[mythtv] Re:jerky live tv with latest cvs(5/28)

Keith Gross grossk at mail.tds.net
Sat May 31 01:16:37 EDT 2003


I've just started using MythTV and I'm also encountering a variations of this 
problem.  I can watch TV just fine with occasional rebuffering messages and 
the accompanying jerk in the video.  But than every now and than :

- I receive a steady stream of rebuffering messages like the following.

rebuffering (150529 256000)
rebuffering (155649 256000)
rebuffering (128001 256000)
rebuffering (133121 256000)
rebuffering (128001 256000)
rebuffering (133121 256000)
rebuffering (138241 256000)
rebuffering (143361 256000)
rebuffering (148481 256000)
rebuffering (153601 256000)
rebuffering (158721 256000)
rebuffering (131073 256000)
rebuffering (136193 256000)

- During this time video either freezes or plays just an occasional frame.  
Sometimes the occasion CPU average jump from 10% - 20% up to 50%+.  Running 
top shows that X is concuming most of the CPU time.

- In at least one case pressing ESC caused MythTV to disappear from TOPs 
display but the frozen video frame was still visible and a entry for MythTV 
was sill listed in KDE's task bar.  When I right clicked the task bar icon 
and chose close the frozen frame disappeared and X CPU consumption dropped.

I've also had one incident of the end of taping causing the entire machine to 
freeze solid.  I could even ping it any more.  The version I'm running was a 
CVS checkout from the afternoon of the 28th.  Anyone have any insight on 
these issues yet?


On Friday 30 May 2003 03:44 pm, Vinton Coffman wrote:
> I too experience the live tv rebuffering issues.  I'm using cvs as of
> 7:00 p.m. 5/29.  I've been having this issue since I installed mythtv
> several weeks ago.  I originally posted the message below over the
> holiday weekend on the user list but didn't get any responses.  So I
> was using cvs versions before 5/27.
>
> Nobody mentioned what capture hardware they're using.  Can you post
> that info?  I'm wondering if this is only being exhibited on pvr-250
> based systems.
>
> I haven't completely ruled out hardware issues but Kingston just
> swapped the memory I was using and it doesn't seem to change anything.
> It rebuffers at random intervals, sometimes recovers, sometimes locks
> the frontend, and sometimes locks the entire system.
>
> When I looked at RingBuffer.cpp I noticed the section of code below.  I
> am not a developer on the project or all that intimate with this code.
> So here goes nothing.  Does the fact I'm using the pvr-250 at the
> default bitrate 8/16Mb/s cause an issue with this piece of code?  Is
> the ringbuffer not reading far enough ahead?
>
> void RingBuffer::CalcReadAheadThresh(int estbitrate)
> {
>      wantseek = true;
>      pthread_rwlock_wrlock(&rwlock);
>      wantseek = false;
>
>      fill_threshold = 0;
>      fill_min = 0;
>
>      if (remotefile)
>          fill_threshold += 256000;
>
>      if (estbitrate > 6000)
>          fill_threshold += 256000;
>
>      if (estbitrate > 10000)
>          fill_threshold += 256000;
>
>      if (estbitrate > 14000)
>          fill_threshold += 256000;
>
>      if (fill_threshold == 256000)
>          fill_min = 128000;
>      if (fill_threshold > 256000)
>          fill_min = 256000;
>
>      readsallowed = false;
>
>      if (fill_threshold == 0)
>          fill_threshold = -1;
>      if (fill_min == 0)
>          fill_min = -1;
>
>      pthread_rwlock_unlock(&rwlock);
> }
>
>  >>>>>> Original message from May 25  follows <<<<<<<<<
>
>   This may be a hardware issue that I'm trying to chase down. I have a
> Athlon 2500 and PVR-250 running RH9 with kernel 2.4.21rc2 that seems to
> lock up completely with the FSB set to 166 MHz and doesn't crash with
> FSB at 133 MHz. Memory bus is 333 MHz with PC2700 dimms speed set by
> SPD in bios in both cases. With the FSB at 133 things don't lock up
> but my audio starts stuttering and video freezes. It usually recovers
> after a few seconds or up to a minute. This happens at random
> intervals and duration.
>
> When I check the output of mythfrontend I get the messages shown below.
> I have an 120 GB ata100 seagate drive and hdparm reports dma is on.
> Top reports mythfrontend is only using 17% of the CPU. The ringbuffer
> was set to 6GB but I can't seem to find this setting since the setup
> menus changed. Did I just overlook it or is not accessible from the
> new menus yet?
>
> The pvr-250 is still using the default settings for audio and video.
>
> I included the piece of code that generates the rebuffering message. I
> couldn't figure out what the variable ateof represented. Can someone
> explain what this code does? It might help me figure out what's going.
>
> TIA
>
> rebuffering (139265 256000)
> rebuffering (144385 256000)
> rebuffering (149505 256000)
>
>  >>>>>>>>> The message below doesn't always show up
>
> Changing from None to WatchingLiveTV
> Over/underscanning. V: 0.000000, H: 0.000000, XOff: 0, YOff: 0
> Using XV port 105
>
>
> Delaying to next trigger: 316423
> Delaying to next trigger: 267058
> Delaying to next trigger: 209758
> rebuffering (133121 256000)
> rebuffering (133121 256000)
>
>  >>>> output from hdparm
>
> [root at localhost mythtv]# /sbin/hdparm -d /dev/hda
>
> /dev/hda:
> using_dma = 1 (on)
>
>  >>>>> from RingBuffer.cpp
>
> if (ateof)
> {
> readsallowed = true;
> totfree = 0;
> }
>
> if (!readsallowed && used >= fill_threshold)
> readsallowed = true;
>
> if (readsallowed && used < fill_min && !ateof)
> {
> readsallowed = false;
> cerr << "rebuffering (" << used << " " << fill_min << ")\n";
> }
>
> _______________________________________________
> mythtv-dev mailing list
> mythtv-dev at snowman.net
> http://lists.snowman.net/cgi-bin/mailman/listinfo/mythtv-dev



More information about the mythtv-dev mailing list