[mythtv-users] Problem with new HD-PVR and 1080i

Keith Pyle kpyle at austin.rr.com
Sun Aug 26 17:18:58 UTC 2012


On 08/26/12 07:00, mythtv-users-request at mythtv.org wrote:
> I have just added a new HD-PVR (1 day out of the box) and am not able to
> use it at all in 1080i mode, or reliably in 720p.  I'm aware of the
> stability issues many have reported and I've seen messages indicating
> people are using a HD-PVR at 1080i.  I've reviewed the MythTV wiki page
> for the HD-PVR and its troubleshooting tips without resolution, and
> Google has not been my friend in delivering any hints.
>
> OS: Gentoo 3.4.7
> Myth:
>
>      MythTV Version   : v0.24.1-27-g30993d6
>      MythTV Branch    : fixes/0.24
>      Network Protocol : 63
>      Library API      : 0.24.20110505-1
>      QT Version       : 4.7.2
>
> HD-PVR hardware:
>
>      - hardware is 49001 LF, rev F2
>      - located on an open, well ventilated shelf (nothing directly above
>      or below, adequate side clearance) in an air conditioned room
>      - top center of the case is 92F when not recording and 94F when
>      recording
>
> Set top box: Scientific Atlanta 4240HDC
> Connections: Component video, RCA audio
>
> 1. With HD-PVR firmware 0x1e (driver package 1.7.1.30059), the HD-PVR
> can accept 1080i input, I can see good video in mplayer, but myth
> doesn't work:
>
>      a. mplayer - works with input allowed to switch between 720p and
>      1080i as determined by source program
>      b. mythtv - gives the following errors, generates a 0 byte
>      recording, and the mythbackend process goes to 100% CPU for the
>      duration of the recording period:
>
>      2012-08-25 11:10:09.885 DevRdB(/dev/hdpvr0) Error: Poll giving up
>      2012-08-25 11:10:09.888 MPEGRec(/dev/hdpvr0) Error: Device error
>      detected
>      2012-08-25 11:10:09.889 DevRdB(/dev/hdpvr0): Stop(): Not running.
>      2012-08-25 11:10:25.736 DevRdB(/dev/hdpvr0) Error: Poll giving up
>      2012-08-25 11:10:25.740 MPEGRec(/dev/hdpvr0) Error: Device error
>      detected
>      2012-08-25 11:10:25.742 DevRdB(/dev/hdpvr0): Stop(): Not running.
>      2012-08-25 11:10:38.152 MPEGRec(/dev/hdpvr0) Error: StartEncoding failed
>                   eno: Resource temporarily unavailable (11)
>
>
> 2. Based on some postings from earlier this year that referenced similar
> messages, I switched to firmware 0x15 (driver package 1.5.7.0) and now
> 1080i does not work at all.  Both mplayer and myth work for 720p input,
> but not reliably.
>
>      a. mplayer - works only if input is fixed to 720p; will not work
>      with input floating between 720p and 1080i*or*  fixed to 1080i.
>      mplayer gives continuous errors like this for attempts to use 1080i
>      input:
>
>      [h264 @ 0x7f36105581e0]sps_id out of range
>      [h264 @ 0x7f36105581e0]non-existing SPS 32 referenced in buffering
>      period
>      [h264 @ 0x7f36105581e0]non-existing PPS 0 referenced
>      [h264 @ 0x7f36105581e0]decode_slice_header error
>      [h264 @ 0x7f36105581e0]no frame!
>      Error while decoding frame!
>
>      b. mythtv - starts recording with input to HD-PVR fixed to 720p, but
>      generally fails in mid-program with the same sort of messages shown
>      in 1b above and the 100% CPU use.
>
> To resolve the 100% CPU spin on mythbackend, I have to stop and restart
> it.  Stopping the recording does not reduce the CPU load. (The process
> is attempting to sleep in QWaitCondition but is waking up instantly.)
>
> So, I do not presently have a configuration that works with 1080i, the
> HD-PVR, and myth.
>
> What versions of mythtv and firmware are known to work?  Are there any
> particular configuration tricks to making it work?
Additional testing has ruled out a few causes.

1. With mythbackend stopped and using only cat to capture from the video 
device, I've tested all four combinations of resolution (720p & 1080i) 
and the two firmware versions (0x15 & 0x1e).  The 720p captures will run 
briefly.  The shortest was just over 1 minute and the longest was 11 
minutes.  This does produce watchable video. The 1080i tests fail in 
about the same amount of time.  The 1080i video is not playable on 
Linux, but is playable with Windows media player.  mplayer shows:
Error while decoding frame!
[h264 @ 0x7fe2822551e0]cpb_count 33 invalid
[h264 @ 0x7fe2822551e0]sps_id out of range
[h264 @ 0x7fe2822551e0]non-existing SPS 0 referenced in buffering period
[h264 @ 0x7fe2822551e0]non-existing PPS 0 referenced
[h264 @ 0x7fe2822551e0]decode_slice_header error
[h264 @ 0x7fe2822551e0]no frame!
Error while decoding frame!
[h264 @ 0x7fe2822551e0]sps_id out of range
[h264 @ 0x7fe2822551e0]non-existing SPS 15 referenced in buffering period
[h264 @ 0x7fe2822551e0]non-existing PPS 0 referenced
[h264 @ 0x7fe2822551e0]decode_slice_header error

When the capture stops producing data to cat, the HD-PVR appears hung 
(red capture LED on) and must be power cycled.

This rules out myth as a proximate cause.

2. I connected the HD-PVR to a Windows 7 system on which I installed the 
Hauppauge provided Arcsoft application.  This test used all of the same 
component, audio, and USB cabling used in test #1. Firmware was 0x1e.  
30-minute captures at 720p and 1080i were both successful and the files 
were playable on the Windows system. However, the 1080i capture was 
*not* playable on Linux using mplayer (same problem as noted above).

This likely rules out HD-PVR power or temperature issues, and cable 
problems.

The remaining possibilities for failing to capture on Linux would seem 
to be:

1. the USB port on the mythbackend system - the port used thus far is 
USB 2.0 on an unused bus (not shared with other USB devices)
2. the Linux USB subsystem - not very likely, IMO
3. the hdpvr driver in Linux 3.4.7

Other than trying a different USB port, I don't see many options for 
success.  I may have a gently used HD-PVR for sale if this fails.

Keith
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mythtv.org/pipermail/mythtv-users/attachments/20120826/9c6c5791/attachment.html>


More information about the mythtv-users mailing list