[mythtv] SVN 8625 Firewire capture glitching while test-mpeg2 doesn't

[Steve Adeff]

On Saturday 21 January 2006 10:15, Steve Adeff wrote:
> On Saturday 21 January 2006 10:08, Steve Adeff wrote:
> > On Thursday 19 January 2006 15:42, Steve Adeff wrote:
> > > With Jim Westfall's help I've yet get a 0bit firewire recording, its
> > > worked everytime. What I notice though is a lot of glitches in the
> > > capture. Sample log output:
> > > 2006-01-19 14:46:12.625 audio waiting for space on soundcard: have 2752
> > > need 6144
> > > [mpeg2video @ 0x2aaaabffa4c0]MPEG motion vector out of boundary
> > > [mpeg2video @ 0x2aaaabffa4c0]MPEG motion vector out of boundary
> > > [mpeg2video @ 0x2aaaabffa4c0]Warning MVs not available
> > > [mpeg2video @ 0x2aaaabffa4c0]concealing 4265 DC, 4265 AC, 4265 MV
> > > errors 2006-01-19 14:46:12.640 audio waiting for space on soundcard:
> > > have 5760 need 6144
> > >
> > > depending on the glitch, I may receive between 4-40+ lines like this
> > > every 2minutes or so.
> > >
> > > As an experiement, I ran a test-mpeg2 capture on the same channel
> > > (though the channel doesn't matter), grabbed 1.8GB and played through
> > > the internal player using MythVideo. This method results in a glitch
> > > free capture for the 17mins of video.
> > >
> > > For MythTV I have the Firewire port set to record via "broadcast",
> > > previously I had it set to "p2p" as it is what worked when testing with
> > > test-mpeg2.
> > >
> > > Here's my plugreport for the capture from test-mpeg2:
> > > Node 2 GUID 0x000e5cfffed720aa
> > > ------------------------------
> > > oMPR n_plugs=1, data_rate=2, bcast_channel=63
> > > oPCR[0] online=1, bcast_connection=0, n_p2p_connections=1
> > >         channel=63, data_rate=2, overhead_id=0, payload=376
> > > iMPR n_plugs=0, data_rate=2
> > >
> > > So then I tried switching mythtv-setup back to a p2p connection to see
> > > if that fixes the glitches, but it still gets the glitches and doesn't
> > > alway get a stream.
> > >
> > > Now I've got mythtv-setup back to broadcast and its working like it was
> > > before, and for kicks did another test-mpeg2 capture this time on
> > > another channel, another 1.2GB and again using the internal player I
> > > get no glitches playing it back (verified by a search through the
> > > frontend log).
> > >
> > > so it looks like somethings wrong with how MythTV captures via the
> > > firewire, and I'll be happy to try and help find the problem if anyone
> > > has anything for me to try.
> >
> > Anybody? The glitches have also caused mythfrontend to crash occasionaly,
> > and I still can't get test-mpeg2 to make a bad capture.
>
> here's the log output from a crash, it occured while in edit mode:
>
> 2006-01-21 10:13:14.099 59520 bytes free on soundcard
> 2006-01-21 10:13:14.099 WriteAudio: Preparing 6144 bytes (1536 frames)
> 2006-01-21 10:13:14.107 54912 bytes free on soundcard
> 2006-01-21 10:13:14.115 56448 bytes free on soundcard
> 2006-01-21 10:13:14.126 58432 bytes free on soundcard
> 2006-01-21 10:13:14.126 NVP: ClearAfterSeek()
> 2006-01-21 10:13:14.126 VideoOutputXv: ClearAfterSeek()
> 2006-01-21 10:13:14.126 VideoBuffers::DiscardFrames(0):
> AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
> 2006-01-21 10:13:14.126 VideoBuffers::DiscardFrames(0):
> AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA -- done
> 2006-01-21 10:13:14.127 NVP: new commBreakIter = 4 @ frame 43142
> 2006-01-21 10:13:14.127 detectInterlace(Detect Scan, Interlaced Scan,
> 29.97, 480) ->Interlaced Scan
> 2006-01-21 10:13:14.127 Interlaced: Interlaced Scan  video_height: 480 
> fps: 29.97
>
> the crashes seem to be related to deinterlacing except that I don't have
> deinterlacing turned on....


could this be a 64bit kernel issue?

-- 
Steve