[vdr] Re: [ANNOUNCE] vdr-xine-0.4.2 plugin

[Reinhard Nissl <rnissl@gmx.de>]

Hi,

Luca Olivetti wrote:

> > A first try to solve the issue might be the following patch:
> >
> >     http://home.vr-web.de/~rnissl/vdr-xine-0.4.2-1.patch.gz
> >
> > It reduces the number of dots written to console to only two (the 
> > first and the last one). Furthermore, it sends no more than 20 speed 
> > changes to xine. These fixes should reduce CPU load dramatically.
> No luck, it still does the same.
As far as I have examined the issue, it seems that sometimes (or quite often) 
xine's metronom doesn't synchronize to the new stream, or at least it 
synchronizes to late. This feature seems to depend on the CPU load too.

The result is that there are frames on the video/audio output fifo where the 
difference of the timestamps between two consecutive frames is atypical large. 
Waiting for the metronom to reach the next timestamp produces the observed 
behaviour.

So I must find a solution, that always leads to synchronization of xine's 
metronom. Until then, it's even likely that release 0.4.1 shows a similar 
behaviour from time to time.

Bye.
--
Dipl.-Inform. (FH) Reinhard Nissl
mailto:rnissl@gmx.de