There's a maximum (50) timer events.
When the EPG data is flagged as updated, the PVR has only to compare each event with the EPG entry for the event time - no searching involved so in itself not very intensive.
Only for events that no longer match the EPG entry, does the PVR have to search the EPG for an updated entry.
That likely uses the same routine as Series Link; next episode, and three-way-clash; alternative instance.

Thanks for the explanation I didn't realise there was a 'change in EPG data' flag This would make sense, except...

Is there an EPG data change flag for each channel?

If FP is employed by every channel (in the future we hope), and that I have, say 20 timers for programmes on 20 different channels and there is a 'EPG data' change, then again my box will be going nuts trying to update...

Obviously this is the extreme case scenario, but... will a Vestel clone be able to cope?!

I remembered reading it somewhere but couldn't remember exactly what it was called.

I think the Central Collator which takes care of the EPG for all transmitters and Mux, would update the EPG version simultaneously on all Mux.

Periodically, new (seven-day) events are added to the carousel, and past events removed. Events already existing for the week ahead if changed, would trigger a version update, (as I suspect would the adding of new events).

I'm only applying my common-sense logic to this, OK the 10,000 possible EPG entries would take a CPU significant processing time to search 50 times, but as I've shown that's unnecessary, it would take a fraction of that to confirm the existing 50 entries then only update those changed. Also for Series Link, only the new events need be looked at.(scratch that, that would be a bug leading to missed episodes if used to replace a changed programme in the week ahead data )

And remember that it needn't be brute force searching, either. As EPG data arrives, it would be quite simple to build a lookup table based on the various CRID values, so it's actually pretty straightforward to check if the details of a specific timer still match what's in the EPG data, which is a lot easier than a brute force search for a text string through all the data.

Very extreme As early receivers designed with only the Now/Next data in mind, could not cope with a FULL 7-day EPG, and needed surplus channels be deleted to show 7-day for the main channels.
The design accomodates a quantity of RAM for storing the EPG.
It needs considering what would happen in the extreme cases:[list][*]The number of Mux broadcast increases (or a receiver is so situated as to be able to receive channels from multiple transmitters)[*]The number of channels squashed into each Mux significantly increases[*]The AVERAGE length of each programme decreases. Imagine if each 30-60 minute entry in the EPG were replaced by 6-12 5-minute entries each with a full synopsis attached.[*]The same receiver is to be marketed in a region where any of the above already applies.[/list]

No, the data is transmitted in tables in which data for all services is group together. There is one set of tables for services in the current mux and one set of tables for all other services.

Within those tables, the data is broken down into sub-tables and segments, with a segment carrying events starting within a 3 hour period.

So if only one event changes, it is only actually necessary to roll the version flag in the table in which its segment is carried, but I don't know whether the Central Collator for UK DTT does that or updates "everything".

Either way, it's a trivial software task to parse the tables for events you're "watching".

The irony of the way the DVB specification partitions the data over the tables based on how far from today the event starts is that events are periodically moving between tables (and all the version numbers changes accordingly) even though their data has not changed.

I must admit that I can't get too excited about series link but your posting has raised a question in my mind. How does series link identify which programmes are part of the series?

What happens in the case of something like Heroes where the BBC broadcasts each episode half a dozen times on a selection of channels? Do you end up recording every episode multiple times?

No - within a series the BBC will have different series links (IE BBC Three premier, BBC Two showing, signed overnight showing etc...), and all episodes marked with that series link you chose will be recorded (eg If you press OK on singed version overnight you'll get the signed versions). There is also an episode id which stops the same episode being recorded twice and allows clashes to be dealt with (which runs across the different series markings, so if a clash prevented the BBC Three showing from recording, the BBC Two one will be instead - actually I think there is a BBC Three repeat earlier but you get the point).

There should be no reason for AR and padding to not be used at the end of a program however. It should be simple to say record for an extra 5 mins after the signal is given.

