1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Is MPEG4 Our Real Issue??

Discussion in 'DIRECTV HD DVR/Receiver Discussion' started by lguvenoz, Nov 4, 2006.

Thread Status:
Not open for further replies.
  1. btmoore

    btmoore Hall Of Fame

    1,079
    0
    Aug 28, 2006
    Yes you get extra points for that. That is part of the agent functionality that determined that you need to watch more tv.
     
  2. lguvenoz

    lguvenoz Icon

    550
    0
    Aug 22, 2006
    I agree with having high expectations. I would certainly have hoped for better reliability from day one. The good news, if there is any :), is that D* is highly motivated to resolve these issues. In particular to get the MPEG4 issues nailed down.

    I know there are those that will say the H20 has MPEG4 so why is the HR20 so different?

    The fact is the HR20 is a full blown computer. It has orders of magnitude more complexity, and so the potential for issues is higher. I started this thread because I really want to try and find the major source of problems. There must be one based upon the issues reported and my own experience. The search is for some commonality.

    MPEG4 seems the most likely culprit in my opinion. Think about this scenario:
    1. The scheduler process in the HR20 launches a new recording.
    2. During the recording the MPEG4 codec hits an unhandled exception.
    3. That exception rolls up the chain to the recording process.
    4. The recording process doesn't know how to handle it so it aborts abnormally.
    5. The exception then rolls up the chain to the scheduler.

    At this point the system does one of two things:
    1. The scheduler handles the exception and as a result whacks your recording or leaves it listed as a partial that is really corrupt because of the abnormal termination of the recording process.
    2. The scheduler doesn't handle the exception and pukes killing the current recording and a bunch of future ones in the process.

    It's a realistic scenario and indicative of where I think D* fell short. These are things that would have only become major problems with the complexity of a DVR so they banked way too heavily on the H20 to be a test bed for MPEG4.

    My personal score for the HR20 is a 2.5 (I would give it a 2, but I do think the picture quality is outstanding).
     
  3. matto

    matto Banned User

    686
    0
    Sep 1, 2006
    This is the best analysis and well thought out opinion I've ever seen on this forum. Thanks for taking the time to work it all out.
     
  4. lguvenoz

    lguvenoz Icon

    550
    0
    Aug 22, 2006
    It's my livelihood so I live for dealing with issues like this. Plus I'm highly motivated to get my wife off my back about this thing so I'm hoping D* is watching.

    PS - Thanks, and Earl feel free to pass along to your buddies at D* to see if it rings any bells....
     
  5. btmoore

    btmoore Hall Of Fame

    1,079
    0
    Aug 28, 2006
    So you are a representative of D* and you can officially state that as a FACT. If not, than it is your opinion, please state is as so.

    My watch, car, phone, microwave, oven, other STBs etc etc etc are all computers, there is nothing in this computer that is complex. It is all very mundane, there is nothing leading edge here, even MPEG4 H.264 is not new, I have been using it to stream "real time" security camera video for 3.5 years. Because it is for security purpose it need to be very reliable because the safety of my employees is dependent on it working.
     
  6. mcl

    mcl Legend

    134
    0
    Aug 16, 2006

    Errr...that's great and all, but it doesn't explain the problems people have when neither tuner is tuned to an MPEG4 stream.
     
  7. RAD

    RAD Well-Known Member

    16,504
    142
    Aug 5, 2002
    Dripping...
    I have no problems viewing live MEPG4 channels. However on only our local Fox channel, which is the only 720p LIL-HD we get, as soon as I do a pause or watch a recorded show there is a bunch of video problems with frozen screens (audio is fine). This happens on both HR20's and 99/103 readings are in the 90's or 100.
     
  8. lguvenoz

    lguvenoz Icon

    550
    0
    Aug 22, 2006
    No I do not work for D* so it is just opinion, but the reality is that with new satellites being launched early next year that are being put up to support MPEG4 they would be fools to not put resolving MPEG4 issues at the very top of their priority list.

    In terms of the analogy to other computers, you are somewhat correct, but not entirely. A watch, car or microwave have much fewer permutations of usage scenarios. In the case of cars though you are seeing weird computer problems with the newer cars that do offer a lot of features (my own car has interior lights that only turn on 2/3 of the time when you open the door due to a problem like this that the manufacturer has been chasing since 2004).

    I agree with you completely on MPEG4 H.264, but who knows where D* got their codec from. I don't, and would expect that they "cheaped out" on this and found the company with the lowest cost solution for decoding. In your scenario I would guess that the same company provided both the encoding and decoding hardware/software so they solved 99% of the problems just by doing that, and the fact it is for security purposes means the company was highly motivated to ensure a bug-free implementation (versus the "it's only TV" approach D* has taken).
     
