain the version number or not?

 > additionally bit_exact disables several 
 > optimizations, IMHO these things should be mentioned too
 > bit_exact should _only_ be used for regression tests, which need binary 
 > idntical files even if the encoder version changes

OK, so what about

bit_exact
  Use only bit exact algorithms (except (i)dct).
  Additionally bit_exact disables several optimizations and thus
  should only be used for regression tests, which need binary
  idntical files even if the encoder version changes.
  This also suppresses the user_data header in MPEG-4 streams
  (that section causes some buggy players such as Apple's
  Quicktime player to declare the stream corrupt).

Diego