[vdr] Re: Re: vdr suggestions & osd colour oddity (Chris
Elsworth) (Stefan Lucke)
stefan at lucke.in-berlin.de
Tue Aug 8 07:56:09 CEST 2006
On Dienstag 08 August 2006 03:56, CR wrote:
> > > > Can you take a look at:
> > > >
> > > > http://www.flickr.com/photos/16371732@N00
> > @CR, I've seen this but I don't know whats the problem with your setup or
> > why our osd drawing is wrong in combination with nvidia cards.
> > At which color depth is your X server running ?
> Stefan, I am running at "24 bpp", 800x600.
Is this 24 bpp with 3 bytes or 24 bpp with 4 bytes ?
Output of xdpyinfo will clear this.
> Can you shed some light on how
> softdevice does it's alpha blending? I assume the decoded video frame is
> in non-RGB color space and from what I read about VDR's OSD that is in
> RGB, right?
Yes. Video is in fourcc YV12 and vdr OSD is RGB via lookup table.
> So, does softdevice do the RGB-->YUV for the OSD and then alpha-blend that
> into the video frame?
And that way, mixing an OSD converted to YV12 color space with YV12 video
(OSD drawing set to software) works with Matrox, Radeon and
VIA-Unichrome with "-vo xv:". In case of Radeon (mentioned as my
current setup) our OSD mixing method works with "-vo vidix:" too.
So its either a choice of software or hardware :-( .
software: use vdr-xine or xinelibout,
hardware: use cards that work: matrox, radeon, Unichrome or savage (with a small
change which is a work-around of a driver bug:
> Supposedly using Cairo/XRender, alpha-blending can be done by the video
> card hardware, I wonder how hard it would be to modify softdevice for
More information about the vdr