  9. mcl

    mcl Legend

    134
    0
    Aug 16, 2006
    No, they would be fools not to fix the current problems with the basic functionality. I.e., the problems with recording and playback.


    The longer these issues remain unresolved, the longer they look like fools.
     
  10. lguvenoz

    lguvenoz Icon

    550
    0
    Aug 22, 2006
    Not 100%. I'm certain the HR20 has some of the same problems the R15 does, but I would expect that those problems are getting better and with regards to the HR20 are probably the minority.

    I still would say that if you watch any MPEG4 content there is the potential for a cascading type of failure effecting even the MPEG2 channels.
     
  11. mcl

    mcl Legend

    134
    0
    Aug 16, 2006
    You can't have a cascading failure from inert content on a hard drive. It can only trigger an exception if you're actively watching it, or the stream is active on one of the tuners.


    Nice try, but it simply DOES. NOT. EXPLAIN. PROBLEMS. WITH. MPEG2.


    And those problems are IDENTICAL to the MPEG4 problems.
     
  12. lguvenoz

    lguvenoz Icon

    550
    0
    Aug 22, 2006
    My opinion is that the HR20 is most plagued by MPEG4 issues though that are a fundamental problem with the codec. Even recording and playback.

    The other issues are probably similar to the R15 problems with recording and playback, and might even be the result of a shared codebase between the two units. Does anyone have any inside information on how much of the actual code is shared between the two units???
     
  13. LameLefty

    LameLefty I used to be a rocket scientist

    12,182
    105
    Sep 28, 2006
    Middle...
    You're ignoring the premise of the original post: that decoding, recording and playing back MPEG4 streams results in errors or failures that cascade through the system and result in lock ups and other faults, regardless of the stream being recorded or played back. I think it's a valid premise, personally, even though I've yet to have a lockup, partial or failed recording (including plenty of MPEG4 recordings).
     
  14. mcl

    mcl Legend

    134
    0
    Aug 16, 2006
    Have you ever written a line of code in your life?
     
  15. lguvenoz

    lguvenoz Icon

    550
    0
    Aug 22, 2006
    You are incorrect that the problem can not be from inert data. All content whether MPEG2 or MPEG4 runs through a single display driver (firmware mind you but still driver software) to output the information to your TV. An unhandled MPEG4 exception could easily corrupt something within this driver that would impact all television viewing.

    This is just like on your PC when it becomes unstable over time because programs do not shut down properly and either corrupt or tie up in memory resources. Once something is played or watched it should leave no residual impact on the system, but it can. Especially when you are discussing improper exception handling since it causes processes to terminate abnormaly and potentially leave artifacts.

    This machine has an OS of some kind under the covers that all content runs through and can potentially impact. It doesn't make it less frustrating, but this again is trying to see if there is some fundamental issue.
     
  16. btmoore

    btmoore Hall Of Fame

    1,079
    0
    Aug 28, 2006
    Ok over reaching in my analogy when it comes to my watch, the point being I have many embedded OS appliances and none of them have ever had range of problems in providing the basic functionality that they were suppose to provide out of the box like the HR20 is having.

    Regarding the codec, that is a D* YP not a MP, they need to use codecs that work period.

    I want to get this strait, you believe, D*, a TV content distributer, that is what they do for business, this means that is what they have to do right 99.999% of the time, that it is not critical that they encode their mpeg4 in a bug free implementation, because it is TV. That one is funny, I am using cheap off the shelf h.264 systems, the mepg4 h.264 systems are far far more mature than anything I am using, in fact some of my h.264 systems are end of support and life. The point being it is a very mature technology, nothing here is new, at best the newest technology in this box is 5 or more years old.
     
  17. lguvenoz

    lguvenoz Icon

    550
    0
    Aug 22, 2006
    If he hasn't, I certainly have, and I've dealt with many a problem causing by a cascading failure. When you have any complex system this type of thing can occur. It never should, but it does.

    Every company is sloppy at some point in development, and D* is no different.
     
  18. PoitNarf

    PoitNarf New Member

    4,880
    0
    Aug 19, 2006
    Seems plausible to me. Ofcourse all of this is speculation unless someone on the forum happens to have taken a peak at the source code.
     
  19. matto

    matto Banned User

    686
    0
    Sep 1, 2006
    I think his point was that they probably used off the shelf decoder libraries from the cheapest vendor possible; and got all the software quality they paid for.
     
  20. LameLefty

    LameLefty I used to be a rocket scientist

    12,182
    105
    Sep 28, 2006
    Middle...
    Actually, I have. Have you? You sure don't act like you've ever seen a boundary condition exceeded or a buffer overrun, given the disdain with which you are attacking the premise of the original poster.
     
Thread Status:
Not open for further replies.

Share This Page