[confirm] Random glitches in LiveTV ("Unexpected LiveTV PAT change" log) (1 Viewer)

peque

Moderator - Spanish Forums
  • Premium Supporter
  • August 4, 2007
    861
    99
    Home Country
    Spain Spain
    • Thread starter
    • Moderator
    • #21
    Hi all,

    I was investigating into old bug reports that could help me in this case... and I've come across this one:

    https://forum.team-mediaportal.com/...96/spanish-tswriter-problem-78380/index2.html

    After reviewing all this issue, I can confirm I'm experiencing exact same issue (in fact, is the same spanish DVB-t multiplex also, and others). First of all sorry, because first time I looked into my logs I dind't see anything in tswriter.log, and that's why I concentrated all my efforts and findings in tsreader... BUT, I've gone to 1.1.0 RC2 again, and I can see tswriter logs regarding pmt changes:

    Code:
    26-04-2010 15:45:07.153 pmtgrabber: got new pmt version:1f 1e, service_id:8c
    26-04-2010 15:45:07.153 pmtgrabber: do callback pid 12c
    26-04-2010 15:45:07.161 analyzer: set video pid:12d
    26-04-2010 15:45:07.161 analyzer: reset
    26-04-2010 15:45:07.163 analyzer: set audio pid:12f
    26-04-2010 15:45:07.163 analyzer: reset
    26-04-2010 15:45:07.167 Recorder: TIMESHIFT paused=yes
    26-04-2010 15:45:07.167 Recorder: TIMESHIFT Received from TvService: pmt pid:0x12c serviceId: 0x8c pmtlength:369
    26-04-2010 15:45:07.167 Recorder: TIMESHIFT Old pids cleared
    26-04-2010 15:45:07.168 Recorder: TIMESHIFT got pmt - tableid: 0x2 section_length: 369 sid: 0x8c
    26-04-2010 15:45:07.168 Recorder: TIMESHIFT PMT parsed  - Pid 0x12c ServiceId 0x8c stream count: 13
    26-04-2010 15:45:07.168 Recorder: TIMESHIFT pcr pid:0x12d
    26-04-2010 15:45:07.168 Recorder: TIMESHIFT SetPcrPid clear old PIDs
    26-04-2010 15:45:07.168 Recorder: TIMESHIFT determine new start pcr
    26-04-2010 15:45:07.168 Recorder: TIMESHIFT add video stream pid: 0x12d fake pid: 0x30 stream type: 0x2 logical type: 0x2 descriptor length: 3
    26-04-2010 15:45:07.169 Recorder: TIMESHIFT add teletext stream pid: 0x12e fake pid: 0x68 stream type: 0x6 logical type: 0x56 descriptor length: 48
    26-04-2010 15:45:07.169 Recorder: TIMESHIFT add audio stream pid: 0x12f fake pid: 0x40 stream type: 0x3 logical type: 0x3 descriptor length: 6
    26-04-2010 15:45:07.169 Recorder: TIMESHIFT add audio stream pid: 0x130 fake pid: 0x41 stream type: 0x3 logical type: 0x3 descriptor length: 6
    26-04-2010 15:45:07.169 Recorder: TIMESHIFT add subtitle stream pid: 0x137 fake pid: 0x50 stream type: 0x6 logical type: 0x6 descriptor length: 10
    26-04-2010 15:45:07.169 Recorder: TIMESHIFT stream rejected - pid: 0x28a stream type: 0x5 logical type: 0x5 descriptor length: 5
    26-04-2010 15:45:07.169 Recorder: TIMESHIFT stream rejected - pid: 0x28b stream type: 0xb logical type: 0xb descriptor length: 49
    26-04-2010 15:45:07.170 Recorder: TIMESHIFT stream rejected - pid: 0x28c stream type: 0x5 logical type: 0x5 descriptor length: 5
    26-04-2010 15:45:07.170 Recorder: TIMESHIFT stream rejected - pid: 0x28d stream type: 0xb logical type: 0xb descriptor length: 51
    26-04-2010 15:45:07.170 Recorder: TIMESHIFT stream rejected - pid: 0x28e stream type: 0x5 logical type: 0x5 descriptor length: 5
    26-04-2010 15:45:07.170 Recorder: TIMESHIFT stream rejected - pid: 0x28f stream type: 0xb logical type: 0xb descriptor length: 49
    26-04-2010 15:45:07.170 Recorder: TIMESHIFT stream rejected - pid: 0x290 stream type: 0x5 logical type: 0x5 descriptor length: 5
    26-04-2010 15:45:07.170 Recorder: TIMESHIFT stream rejected - pid: 0x291 stream type: 0xb logical type: 0xb descriptor length: 49
    26-04-2010 15:45:07.171 Recorder: TIMESHIFT paused=no
    26-04-2010 15:45:07.184 Recorder: TIMESHIFT start of video detected
    26-04-2010 15:45:07.185 Recorder: TIMESHIFT clear TS packet queue
    26-04-2010 15:45:07.185 Recorder: TIMESHIFT Info : Next broadcaster program clock reference rollover : 0 days 16:43:42 0
    26-04-2010 15:45:07.263 Recorder: TIMESHIFT start of audio detected
    26-04-2010 15:45:07.273 Recorder: TIMESHIFT start of audio detected

