On Thu, Feb 13, 2003 at 10:34:19AM +0200, Ville Syrj�l� wrote: > I'm sure everyone remembers Brian's CRTC2 troubles. Well I think we > managed to fix it off list some time ago. The problem was caused by > displaying the fields in the wrong order.
And the fix that we found has proven to work well.
Ville, I must appologize for not getting back to you on the included
implementation, but I have not found a moment to hack up all my code
here to do it.
> Here's a patch with the solution I came up with. It adds a new method
> IDirectFBDisplayLayer::SetFieldParity() to select which field is displayed
> first after a Flip(). This behaviour isn't always the best choice so I
> added {DLCAPS,DLOP}_FIELD_PARITY flags.
Cool.
> PS.
> I've also included a patch to mplayer to make use of this. Just use -vo
> dfbmga:fieldparity=0 or 1 to select top field first or bottom field first
> respectively.
Indeed. What is missing, of course, is being able to determine the
field dominance from the stream demuxxer. My experience seems to
reveal that it's quite reliable to assume that field "0" (what
our current implementation in my code tree here is doing) is assumed
to be dominant, but that will of course not always be the case.
It would be nice to be able to replace the command line args for field
dominance with being able to obtain that information from the stream
demuxxer, but alas, I don't think that is currently possible with
MPlayer.
b.
--
Brian J. Murrell
msg01348/pgp00000.pgp
Description: PGP signature
