[mythtv] [patch] AC3 patches

[Mark Anderson]

On Monday 07 February 2005 08:21, Mark Anderson wrote:
> On Sunday 06 February 2005 15:30, Isaac Richards wrote:
> > On Monday 31 January 2005 04:28 am, Mark Anderson wrote:
> > > On Monday 31 January 2005 16:53, Isaac Richards wrote:
> > > > On Saturday 29 January 2005 05:26 pm, Mark Anderson wrote:
> > > > > For those following the AC3 audio saga, here is a patch against the
> > > > > CVS from yesterday to get AC3 audio up and running with DVB. This
> > > > > patch is different to previous patches in a number of ways, mostly
> > > > > I have changed things after advice from Isaac.
> > > >
> > > > Before I apply this - could you test moving the av_remove_stream to
> > > > the Reset() function instead of SeekReset()?  That will only get
> > > > called when channel/inputs change, instead of on every seek..
> > >
> > > Made the changes and everything seems to function correctly.
> > >
> > > I have included a new version diffed against CVS, this removes the
> > > change in sitypes.cpp as indicated by Jesper. I also added a couple of
> > > comments above the av_remove_stream code block.
> >
> > I've just reverted the av_remove_stream part of this patch.  It's causing
> > normal XvMC to die on a channel change, and I haven't had time to track
> > down why.  Can you come up with a better way to handle this?
> >
> > Isaac
>
> Attached is a new version of the patch that does not clear out the video
> stream on a channel change, it only removes audio streams. Could someone
> who can reproduce this problem apply the patch (to the reverted CVS) and
> see if the channel change problem comes back. Also could someone confirm if
> the problem is exhibited when using TS rather than PS recording?
>
> I just tested this patch with my system running nvidia XvMC and am able to
> change channels fine. I have an EPIA but haven't got around to setting up
> XvMC so am not in a position to test the unichrome XvMC at the moment.
>
I have now tested the fix with via vldXvMC and it seems to work fine. With the 
original av_remove_stream calls there was garbage on the screen after 
changing channels, I then addeed the test to only remove audio streams and 
changing channels then worked fine.

I can make a clean patch if needed. 

Cheers,
Mark