    What's more, I've used tswriter provided in that post, with no pmt change detection, and I don't have any stutter or glitch after more than an hour of livetv in conflicting multiplex.

    Now my questions:
    - is Tswriter in that post 100% compatible with RC2, or I must ask for a RC2 compatible tswriter with this feature deactivated?
    - Is there any mantis open about this? What's more... is there any solution for our case, is a real bug? Any way to rescan PMT table with no glitch (I assume it's not possible)? Are we condemned to ask for compatible tswriter for each new MP TV Server release? Could it be configurable?

    I repeat: No problem to record entire .ts and upload it so you can test this... In fact, I'll do it anyway.

    Thanks so much for your patience.

    EDIT: Perfect. I could grab entire ts while one of these pat change, using TV Server debug option. 1 Gb file. You'll have it uploaded to your ftp

    Address: ftp.team-mediaportal.com
    Login: tsdump@team-mediaportal.com
    Password: mpuser88

    tomorrow morning.

    bye!
     

    peque

    Moderator - Spanish Forums
  • Premium Supporter
  • August 4, 2007
    861
    99
    Home Country
    Spain Spain
    • Thread starter
    • Moderator
    • #22
    Oops... it seems ftp has no "append" (resume) function enabled... Upload has stopped (I don't know why) and I can't continue from that point... So you have a 585 Mb incomplete file... If any here can erase this file:

    Spain_Telecinco_850MHz_typeT_EntireTS_FTA_AboutPMTChangeVersion.ts

    It's useless... I'll try it again tomorrow morning.

    Thanks a lot.
     

    morpheus_xx

    Retired Team Member
  • Team MediaPortal
  • March 24, 2007
    12,073
    7,459
    Home Country
    Germany Germany
    AW: Random glitches in LiveTV ("Unexpected LiveTV PAT change" log)

    Hi,

    I started some modifications on TsWriter code:

    The callback to tell tvserver about new PMT versions does now a check if the a/v pids have changed. If all are still equal, no callback is done.

    Please test this attached tswriter (backup before!) if this already fixes the recording issue. It maybe also fixes some playback issues.

    If playback is not yet properly, I have to adopt the same code changes to TsReader side.

    Please give feedback!

    Morpheus

    edit: removed attachment, try newer build on next page!
     

    iosub

    Portal Pro
    April 26, 2006
    573
    30
    San Sebastian
    Re: AW: Random glitches in LiveTV ("Unexpected LiveTV PAT change" log)

    Hi,

    I started some modifications on TsWriter code:

    The callback to tell tvserver about new PMT versions does now a check if the a/v pids have changed. If all are still equal, no callback is done.

    Please test this attached tswriter (backup before!) if this already fixes the recording issue. It maybe also fixes some playback issues.

    If playback is not yet properly, I have to adopt the same code changes to TsReader side.

    Please give feedback!

    Morpheus

    Thanks!! just instaled.. will report tomorrow

    Thanks!
     

    peque

    Moderator - Spanish Forums
  • Premium Supporter
  • August 4, 2007
    861
    99
    Home Country
    Spain Spain
    • Thread starter
    • Moderator
    • #26
    Re: AW: Random glitches in LiveTV ("Unexpected LiveTV PAT change" log)

    Hi,

    I started some modifications on TsWriter code:

    The callback to tell tvserver about new PMT versions does now a check if the a/v pids have changed. If all are still equal, no callback is done.

    Please test this attached tswriter (backup before!) if this already fixes the recording issue. It maybe also fixes some playback issues.

    If playback is not yet properly, I have to adopt the same code changes to TsReader side.

    Please give feedback!

    Morpheus
    Thanks so much Morpheus. Should we see any change in log files? Or behaviour is the same in log? I'm not at home now, so I can only switch on TV and see logs, till lunch time... :). I'll report back in some minutes...

    bye!

