I was thinking of giving xine/vdpau/xineliboutput another try after failing earlier in the year. I came across these patches:
http://www.nfvdr.net/bbs/thread-5856-1-1.html
Is anyone using these?
On Mon, 24 Aug 2009 12:49:18 +0100 scott scott@waye.co.uk wrote:
Looks like this: http://vdrportal.de/board/thread.php?threadid=86804
And yes, I use the patches and they improved playback considerably.
On Mon, 24 Aug 2009 18:48:21 +0200, Malte Schröder maltesch@gmx.de wrote:
Oh, yes its just a translation, thanks.
I take it you are using them on xinelib 1.2, are you using xineliboutput 1.0.4 or cvs?
-- Scott
On Tue, 25 Aug 2009 11:39:51 +0200 (CEST) "Gerald Dachs" vdr@dachsweb.de wrote:
I use CVS with only slight glitches (sometimes the picture is not deinterlaced properly) problems. But I must add, I only use it for PAL-TV with vdr-1.6.0-2.
Hi,
I am still using 1.6.0-2 and have patched it for FullHD OSD (http://www.vdrportal.de/board/thread.php?threadid=79647&threadview=0&... ). Frontend is xineliboutput-cvs (vdr-sxfe in other computer; tcp connection) and skinenigmang-0.1.0 is in use. Everything seems fine except that if I use antialiased fonts there is one problem that I have noticed only with epgsearch plugin "Schedule" view. The problem is that Schedule is not shown if the OSD should be really full. With short EPG info the schedule is shown.
The problem seems to be in buffering as I get following message in / var/log/messages: Aug 25 18:04:35 server vdr: [7573] [xine..put] write_osd_command: socket buffer full, OSD send skipped (got 262142 ; need 302366
256k seems to be the limit.
Is it the problem in vdr, xineliboutput, epgsearch or somewhere else? OK I understand that these versions were not designed for fullHD support but anyways it would be nice fix this.
I can avoid problem by disabling anti-alias fonts or setting "Try 8bpp single area: no" from skinenigmang options.
- Kimmo
Kimmo Taskinen wrote:
xineliboutput limits compressed OSD bitmap size.
OK I understand that these versions were not designed for fullHD support but anyways it would be nice fix this.
Does the attached patch fix the problem ?
- Petri