That is not correct! As I posted on the other thread you quoted, a series can refer to programmes across a range of channels. The limitation being the scope of the CRIDs - ie: for BBC, all BBC channels, for ITV, all ITV channels etc.

Don't knock specifications if you haven't seen/read them.

I have not 'read the spec's' as you need to pay for them - i.e. they are made available to manufacturers who pay for them (subs?) to be able to implement FP systems. I am not a PVR manufacturer or work for such... I was only referencing info that have been made available to the public as it were...

Beg your pardon - where did you say you posted the info? I don't think I've ever read any detail argument why this is or isn't possible...

EDIT: found this:

Judging from the above, I don't think you've read the specs either. It is not exactly sure statement and obviously open to interpretation...

Not that I don't believe you... From what you say, I would assume that the SCOPE of a CRID might be restricted to mux'es as well as the broadcaster's channels - no?

Post 11 on this thread . Wasn't detailed - just stating the facts (having read the specs!).

I can't see anything in the data that would preclude series links from working across channels. It's probably more a case of whether or not it's required, and whether or not any particular implementation checks channels as well as the various CRIDs when looking for programmes.

All the CRIDs should be unique across all channels, so I don't see why in principle a box couldn't follow a series from BBC2 to Channel 4 (the last 2 Dennis Potter dramas were shared between those two, I recall).

Since all the data is available on all channels, it should be quite feasible - but there's the question of time, too. A full Freeview EPG can have around 10,000 entries in it, and I suppose in some circumstances, you might decide to just look at a subset, either based on 'provider' (which is part of the data stream), or on a single channel only.

Nigel.

So would series link know if a new episode is broadcast, or would it, for instance, record the same episode on E4 and also on E4+1?
Last edited by ChrisE : 25-11-2007 at 18:52

So far such SR metadata has been terrible. E4, etc in particular... as an example: the series data for 'Friends'. If you set this to record the series, they've not actually distinguished any of the episodes, so you get all of them recorded no matter what order or from which series...

You will only get one instance of the episode - whether it's come from E4 or E4+1. The benefit of Series Linking is that if all your tuners are busy when the first instance of a new episode is broadcast, your recorder will go for an alternate instance if it's got a free tuner at that time.

But it is all down to how good the metadata from the broadcasters is

OK, I take it back. I'm sorry for going around missing informing about the FP specs.

Thanks Nigel for making it a bit clearer.

The specs are probably defined so broadly that makers and broadcasters are left to their own devices I suppose...

The broadcaster can't even get the AT flags set properly or at all at the moment - I'm surprised that anything works!

My goodness, but you are very boring

Hi Phil

Is it general view that 'auto padding & series link' won't work together? if it is the case all series link development is a waste of time. We all know that 'accurate recording' is not accurate neither reliable at all so I don't think I will ever drop 'auto padding' for accurate recording.

PVR users can't just rely on TV stations to send their data in time. Who would use series link and risk missing the final minutes of a movie or footy? I don’t think many users would risk that

I just hope that PVR users won’t have to make a choice between ‘auto padding’ and ‘series link’


Auto padding is not part of the freeview specs but it is working perfectly and should also work with series link. Series link should be looking for the programme name & info not the time it is set to record.

To be honest, I don't know the full details of how it works, as has been shown above, and that I do not have FP specs. I got a feeling that the specs are being written on the fly anyway as manufacturers get more involved... and that they have/are changing as they get worked on...

All I know (or gathered) is that Humax gave up on fulfilling the ten items in the 'owners wish list' to proceed with implementing the Freeview Playback facilities. And because AP is not in the FP specs (i.e. you don't need AP if AT+SR work properly) then it has not been implemented.

There might me technical reasons why AP+SR(+AT) is/can not be implemented together and that most PVR makers have avoided this route and rely on the fact that the specs defines you only need to implement AT+SR for things to work.

I am really grabbing at straws here as I don't know the real ins and outs but I would hazard a guess that you probably need a PVR slightly cleverer than then 9200T is at the moment to be able to have all we desire.

That is why I ask about the Toppy as it has the TAPs capability which a clever coder could put something together to make the machine do the things you want... albeit at a lost of something else... I dunno...

I suppose in theory AP could work with series link (as the Toppy TAP shows), but then what happens if, say, the cup final goes into extra time for 20 minutes and Dr Who starts 20 mins late as a result.

