[mythtv] Weird framerates on CVS with DVB

[Bruce Markey]

Isaac Richards wrote:
> On Sunday 22 August 2004 08:25 pm, Andy McMullan wrote:
> 
>>>if (frameheader.frametype == 'S')
>>>{
>>>    if (frameheader.comptype == 'A')
>>>    {
>>>        if (frameheader.timecode > 0 && frameheader.timecode < 5500000)
>>>        {
>>>===>        effdsp = frameheader.timecode;
>>>            m_parent->SetEffDsp(effdsp);
>>>        }
>>>   }
>>>}
>>
>>BTW Commenting out the m_parent->SetEffDsp line is enough to let me
>>watch my 'bad' recordings.
> 
> 
> Interesting.  Nothing that I'm aware of has changed in that code at all 
> recently (or in the past year or so, even). =)  It shouldn't be tied to the 
> avcodec changes, as audio for software recordings is completely internalized 
> to myth.

This bug had been around for a long time. I'd been getting one
of these "bad" files once every two or three weeks. It was always
on the same machine so I'd always assumed that was an intermittent
hardware problem with the sound card on that machine. I've had
this debug statement:

             if (frameheader.comptype == 'A')
             {
+               VERBOSE(VB_AUDIO, QString("Audio timecode = %1")
+                       .arg(frameheader.timecode));
                 if (frameheader.timecode > 0 && frameheader.timecode < 5500000
)
                 {

for months and could see that on the "bad" files the timecodes
started near zero which was completely different than the "good"
files.

I have one of those "bad" files right now and can be the third
or fourth person ;-) to confirm that the change in current CVS
fixes this problem.

Thanks to each of you for tracking down this bug and squashing it!

--  bjm