[Bug] TV: missing audio stream in records (Sky) (2 Viewers)

HTPCSourcer

Retired Team Member
  • Premium Supporter
  • May 16, 2008
    11,418
    2,335
    Home Country
    Germany Germany
    The above as stated by @Stéphane Lenclud has been discussed in depth at the beginning of this thread:

    Everything works fine if you start recording at a time where Sky is already broadcasting a multi-language stream, If, however, you begin your recording with sufficient pre-recording time while a mono-language stream is broadcasted, then the Sky switch to the multi-language screen is not effectively registered by MediaPortal and only the first language (German for Sky Germany) gets recorded.

    @mm1352000 , if your request is intended to generate logs for DigitalDevices, then we should already have plenty of them collected and uploaded with different answers in this thread. I rather see the risk of uploading "no issue" logs that will confirm DD's understanding that there actually is no issue. I believe that the only way of demonstrating that there is a problem is to repeat the recording of the pre-show mono-language setup. You may remember the discussion with @FreakyJ and myself.
     

    mm1352000

    Retired Team Member
  • Premium Supporter
  • September 1, 2008
    21,577
    8,224
    Home Country
    New Zealand New Zealand
    What difference do you make between decrypting and streaming?
    Streaming = timeshifting or recording.
    Decrypting = decrypting an encrypted channel for the purpose of timeshifting or recording.

    You can call it decrypting if you want. I don't think there is a meaningful/significant difference.

    How would that translate in terms of use case?
    It doesn't matter if you do the test with timeshifting or recording. Just as long as you test with two tuners, one CI/CAM (ie. MTD), at least two different encrypted channels, and you catch a dynamic PMT change. Timeshifting and recording give the same interactions with the CI/CAM, and that is the main thing that we need to worry about.

    For instance...
    Yes, I think that is okay... but to be very clear:
    Watch another channel until recording of the first one starts.

    1. There should be some time where both channels are being decrypted.
    2. The dynamic PMT change (audio stream added) should occur on the channel that you are watching, not the recording.
    3. After the new audio stream is added (assumption: the new stream remains encrypted = silent) happens you should stop viewing that channel, then restart it and confirm that the new audio stream is okay now.
    4. Finally, stop viewing the channel and stop the recording, then check that the recording is okay from start to finish.

    Make sense? :)
     

    mm1352000

    Retired Team Member
  • Premium Supporter
  • September 1, 2008
    21,577
    8,224
    Home Country
    New Zealand New Zealand
    The above as stated by @Stéphane Lenclud has been discussed in depth at the beginning of this thread...
    Sure, but this test case is slightly different than what was previously tested.

    Everything works fine if you start recording at a time where Sky is already broadcasting a multi-language stream,
    Agreed.

    If, however, you begin your recording with sufficient pre-recording time while a mono-language stream is broadcasted, then the Sky switch to the multi-language screen is not effectively registered by MediaPortal and only the first language (German for Sky Germany) gets recorded.
    It is registered fine by MediaPortal/TV Server. Evidence for that is the fact that the stream is included in the PMT. The issue is that the stream is not decrypted and that is why it is not recorded. Whether the fault is with the CI hardware, driver or CAM is what we're trying to clarify. DD are effectively saying it is not their fault so it must be the CAM or MP.

    if your request is intended to generate logs for DigitalDevices
    It is.

    then we should already have plenty of them collected and uploaded with different answers in this thread.
    We do not have the log files that we need because we did not previously test this scenario. The point that I'm trying to show is that it is enough to simply restart timeshifting (or recording) after the audio stream is added and... "ta da": everything is okay and all streams are decrypted. Therefore the problem is not with the CAM and must be with the driver.
    We have log files showing failure to decrypt the new stream... but we do not have log files showing that stopping and starting timeshifting solves the problem.

    Does that make sense?

    I rather see the risk of uploading "no issue" logs that will confirm DD's understanding that there actually is no issue.
    DD accepts there is an issue. They just don't believe it is their fault.

    I believe that the only way of demonstrating that there is a problem is to repeat the recording of the pre-show mono-language setup.
    That's exactly what I asked for except with timeshifting rather than recording. I asked for a demonstration with timeshifting because it is easier to see when the new audio stream has started, and easy to stop and restart. If you want to try to produce logs with recording then please by all means do that... but just make sure you include the essential points:
    1. Two tuners.
    2. Two encrypted channels being decrypted in MTD mode (1 shared CAM).
    3. After the dynamic PMT change on one of the channels, stop and restart that channel while leaving the other channel running.
    Does this make sense?
     

    Stéphane Lenclud

    Retired Team Member
  • Premium Supporter
  • April 29, 2013
    2,576
    1,294
    Home Country
    Germany Germany
    Should we, did we ask DD to try on reproduce the issue without MP?
    Should we, could we try and reproduce just by using DD software? That's a question to myself I guess :)
     

    mm1352000

    Retired Team Member
  • Premium Supporter
  • September 1, 2008
    21,577
    8,224
    Home Country
    New Zealand New Zealand
    Should we, did we ask DD to try on reproduce the issue without MP?
    Should we, could we try and reproduce just by using DD software? That's a question to myself I guess :)
    An excellent question. (y)
    Testing with DVBViewer or some other software was going to be my next suggestion if DD can't accept the results of this test.
    Probably we can't ask DD to reproduce; they are busy with many things. If we want the problem to be fixed, we must do as much of the work as possible.
    AFAIK DD do not have their own software. DD-TV is a rebranded/reskinned version of DVBViewer.
     

