back to article Mass Sony DVR seppuku riddle: Freeview EPG update fingered

There’s still no word from Sony to indicate when it will fix a glitch that put many of its hard-drive-equipped Freeview video recorders out of action this weekend. The problem affects Sony’s HXD series of digital video recorders (DVRs) located across the length and breadth of Britain. Many owners arrived home on Friday evening …

COMMENTS

This topic is closed for new posts.
  1. Colin Millar

    Broadcast EPGs

    Always seem to be a bit crap in my experience. If your signal is a bit off or your tuner a bit old or the box a bit underpowered (not a rarity in the world of consumer electronics) you can expect to wake regularly to the frozen update screen - its been happening for years.

    Manufacturers could resolve a large part of it by supplying properly balanced PSUs but the broadcasters need to work on the robustness of the broadcast EPG - not everyone can achieve a perfect signal and tuners that are perfectly acceptable for viewing shouldn't need to be replaced because the EPG update requires perfection.

    Me I switched to xmltv and never looked back on all that EPG crap.

    1. handle

      Re: Broadcast EPGs

      What's a "properly balanced PSU"? Something that can supply enough current? In my experience if the PSU is not built-in it will be likely to have more power than required simply because it's cheaper to buy a generic one than one designed to match exactly the maximum power draw of the appliance.

      1. Colin Millar

        Re: Broadcast EPGs

        A properly balanced PSU will have a maximum continuous rating at above the maximum power draw of whatever it is supplying.

        Unfortunately lots of manufacturers in CE tend to cut the overrating to a bare minimum and even spec to the peak rating instead (cos it's a lot cheaper). If they are buying in generic external PSUs it still applies - it will be even cheaper to provide a lower powered unit.

        It applies across the range. Sky boxes are notorious for needing to have a bit of a rest for a while. Digifusion were notorious for it to the point where people were actually advertising PSUs as direct replacements for named boxes. Thompson and Philips are other brands which have been plagued by this.

        This from the DF FVRT200 user manual on EPG freeze ups

        "The unit has experienced a problem in downloading a programme guide or software update.

        and does not respond to the remote control - Power cycle the unit by switching it off at the mains for 10 seconds"

        Translation = "Not enough go juice to complete the requested operation - need to cool down a little bit first"

        1. handle

          Incorrect translation

          "Not enough go juice to complete the requested operation - need to cool down a little bit first"

          No, the power cycle is simply because it has crashed, and the 10 seconds is for a cooling down unrelated to overheating: it allows the inrush current limiter (an NTC thermistor) to cool down enough to limit the inrush current into the discharged smoothing capacitor when you reconnect the mains. It then intentionally heats up again to minimise resistance during normal operation. (I also suspect that stating 10 seconds is to discourage the impatient from switching off and on so quickly that the power rails are maintained and so the device doesn't reset.)

          Also, the PSU is not "balanced" - it is simply "sufficient". Something that is balanced is at a point between two unwanted positions; with this, there is only one unwanted position - not enough current capability.

          My comments about power supply ratings were for external power supplies; for internal ones you have to make a bespoke design so I agree with you that you might as well shave off a fraction of a penny by reducing their rating to within an inch of their life.

    2. prefect42
      Meh

      Re: Broadcast EPGs

      That's just nonsense. An EPG should be trivial to standardise and extend in such a way that boxes that met the standard don't get upset by new extensions. Just basic checksumming on the EPG should be fine to cope with poor signal, as if it's got an invalid block of EPG data it can just throw it away and wait for it to be resent. If your signal's that crap, you'll be complaining about the lack of picture...

      I'm a very happy humax user, and have never had any problems with EPG on either freeview or freesat. Seriously, why would you want the EPG to come from anywhere else other than inline with the content?

      As for frozen update screens... That just seems to be random finger pointing, as it's not a general problem with set top boxes, and is surely unrelated to PSU issues.