I had MotD2 set for SR and can confirm the BBC1 event was not linked - it did not appear in the schedule, nor was there a series option on the timer once selected. If I understand the explanations given, this was an error in the broadcast data.

It's not 1 either, it's the TV channels that send the broadcast data to the machine, the data was updated later in the evening.

To go back to the original postings on this thread, the email from Humax said the updated software "should be completed in the next 2-4 weeks and as soon as it is, it will be made available for manual download and installation from the support section on our corporate website"

That email came around 24 November, so can we expect something on their site before the end of the week? I'd suspect not, but their email implied this. I wouldn't think they'd release new software just before Christmas, as any issues would have minimum support.

That is the completion of the firmware coding. Next the testing phase, then QA phase, then... and then the booking of the slot for OTA transmission:

So no, the firmware will not be released at the end of this week, or month... but at the stated time...

I wonder what their testing and QA phase includes?

Will be great to get some bug free software for a change......

Your's is an interesting interpretation of the original email.
I agree with scoggy that it suggests it will be available for download after 3-4 weeks, not at the same time as the OAD in January.

The idea for the FPG2 firmware is to simultaneously release this OTA and HCSA. That was the idea for v20 too, but as it got installed on the TBX and those machines rolled off the lines... the betatesters discovered the problems with it and it was pulled from OTA and HCSA. But it got released via Hummy.org.uk later to reach a wider audience. This is well documented in various threads on various forums...


I hope someone with more authority... Barry? will agree with me and put you and others out of missery by hoping for something that won't happen...

But never the less probably correct.


Edit: Don't know about the authority bit son_t, but I would agree with your posts.
Last edited by Barry : 18-12-2007 at 10:57

Am I being pedantic? The email does say it will be complete in 2-4 weeks (i.e. by the end of this week) and as soon as it is (complete) it will be made available on the website.

Can it be interpreted any other way?

They said they'd wanted to get an OTA before Christmas, but couldn't. Nowhere do they say (in the email) that the download would be available at the same time as the OTA update.

Is your information coming from a source other than the email? Was the email superseded or just wrong?

(I'm not chomping at the bit for this update btw, I'd rather it was tested properly and released when ready, I'm just going by what Humax sent me.)

I understand little..none !!..of the above....so......

I have the 9200T.

Hardware ver....REV 1.0

Software ver...PEXTF 1.00.08

Loader....a 4.09

Update...09:08:06.

What differences will there be with the new update. The box works fine, and I will be a dead duck if it goes belly-up !!!!!!!!!!!!!


I cannot find "auto-update" in the menu...please advise..

The FAQ is a good place to start: http://www.hummy.org.uk/invison/index.php?showtopic=473

Maybe...

I think you are right - having re-read the email...

Like Barry says, he's not an 'authority' but we are hinting that it might not turn out like the email says...

I am basing my assuptions on the previous aborted release of v20, so I might be wrong for this one (too), but I think the idea of FPG2 is to have simultaneous OTA and HCSA release...

Unless someone at Humax want to give us an early Xmas present... like last year

I'd forgotten last year's came around this time. I doubt people will want to risk testing new software while they're away for Christmas.

Thanks for the link, appreciated...

It does not refer to the latest update in detail, but it does say it can cause issues, which will be addressed..sometime.......is there anywhere I can find any details...

I assume that if I prefer to "leave well alone" I do not set it to auto....

You will get lots of new features and hopefully fewer bugs than the version you are using. None of the new features will force you into using the box in a different way but will offer what are, in my view, easier ways of doing things. The only inconvenience is that you will have to reset to defaults which will mean you will lose your future recording schedule. The stored recordings will not be affected.

Why Is the data broadcast for SR etc so poor? This is irrelevant but I have windows media center on my laptop and that seems to get a decent tv guide with episodes etc rather than just 'Friends' like my freeview box. The broadcasters must have this info lying around to be able to sell it to MS for media center. How hard is it to put on the air in the EPG?