[vdr] Re: wrong entry in EPG - Update

[Oliver Endriss <o.endriss@gmx.de>]

On Thursday 29 August 2002 12:51, Oliver Endriss wrote:
> On Thursday 29 August 2002 10:51, Christian Schuld wrote:
> > On Thursday 29 August 2002 06:19, you wrote:
> > [..]
> > > Therefore, we have to use some additional info to separate these
> > > EPG data. Any suggestions?
> >
> > Apparently on allmost any other satellites service_ids are not
> > unique. So i suggest to use the
> > a.) DiSEqC number
> > b.) Transponder
> > c.) service_id (aka ProgramID)
> > to uniquely identify a program.
>
> Of course, that could be done, but then you have to tune to each
> transponder to get EPG data for a given transponder.
> That's not optimal because service providers often use more than one
> transponder. On each transponder EPG data for all channels of the
> bouquet is transmitted. We need something like a unique provider id.

If I understand ETSI EN 300 468 correctly, the combination  
(original_network_id, transport_stream_id, service_id) should identify 
the event information in a unique way.
Perhaps this could be changed in a future VDR release, should be easy.

But this will not solve our problem because La Cinquieme transmits bogus 
EPG data for some channels. NetworkID, TransportStreamID and ServiceID 
match. For example:

EPG data transmitted by ZDF:
   EventID: 5509
   ServiceID: 28006
   TransportStreamID: 1079
   OriginalNetworkID: 1
   EitVersion: 12
   StartTime: Fri Aug 30 03:50:00 2002
   Duration: 5 Minuten
   Status: RUNNING_STATUS_NOT_RUNNING
   Descriptor: Short Event
      Name: heute nachtausgabe
      LanguageCode: deu

EPG data transmitted by La Cinquieme
   EventID: 24201
   ServiceID: 28006
   TransportStreamID: 1079
   OriginalNetworkID: 1
   EitVersion: 30
   StartTime: Fri Aug 30 06:00:00 2002
   Duration: 1440 Minuten
   Status: RUNNING_STATUS_NOT_RUNNING
   Descriptor: Short Event
      Name: PROGRAMMES ALLEMANDS
      LanguageCode: fra

--> There a two sources of EPG information for ZDF.
 
So what could be done?
a) do nothing, it's a feature :-)
b) bug fix the way Klaus suggested
c) ignore EIT data for 'other TS' (table ids 0x4f, 0x60, 0x61)
   --> you have to tune to each transponder to collect EPG data
d) remove *all* EPG data received from 'other TS' when valid data for
    'actual TS' is being received
e) use the LanguageCode to select which EPG data should be displayed
...

Oliver