    3. mccp

      Re: Broadcast EPGs

      "the broadcasters need to work on the robustness of the broadcast EPG - not everyone can achieve a perfect signal and tuners that are perfectly acceptable for viewing shouldn't need to be replaced because the EPG update requires perfection."

      This is simply wrong.

      The EPG is broadcast in system information tables specified by the DVB standards. It is easy enough to download copies of these either from ETSI or dvb.org.

      The tables that are used for the EPG are broadcast repetitively at intervals that vary between 10 and 120 seconds. It doesn't matter one bit if your signal is dodgy, or that your installation is rubbish. In fact, there is so much redundancy and error detection in DVB signals that if you can see video you can guarantee that you will be receiving good system tables.

      As is mentioned elsewhere, it is far more likely that a receiver's firmware doesn't implement the standards completely or correctly. Freeview are often at the bleeding edge of adopting the more difficult techniques that are described in the standards but are generally not used by other (less experimental) broadcasters. I suppose that may mean that STBs show up their shortcomings with Freeview more often than elsewhere.

      I have written and maintain a commercial DVB stack so that's how I know this.

  2. Anonymous Coward
    Anonymous Coward

    All I do know

    If it was ANY other brand of consumer electronics that had a update burp, it wouldn't be getting this amount of press...

    1. handle

      Re: All I do know

      Oh really? What about Apple, for instance?

    2. Anonymous Coward
      Anonymous Coward

      Re: All I do know

      Humax got a thrashing when their FoxSat HDRs wouldn't record from the BBC's temporary olympics channels last year, but they rushed out a software patch.

    3. Fatman
      FAIL

      Re: If it was ANY other brand of consumer electronics that had a update burp,

      Agreed!

      BUT IT IS SONY

      A company run by some of the most arrogant assholes on Earth.

      The only mistake, at least in my mind, was that the product committed seppuku.

      It should have been SONY executives that committed seppuku.

      (Coming from someone who had a PC bricked by a Sony rootkit!)

  3. PipzUK

    Anonymous Coward is so right, there was a reboot bug affective many Vestel boxes from Friday afternoon until Sunday afternoon! Vestel is a huge player in the market but mainly make re-branded boxes for the cheaper end of the market so presumably nobody at El Reg cares!

    http://www.vestelpvr.futaura.co.uk/news.php?id=1374347058

    1. Anonymous Coward
      Anonymous Coward

      Of course we care...if we know about it

      We got several tips about Sony over the weekend but none about Vestel.

  4. Anonymous Coward
    Anonymous Coward

    I thought it was just me

    Cue much thumping, powering off and on, hard resets, pulling the lid off and tweaking all the connectors. Eventually started working again after 24 hours unpowered, a couple of hard resets, and powering back on with no aerial connection, and then putting back into service. At the time I put it down to my skill in tweaking the connectors, but obviously not.

    1. monkeyfish

      Re: I thought it was just me

      Me too, I assumed the old bugger had just died of natural causes. Found a soft re-start would get rid of the update message, enough to be able to watch recorded stuff again, but not fix the tuner (hold stop and power on the front). Oddly mine would show a very brief (a single frame?) image when re-tuned but nothing else. Also found a firmware update from 2010 that said something about fixing a tuning problem, but not tried it yet. I've got quite an old model, and not ever updated it. Kind of glad to find out it's not just me.

      1. Anonymous Coward
        Anonymous Coward

        Re: I thought it was just me

        I'd already lost a couple of programmes earlier in the week (coat hanger aerial fell from its precarious precision alignment) which corrupted some of the timers, and I initially thought this was an extension of that, then convinced myself it had just died the death. Unplugging/replugging got rid of the 'update' message long enough to try a retune. Rinse and repeat half a dozen times over Saturday + a manual update of the firmware (took a while so something might actually have been done although the revision number was the same as before) and suddenly all was ok again (too late sadly for the Tour highlights of Alp d'huez).

