[mythtv] 0.18.2, IOBOUND, ringbuf, and Live TV Jitter
Dave Sp
dspmisc at speakeasy.net
Sat Jul 23 22:02:51 EDT 2005
Lutz,
Sorry -- I applied both of your suggested changes and the issue remained.
The jitters only subsided after the size of the ringbuffer stabilized.
So my original question remains.
--davesp
Lutz Mändle wrote:
> Hi,
> sometimes ago I played around with the following settings in
> RingBuffer.cpp to have better result using mythfrontend via 54Mbit
> wireless LAN.
>
> //#define TFW_DEF_BUF_SIZE (2*1024*1024)
> //#define TFW_MAX_WRITE_SIZE (TFW_DEF_BUF_SIZE / 4)
> //#define TFW_MIN_WRITE_SIZE (TFW_DEF_BUF_SIZE / 8)
> #define TFW_DEF_BUF_SIZE (64*1024*1024)
> #define TFW_MAX_WRITE_SIZE (TFW_DEF_BUF_SIZE / 4)
> #define TFW_MIN_WRITE_SIZE (TFW_DEF_BUF_SIZE / 128)
>
> The comment lines are the original settings, with these new settings
> I've got good results for recorded programs, but Live-TV had the old
> behaviour.
>
> After searching the code I've found this in tv_rec.cpp:
>
> void TVRec::SetupRingBuffer(QString &path, long long &filesize,
> long long &fillamount, bool pip)
> {
> ispip = pip;
> filesize = gContext->GetNumSetting("BufferSize", 5);
> fillamount = gContext->GetNumSetting("MaxBufferFill", 50);
>
> path = gContext->GetSetting("LiveBufferDir") + QString("/ringbuf%1.nuv")
>
> .arg(m_capturecardnum);
>
> outputFilename = path;
>
> filesize = filesize * 1024 * 1024 * 1024;
> fillamount = fillamount * 1024 * 1024;
>
> rbuffer = new RingBuffer(path, filesize, fillamount);
> rbuffer->SetWriteBufferMinWriteSize(1); //<<- these line is now
> commented out
> }
>
> After commenting out the last line in this function, Live-TV over
> wireless LAN works reasonable.
>
> Maybe this helps.
>
> Und Tschüß...
> Lutz Mändle Internet: lmaendle at gmx.net
> Berlin/Germany
>
> _______________________________________________
> mythtv-dev mailing list
> mythtv-dev at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
-----
Original Email:
Regarding the various emails about video jitter...
I see this issue while the file ringbuf1.nuv is growing. Once the file size
stabilizes, I stop seeing IOBOUND errors.
This didn't happen to me when I was running 0.16. Nothing else on my system
changed since then. (DMA has always been enabled, and the ext3 filesystem
disks worked fine back then.)
The IOBOUND and video jitter issues are not seen during scheduled
recordings, so I suspect something is odd about the live TV and/or ring
buffer code.
Is there any debug data I can collect to help resolove this issue?
--davesp
More information about the mythtv-dev
mailing list