Unless you have 20 mins padding, you're going to miss some of it - whereas with Accurate Recording, you don't. Or do you complicate things by waking up in time for the scheduled (padded) recording, and then extend it based on the EIT p/f info that's used for accurate recording?

Since the point of series link is to record things whatever time they're on, I think the average user would naturally expect that it would still cope if the programme was delayed, and presumably that's why most vendors seem to be going down that route - on the Toppy's Group 2 firmware, when you set a recording via the EPG for series link, you can't alter the options either.

The real solution, of course, is for proper updating of the EPG. The BBC do this by controlling the EIT information via synchronisation with their playout systems. Channel 4 are moving their playout ops to the same company, Red Bee, that handles the BBC, so hopefully they'll improve too.

Ideally, of course, the EIT conditions in the license need to be tightened up - at present there's not even an obligation on broadcasters to maintain the 7/8 day guide, let alone accurate information for Playback.

Nigel.

AP works in the main, but there are instances when it isn't useful as I found out once with the Junior Mastermind final. A weekly timer had been set for the half hour programme, but the final was 30 minutes longer and I missed the rest of that show. I guess I should have checked that out myself. Delays are another thing, and this is probably why AR/AT is supposed to help when the right data is transmitted.

Ultimately, I think the best thing would be setting minimum padding (dropped if impossible still).

EG with 2 mins before, 5 after, and The Simpsons (say) if AR doesn't get it to start recording by 5:58 it starts anyway. Likewise if AR means it would stop before 6:35 it would carry on until 6:35 anyway. If recordings are back to back on the same channel then use the AR to decide when to change it. That way it would always record at least what it does with either option so would always be the most reliable option.

It could also be programmed to store when its signalled to start, and make it default to this (IE first play after finished, and when play pressed twice) so it's only the (none to scarce) disk space that is wasted, not time, if AR is implemented by the broadcaster.

Still, that's a perfect implementation, and seeing as how I neither bought a PVR with auto-padding or accurate recording I'm well up on the deal. It just seems a shame that in order to use new really useful functional I'd have to sacrifice wholesale other really useful functionality.

Hi


Yes accurate record will hopefully help with large overruns, the problem is the majority of the time 'accurate record' isn't needed as must things run to schedule and the entire program can be caugt with a few minutes of padding.

My issue with it is 'accurate record' tends to go wrong even when things run to schedule, you can lose a few minutes at the beginning and sometimes at the end, or find nothing records at all when the TV company is having issues. So the choice seems to be use accurate record and continually find niggles with missing the odd start and end or having a completely lost recording, just so once in a blue moon it catches a massive overrun in your program, assuming of course at that particular time the TV companies update the data with that overrun.

The net result with accurate record is overall you get less programs recorded entirely than you would have using padding.


But series link will still work without accurate recording. AR is only a fine tuning of the start and end times, that could go leaving the rest fully operational.

The biggest improvement with the Freeview Playback specs is schedule tracking. You set a recording 7 days in advance, either manually from the EPG or it's done automatically via series link, and you get a recording schedule added with the times from the EPG. If a few days later the EPG is updated and the program is now on a day later, an hour earlier or 3 hours later (or any other time interval change), the recorder can see this and updates the recording schedule with the new times. This is what copes with Doctor Who being on 8:45 one week, then 9:30 the next week using series link, accurate recording doesn't come into it at this stage. Accurate recording's job is to cope with a program starting a few minutes early, or overrunning by a few minutes, but we already have a reliable system to cope with that, which requires no additional data broadcast or being reliant on someone throwing a switch, we call it "auto padding", and it's entirely under our own control.

Accurate recording is always going to be less reliable than padding. We already have issues with people not getting some recordings when others do get the same program recorded, this being put down to "data issues" with the TV company in certain regions, and this situation with the various regional opt outs, different schedules by region and other complications mean accurate recording is never going to be.

The Freeview Playback specifications need changing so that the real-time element of accurate recording has to be switchable on or off by the end user without disabling the other features.

Without this change people will be turning off from Freeview Playback completely.

Regards

Phil

That would be far from perfect as it would nullify the main reason for AR, sport overruns. You would end up recording possibly hours of extra footage (the end of the sporting event) just because you didn't trust the flags, then still depend on those flags to mark the end.

It seems to me the best implemenation would be to allow the user to decide which channels they trust the AR data and which they don't. In the latter case us AP.

This would then work in a very similar way to PDC did on my old VCR where you got to decide per channel whether to default to using PDC or not - and then per recording you could toggle that too.

