[mythtv] EIT fix for Czech DVB-S/T

Stuart Auchterlonie stuarta at squashedfrog.net
Sat Jan 1 21:05:45 UTC 2011


On 26/12/2010 13:01, Petr Stehlik wrote:
> Hi all,
>
> I have written and attached a patch that fixes the worst anomalies in
> the DVB-S/DVB-T EPG for Czech TVs. It's against an old 0.23fixes for now
> - I'll update it for 0.24 and trunk when I upgrade to QT 4.6 equipped
> Debian release.

Patch itself looks good.

>
> I have a question regarding the way the EIT fixes are enabled - I don't
> like the transport ID/network ID mask as a switch. Is it unique at all?
> How is it working for DVB-T when there are many transmitters - are the
> TID/NID same across the whole country? How about enabling it per channel
> somehow so that users can turn it on/off on their own? And how can I

The TID/NID + service ID is the unique identifier for any given channel.
Since generally the fixups need to be applied to all EIT data across all
the channels the TID/NID specifier is sufficient to uniquely identify
the network (provided they aren't 0).

EIT data will only be collected for a given channel if
1) EIT data is enabled on the data source
2) The channel itself is enabled for EIT

> find out where some channel is getting EPG from? Since it seems like for
> a particular channel it's reading its EIT from a different transport
> because my EIT fix is not enabled for it.

EIT data is often "cross-carried". That is, EIT data is broadcast for
channels which may be on a different mux. This is common in DVB-T but
not DVB-S.

Your fixed gets triggered on matching the TID/NID. If your fixup isn't
being triggered then this must not be matching.

A bit of dvbsnoop will reveal what eit data is being broadcast.

>
> Another question is about sharing of EIT data between same channels
> (same channum, same callsign) of different sources. It seems wasteful to
> scan for EIT on each card when they are all pointing to the same
> satellite. If the callsign is same why not share the data for that
> channel.

If they all point to the same satellite, then they recieve the same
channels and should use *1* data source. All cards are considered
equal in this scenario.

If they point to different satellites or recieve different channels
(for example if you only have a CAM on 1 card) then they should
use different data sources and will collect EIT data individually.
There is currently no support for feeding EIT based data across
to more than one channel.

>
> Last thing is that since I switched from manual mythfilldatabase filling
> to this automatic EIT fetching Myth scheduler no longer uses the other
> two DVB cards and keeps scheduling all recordings on first card only. I
> suppose this is some kind of misconfiguration but when I was feeding
> Myth the EPG data from XML file it was recording OK from all three
> cards. Perhaps some channels on the other cards don't have its EPG data
> but I don't see it because all channels with same channum+callsign are
> shown just once so they might be masking out the 'empty' channels. Not
> sure how to debug it without renaming the channums and callsigns.
> Possibly by posting to -users list.

Sounds like a data source issue.


Stuart



More information about the mythtv-dev mailing list