[Softdevice-devel] Juddering OSD with mgatv output

Tomi Hautakoski thauta at iki.fi
Wed Mar 23 23:54:28 CET 2005


On Wed, 23 Mar 2005, Stefan Lucke wrote:
> So your setup works better than mine (G400+celeronIII 1.2GHz, very old dfb).
> Whats your CPU power ?

The same. I'm not sure if it was the new dfb or softdevice which made 
the difference but now I tested the system with live hockey (which seems 
to be good material for testing) and the picture was perfect.

> OSD drawing with dfb has to be rewritten, as it always draws complete osd
> if only a small area has to be redrawn. To my experince this effect seems
> to be worse with the channel info osd, which in term is refreshed for about
> every frame. Vdr main menu menu is only refreshed after a few seconds.

Yes, I have exactly the same symptoms here. I had to remove the EPG 
description field from the channel info and change the display time to 1 
sec to be able to change my channel without loosing the a-vsync 
completely. Still, nice to know that you have a hint what is causing the 
problem. =)

>> Maybe these are somehow related to the problem. Noticed from the logs
>> after a longer run:
>> alsa-audio: Xrun
>> [mpeg2video @ 0x407f56b0]00 motion_type at 16 15
>> [mpeg2video @ 0x407f56b0]MPEG motion vector out of boundary
>> [mpeg2video @ 0x407f56b0]MPEG motion vector out of boundary
>> [mpeg2video @ 0x407f56b0]MPEG motion vector out of boundary
>> [mpeg2video @ 0x407f56b0]MPEG motion vector out of boundary
>> [mpeg2video @ 0x407f56b0]ac-tex damaged at 1 16
>> [mpeg2video @ 0x407f56b0]Warning MVs not available
>> [mpeg2video @ 0x407f56b0]concealing 135 DC, 135 AC, 135 MV errors
>
> This is either related to a bad signal or some buffer problems.
> Do you get buffer overflow messages in syslog ?

Now that I know these don't have something to do with OSD I'm sure it's 
the fridge which makes the DVB-T signal go wild every time it starts. So 
we can ignore these. =)

Tomi Hautakoski
thauta at iki.fi
                            Queen Elizabeth rules, UK?



More information about the Softdevice-devel mailing list