Personally I am happy with AR as it stands. I've been using it since v15 and never missed any endings. At worst it was a few seconds at the beginning and that was due to the bug of reading the wrong flags. Since v20 everything has started nicely in the opening credits. Having said that I don't watch anything on ITV at the moment so YMMV.

You could probably rely on a cup final over-run being entered as it's sort of predictable but on Friday night/Saturday morning BBC News 24 overran by 15 minutes or so to report on the Australian election result and Click was late starting. However, despite being AR scheduled (under 1.00.20) my Click recording started at the scheduled time and only ran for 30 minutes so someone forgot to press a button!

The real test will be the next time there's a major disaster that causes main channel schedules to be interrupted with lengthy news flashes. I wonder if the system is geared up to reflect that sort of unpredictable late start/over-run?

So, everything would still be recorded, the only thing that would be wasted is disk space, which I don't find an issue. Hours is unlikely anyway - it's part of the specs to make sure the recording ends within 2 hours of it's (autoupdate) EPG time, so if a system crashes and it stops updating it doesn't overrun infinity.

Actually if you think about it, my plan is almost exclusively for those cases - without a live event it's very rare for programmes to be more than a few mins out anyway. Otherwise it just acts like autopadding.

Another option is to start recording 10 mins early, and up your padding time after the end is signaled, then automatically trim it to your autopadding around the AR "flags", essentially applying autopadding to the AR signals (which would presumably either be accurate or to time) so as not to waste as much space.

Your idea would be more complicated (for the user) as it needs people to work out in advance which channels are trustworthy or not. Where as mine would be a lot simpler to set up. A combination of the two would be even better (my way + a "trust list", which runs on pure AR, so if you want to you can put more time into getting it just right). Best of all would have been if tags were included in the EPG indicating the availability of AR seeing as how the DTG was adding additional data anyway...

There is also the issue that recordings don't start/stop instantly, which exclusily effects the start of the programme and would end up with problems, unless you start recording early.

Anyway, the vast majority of my viewing is on BBC/C4 channels so it's not all the important.

So... Match of the Day 2 tonight, can't be shown on its normal channel BBC2 due to the snooker and has moved to BBC1... The Humax is not currently showing it as anything - just a normal programme...

So what is wrong here?
1. Humax?
2. BBC?
3. FP Specs?

Take your pick... probably (4) All of the above?

Unfortunately I have not set MotD2 as a SR timer, so can't tell if it would have recorded if I had...

Looking at the CRID data, if you mean the programme that's on at 2250 tonight on BBC1, then it has one CRID entry, CRID31 = /4KS413

It's not got a CRID32 value, which would mark it as part of a series, so it wouldn't be series linked.

But the CRID31 should be sufficient to ensure that, if it had previously been scheduled on a different channel, the timer would update to record it from BBC1.

What you are saying is CRID31 is the programme ID? and CRID32 is SR info. If so do all programmes have CRID31?

So if at the start of the week MotD2 was on BBC2 and I scheduled that to record. Then during the we it got moved to BBC1 then the schedule timer should up that it changed channels..?

MotD2 is normally a SR so why should this particular showing not be..? Sounds like something can't cope with SR and programmes jumping channels...

Yes; CRID31 is the programme/episode ID, and CRID32 is the series ID, while CRID33 is for recommendations.

All programmes should have a CRID31, on a playback enabled channel.
If there's no CRID32 then the most likely reason is that it's not been entered, rather than anything else.

There are errors in the data at timer - for example, the Dr Who on BBC3 at the moment has a CRID32 of /KCIWNS on tuesday (and last week), but Wednesday's is /KCJ6W4, as are later ones in the week, though it's the same actual series, and the episode IDs themselves are sequential. So, a Series link set last week won't work after Tuesday (and will have missed out tonight's episode too, as far as I can recall).

There are going to be issues like this for a while, I guess. That doesn't mean that somehow the design of Playback is flawed - it means there are still process issues that need to be resolved by the broadcasters.

Are you sure about this? If so, then a FP box would have to check all (FP) programme timers in its list against the EPG to see if the times and channels have changed... constantly and update the scheduled timers? No? I would assume this is more intensive (for the box) than doing this for just SR timers...

What is this about the scope of an ID or a SR ID..? Is the scope limited by channels or mux? So what is the scope for, say BBC1 programmes?

Not (3).

Yes, but not constantly, only when the data changes - signalled by a change of version number in the data.



By CRID authority, so broadcaster in the FP model.


Any BBC channel.