    HTPCSourcer

    Retired Team Member
  • Premium Supporter
  • May 16, 2008
    11,418
    2,335
    Home Country
    Germany Germany
    Here is my contribution, run on Cine S2 + Duo Flex S2 + CI + MaxcamTwin V1 (V5.25)

    16.56 - started recording on Sky Action HD
    16.58 - started watching Motorvision TV
    16.59 - stopped watching Motorscience TV
    then restarted Motorscience TV - and got a channel encrypted message (this happens occasionally :unsure:)
    tried another Sky Channel, which started OK, then switched back to Motorscience TV
    Verified the recording on Sky Action HD - both language streams correctly recorded

    Let me know if you need anything more.
     

    FreakyJ

    Retired Team Member
  • Premium Supporter
  • July 25, 2010
    4,024
    1,420
    Home Country
    Germany Germany
    Thank you, but unfortunately you did a little mistake :/

    You have to do it like this:

    1) Start a recording on an encrypted channel
    2) Start watching Tv on a channel which will get a dnamic pmt change soon (e.g. gets an extra audio track).
    3) check if you can switch to that audio track (should fail)
    4) stop timeshifting
    5) start timeshifting the same channel again
    6) check if you now can listen to the audio track
    7) stop timeshifting
    8) stop the recording
    9) check if the recording is okay :)
     

    Stéphane Lenclud

    Retired Team Member
  • Premium Supporter
  • April 29, 2013
    2,576
    1,294
    Home Country
    Germany Germany
    @HTPC_Sourcer That's not the expected result I'm afraid :(

    The bug should have left your recoding with an empty English stream.
    Are you sure that an English stream was dynamically added to your Sky Action HD between 16:58 and 16:59.
    That's typically happening before the beginning of a movie.
     

    HTPCSourcer

    Retired Team Member
  • Premium Supporter
  • May 16, 2008
    11,418
    2,335
    Home Country
    Germany Germany
    The English Stream is present from the very beginning of the recording and is not perturbated by the second TV stream. That's why I wrote that I don't understand the purpose of the exercise. This behavior was never in question.
     

    mm1352000

    Retired Team Member
  • Premium Supporter
  • September 1, 2008
    21,577
    8,224
    Home Country
    New Zealand New Zealand
    The English Stream is present from the very beginning of the recording and is not perturbated by the second TV stream. That's why I wrote that I don't understand the purpose of the exercise. This behavior was never in question.
    Let me try to explain this a third time. Please read carefully. :)

    It is important to understand: from the perspective of the CI/CAM/driver, adding a new audio stream looks basically the same as stopping and restarting timeshifting (or recording) on the same channel. If you don't understand this then you'll never understand the purpose of the test.

    We know from previous tests that when a new audio stream is added while timeshifting/recording, that new stream is not decrypted. Part one of the test is to confirm this.

    As I said above: adding a new audio stream looks similar to stopping and restarting timeshifting/recording on the same channel. If one works then so should the other; if one doesn't work then both shouldn't work. In the past we have not tested whether it is possible to do this. We assumed it would work but actually we never tested it. Part two of the test is to check this.

    Finally, from previous tests with the patch we have seen that sometimes starting decrypting on one channel can cause other channels to stop decrypting. Part three of the test is to check if/when this happens.

    So there are three parts to the test. All are critical.

    How are we testing these three parts?

    The way we're going to test part one is to timeshift or record an encrypted channel, wait for an audio stream to be added, and see whether the new audio stream is decrypted or not. Of course we expect it will not be decrypted because we have done this test before. However we must do it again. DD drivers have been updated; CAM firmware could have changed; etc. so the problem could have gone away.

    The way we're going to test part two is to stop and restart the timeshifting/recording from part one. The point is that the audio stream is not decrypted at the start, but after stopping and restarting we expect it will be decrypted successfully. To be clear: part one and part two should be done with the same tuner and same channel.

    The way we're going to test part three is to have a second encrypted channel recording or timeshifting while doing parts one and two. After part one and two are completed we stop

    Does it make sense now?

    Here is my contribution, run on Cine S2 + Duo Flex S2 + CI + MaxcamTwin V1 (V5.25)
    16.56 - started recording on Sky Action HD
    16.58 - started watching Motorvision TV
    16.59 - stopped watching Motorscience TV
    then restarted Motorscience TV - and got a channel encrypted message (this happens occasionally :unsure:)
    tried another Sky Channel, which started OK, then switched back to Motorscience TV
    Verified the recording on Sky Action HD - both language streams correctly recorded
    Here you have test part two and part three without part one.
    For part two: we should expect restarting Motorscience TV to have worked. The fact that it failed is really really interesting.
    It would be interesting even just to test part two again with other channels and see if the result is repeatable.
     

    Users who are viewing this thread

    Top Bottom