[mythtv-users] HD-PVR Audio Sync Issues After Channel Change

[Alan Young]

On 04/12/2010 04:11 PM, John P Poet wrote:
> On Mon, Apr 12, 2010 at 4:56 PM, Alan Young<ayoung at teleport.com>  wrote:
>>
>> The patches in the wiki are a band-aid.  They stop and restart the recorder
>> after the channel change.  It flushes the HDPVR buffers of data and any
>> pending data on the PC.  That's why back to back recordings end up missing
>> several seconds.
>
> What those signal monitor patches do:
>
> 1) Tell the HD-PVR to start recording
> 2) Test the output of the HD-PVR to see if a valid resolution is being reported
> 3) If a valid resolution is not reported, or if data is not
> forthcoming in a timely manner then reset the HD-PVR driver and goto
> step 1
> 4) Tell the HD-PVR to stop recording, and continue on to the actual
> recording process.
>
> As part of the process, the "state" is reported to the frontend, so
> the user can see what is going on and abort if there is a serious
> problem.  This also prevents the frontend from timing out, if the
> channel change process takes more than 15 seconds.  The user can abort
> LiveTV or try another channel at any time during the process.
>
> Because the signal monitor patches effectively try to briefly record
> something, someone watching the status light on the HD-PVR would see
> it record, stop, and then record again.  The idea is, that when the
> actual recording process starts up, there is a very high chance that
> the STB and HD-PVR are in a happy state.
>
> Even without those patches, Myth's HD-PVR recording process will issue
> "stop recording"/"start recording" commands to the HD-PVR driver, if
> it does not see a steady stream of data from the HD-PVR

Yes, but something doesn't seem right because even when it has a valid 
resolution it still stops and starts the recorder.  I see that on back to back 
recordings of the Daily Show and The Colbert Report.  And I've tweaked my 
channel change script to remember the current channel and not change it if 
it's the same channel.  So in theory the signal should not be interrupted much.

Looking at the old copy of the patch I have (says it's v2), i think it should 
start at m_stage==3 first to see if the signal's valid then if that's not 
true, try a start and stop.

I also have ATSC card in this box, so I can test with that too.  And if need 
be, I can put a PVR150 back in to test too.

Alan