On Fri, Feb 25, 2005 at 01:09:54PM +0100, Claudio Ciccani wrote:
> Lucian Muresan wrote:
> >Hi,
> >
> >after Ville found out why df_xine inverts colors when using a Matrox
> >CRTC2, I was able to get correct output by commenting out the line in
> >DirectFB-extra/samples/df_xine/control.c in dfx_control_correction where
> >"value" is assigned 0x8000. Well, this may work for my CRTC2, but I
> >haven't found out what it does on the primary head. Just wanted to
> >report that something like this would work for CRTC2, and Claudio should
> >consider this (please ;-) !).
>
> Yet, I don't understand why setting color adjustments to 0x8000 on start
> may cause problems with the CRTC2.
> Ville, could you explain me, please?
Because CRTC2 default color adjustment values are not 0x8000.
>From matrox_crtc2.c:
if (mdev->g450_matrox) {
adjustment->brightness = ntsc ? 0xAA00 : 0x9E00;
adjustment->saturation = ntsc ? 0xAE00 : 0xBB00;
} else {
adjustment->brightness = ntsc ? 0xB500 : 0xA800;
adjustment->saturation = ntsc ? 0x8E00 : 0x9500;
}
adjustment->contrast = ntsc ? 0xEA00 : 0xFF00;
adjustment->hue = 0x0000;
Those are the values used by the Windows driver btw.
You should query the default values when df_xine is started and use those
when a reset is requested.
> >However, I have other issues with df_xine on CRTC2. As the CRTC2 is
> >outputting an interlaced signal for a normal TV (PAL 4:3 in my case),
> >I'm expecting df_xine to be able to play interlaced material (like I
> >have from several DVDs, or with DVB material when using df_xine as a
> >viewer for the VDR-xine plugin) on the CRTC2 by just leaving it
> >interlaced and just take care of the field parity order I'm specifying,
> >as the output device, my TV, is interlaced too, anyway. Nevertheless,
> >there is no difference in the specified field parity option in df_xine,
> >the picture has wrong field order specific artifacts, especially
> >viewable on quick horizontal movements, scrolling text banners for
> >example). Such a thing does not happen with mplayer/dfbmga for example.
> >I even tried single/double/triple buffering, with and without V-Sync,
> >but there is no change. Is there something specific I could try, or is
> >this part of the code not yet working as expected?
>
> Currently the -f,--fieldparity option takes only effect when
> you specify -d,--deinterlace.
It looks like your code is wrong. You enable DLOP_DEINTERLACING when in
fact you should enable DLOP_FIELD_PARITY.
--
Ville Syrj�l�
[EMAIL PROTECTED]
http://www.sci.fi/~syrjala/
_______________________________________________
directfb-dev mailing list
[email protected]
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