        I had two previous Argos special freeview boxes bricked in the past by changes to the way they squeezed the channels a couple of years ago.

  5. Danny 4
    Boffin

    bad EPG data from Sky

    I'm one of those Sony users affected by the broadcast of out-of-spec EPG data.

    It's not and has never been about a firmware update from Sony. The BBC article is factually incorrect.

    Some knowledgable people have tracked the error down as corrupt EPG data from Sky, or its EPG data supplier, being broadcast on COM5. It would appear to boil down to a missing "/" in the data.

    Other makes of Freeview devices gracefully handled this error; the Sony HXD boxes made by Pioneer and some Pioneer boxes (and a few others) choked when presented with this data stream.

    Retuning using France as the location was a temporary workaround. They must use a different EPG ident or structure or something so the UK EPG packets in the broadcast stream were discarded and hence no recorder crash. You could select either Guide++ or digital EPG (though the EPG was still empty) to affect which timer controls to use. Selecting Guide++ while using a UK location did not help.

    By Sunday afternoon, Sony, DTT and other big nobs had got together, kicked bottom and the broadcast stream was fixed. Everything seems to be back to normal now.

    You can follow the whole EPG Black Friday saga here:

    http://www.avforums.com/forums/blu-ray-dvd-recorders/1795677-sony-870-all-sony-hxd-some-pioneers-dvd-recorders-no-freeview.html

    As a developer I know it's the dull boring bit but, hey, to all involved: more testing guys!

    1. Lee D Silver badge

      Re: bad EPG data from Sky

      Sorry, but that sounds like a software issue - failing to properly verify data it acts upon.

      A '\' in the data - yes, I quite agree, that could balls up the EPG. But it should NOT balls up the device to this extent. They have obviously written the software without any care what happens if it's fed even slightly incorrect data (which, with all the checksumming in the world, can still happen over a radio broadcast medium enough that you have to deal with it).

      But then it's up to the SOFTWARE to detect that something's not quite right, not charge through the rest of the string and hang on an "update" screen, or even do anything more than say "EPG programme error data - reusing valid cached EPG data - contact your DVB transmitter to resolve the problem" and carrying on as it had before.

      This is about sanity-checking in code, not the fact that there was an extra '\'. The slash was the TRIGGER. The cause is really writing crappy EPG-parsing code that doesn't error out nicely when it hits the unexpected.

      Hell, I learned not to expect ANY data from another system or user to be correct back when I was a teenager. It's just too easy to make many modern devices like this crash where they should really just say "Whoops, I found a problem here - I can't use this EPG data, what would you like me to do?". NOT hanging on a useless screen that the user can't interact with at all except to turn it off.

      1. Danny 4

        Re: bad EPG data from Sky

        @Lee D

        I don't understand why you're sorry?

        Did you read my post? The casual fault is bad EPG data. Yes, this should have been avoided by better software, both the software used by the EPG preparer - it should verify the data before committing to broadcast - and the firmware in the recorder should not trust external data even from non-belligerent sources. I reiterate, "to all involved: more testing, guys!"

        Also, there appear to be a fault on your keyboard. It seems to be outputting occasional words in all-caps. Unless this is a wetware fault?

        1. Bob H
          Boffin

          Re: bad EPG data from Sky

          The specifications say you should ignore invalid data, the manufacturers should handle edge cases but accounting for everything isn't always easy so there is a tacit agreement that the broadcasters will keep the data within spec. Some broadcasters do fix problems when they are raised, some drag their feet or deny they are doing anything wrong "the spec says you should ignore it so we aren't wrong". Over the years the broadcast suppliers have found many creative ways to break the specs and manufacturers have found many ways to misinterpret the specification.

          1. Richard 12 Silver badge
            FAIL

            Re: bad EPG data from Sky

            The Sony box still has an unforgivable error.

            It should not be possible for any data, valid or invalid, to actually crash the box.