    EDIT: Good news... :) This is piece of new tswriter log of a pmt change:

    Code:
    29-04-2010 09:47:31.140  pcr      pid:   cb 
    29-04-2010 09:47:31.140  pmt      pid: ffffffff 
    29-04-2010 09:47:31.141  video    pid:   cb type: MPEG2
    29-04-2010 09:47:31.142  audio    pid:  12f language: spa type: MPEG2 - audio
    29-04-2010 09:47:31.142  audio    pid:  132 language: mul type: MPEG2 - audio
    29-04-2010 09:47:31.142  pcr      pid:   cb 
    29-04-2010 09:47:31.143  pmt      pid: ffffffff 
    29-04-2010 09:47:31.143  video    pid:   cb type: MPEG2
    29-04-2010 09:47:31.143  audio    pid:  12f language: spa type: MPEG2 - audio
    29-04-2010 09:47:31.143  audio    pid:  132 language: mul type: MPEG2 - audio
    29-04-2010 09:47:31.144 pmtgrabber: got new pmt version:2 1, service_id:b4
    29-04-2010 09:47:31.144 pmtgrabber: NO callback done because a/v pids still the same.

    This doesn't trigger any equal tsreader log, so I assume there's no glitch (I will confirm at lunch time) :).

    Gooooodddd! :)
     

    iosub

    Portal Pro
    April 26, 2006
    573
    30
    San Sebastian
    Re: AW: Random glitches in LiveTV ("Unexpected LiveTV PAT change" log)

    Hi,

    I started some modifications on TsWriter code:

    The callback to tell tvserver about new PMT versions does now a check if the a/v pids have changed. If all are still equal, no callback is done.

    Please test this attached tswriter (backup before!) if this already fixes the recording issue. It maybe also fixes some playback issues.

    If playback is not yet properly, I have to adopt the same code changes to TsReader side.

    Please give feedback!

    Morpheus
    Thanks so much Morpheus. Should we see any change in log files? Or behaviour is the same in log? I'm not at home now, so I can only switch on TV and see logs, till lunch time... :). I'll report back in some minutes...

    bye!

    EDIT: Good news... :) This is piece of new tswriter log of a pmt change:

    Code:
    29-04-2010 09:47:31.140  pcr      pid:   cb 
    29-04-2010 09:47:31.140  pmt      pid: ffffffff 
    29-04-2010 09:47:31.141  video    pid:   cb type: MPEG2
    29-04-2010 09:47:31.142  audio    pid:  12f language: spa type: MPEG2 - audio
    29-04-2010 09:47:31.142  audio    pid:  132 language: mul type: MPEG2 - audio
    29-04-2010 09:47:31.142  pcr      pid:   cb 
    29-04-2010 09:47:31.143  pmt      pid: ffffffff 
    29-04-2010 09:47:31.143  video    pid:   cb type: MPEG2
    29-04-2010 09:47:31.143  audio    pid:  12f language: spa type: MPEG2 - audio
    29-04-2010 09:47:31.143  audio    pid:  132 language: mul type: MPEG2 - audio
    29-04-2010 09:47:31.144 pmtgrabber: got new pmt version:2 1, service_id:b4
    29-04-2010 09:47:31.144 pmtgrabber: NO callback done because a/v pids still the same.

    This doesn't trigger any equal tsreader log, so I assume there's no glitch (I will confirm at lunch time) :).

    Gooooodddd! :)

    no more glitch here. :) will test more tonight. thanks
     

    cfforce

    MP Donator
  • Premium Supporter
  • March 4, 2008
    241
    21
    Home Country
    Netherlands Netherlands
    I have a working system, and wanted to test this also out; even if i had no problems; but what i noticed at start(~1min in) when all my 3 tuners where running (recording/timeshifting) tvservice crashed; happend 5 times; when i put the old tswriter.ax back; all was ok. I'll post logs @ later time (tested on 25359 Disaster-ed.)
     

    peque

    Moderator - Spanish Forums
  • Premium Supporter
  • August 4, 2007
    861
    99
    Home Country
    Spain Spain
    • Thread starter
    • Moderator
    • #29
    It's a little strange, that so tinny change can break tswriter this way in your system... ¿?
     

    morpheus_xx

    Retired Team Member
  • Team MediaPortal
  • March 24, 2007
    12,073
    7,459
    Home Country
    Germany Germany
    AW: Random glitches in LiveTV ("Unexpected LiveTV PAT change" log)

    The change itself needed quite many code changes: I had to adopt 2 classes from tsreader and use them. So it can be indeed the source of hard crashes!

    A way to find to source, would be the tswriter.log. It maybe tells us more...
     

    Users who are viewing this thread

    Top Bottom