            The box should have simply said "Whoops, the EPG data is corrupt." After that, showing either a blank EPG or the last-known-valid EPG data would be reasonable.

            What does your browser do when you go to a malformed webpage?

            Mine recovers from many errors (if the intent remains obvious), and shows a page saying what the error is and exactly where it happened if the intent has been lost.

  6. Cliff

    Sanitise your inputs

    This is 'Bobby Tables'* but on a broadcast scale. You should always assume data you don't control may be hostile, or just corrupted by accident.

    *http://xkcd.com/327/

  7. Test Man

    Ah, so this killed several channels on my Digital Stream last night then?

    1. Anonymous Coward
      Anonymous Coward

      Digitalstream designed some recorders for Sony.

  8. Anonymous Coward
    Anonymous Coward

    You're all wrong!

    Well, ok, not all of you.

    @ Colin Miler - I would bet a small wager that there's a 99.9% chance that the reason the manual says "...switch it off, wait 10 seconds, and then try again" is not because of a PSU problem, but because the software is buggy and leaking memory like Dear Henry's bucket. This goes for most things; phones, PVRs, TVs etc etc.

    As for the corrupted EPG data, there are very strict guidelines for what should be broadcast. If Sky are indeed at fault, then they should indeed have their hand slapped. Having said this, the software in the box should have been able to handle the problem. It's probably not realistic to "fix" such an issue, but it should certainly have thrown the data away. Maybe it DID throw the data away and that was the problem.

    Although I'm not involved in it any more, I am quite familiar with this particular family of Sony products (maybe it was my code that failed :-) - actually no - I know it wasn't. Like many consumer electronics products, some bits of them (usually software - again!) are a complete mess, and I would be the first to lambast such poor workmanship. Other bits are quite good, but such systems are very complex, and I think many people would be staggered at the amount of code in them - they have to do an awful lot of work to convert that satellite feed into a pretty picture and sound, and it is unfortunately a fact of life that things go wrong now and then.

  9. Anonymous Coward
    Anonymous Coward

    Oh...

    @prefect42

    ...and can I just point out to anyone suggetsing that the EPG data is trivial, errr... no it isn't. The EPG data is part of a very lage meta-data set, without which not much would work. The documentation for all this runs into many volumes. It is far from trivial.

  10. Anonymous Coward
    Anonymous Coward

    Well done Sony

    ...another stunning partial success.

  11. MarketingTechnoDude

    It's complicated both technically and commercially

    There seems to be an assumption that the manufacturer has access to every line of code in these Set Top Boxes ... Well this is rarely the case. Quite often, middleware components such as the "DVB Stack" may come from one supplier, the "MHEG Engine" from another and key hardware device drivers from the silicon (SOC / Demodulator) /Tuner) vendor(s) . Guess what? ... most of these chunks of code are supplied in Object Binary format (a bit like a Windows SDK or DDK).

    I agree the software should be tolerant of the error (and also the caller of these functions should have put some watchdog timer protection around them to recover from errant middleware).

    Most of these boxes are NOT connected to the Net (even if they are capable .. most users DO NOT BOTHER to connect them). So when the broadcaster makes changes that cause previously working boxes to crash or lock up etc there is the added problem about updating the boxes using a very expensive Over the Air Download service or hoping that users will get a USB stick to download the update from the net. Most instances should be resolved (and checked) at the broadcaster end though!

    These instances have happened several times in the past (ITV Teletext bricked many boxes several years ago, changes to the NIT table structure was another case)...

    It will not be the last time this kind of thing happens unfortunately...

    It is also totally impossible to build such a complex software system that is tolerant to every possible error condition with so many factors outside of the control of the manufacturer (and indeed middleware vendors).

    So perhaps in this instance Sony were merely a victim of broadcast stream errors and lack of defensive programming techniques in the middleware?

This topic is closed for new posts.

Other stories you might like