[mythtv-users] Automatic channel scans

[Stuart Auchterlonie]

On 24/08/17 11:49, Mike Perkins wrote:
> On 24/08/17 11:37, Stuart Auchterlonie wrote:
>> On 10/08/17 12:19, Mike Bibbings wrote:
>>> On 08/08/17 19:01, Jeff wrote:
>>>> It seems to me that there are changes to the channel line-up on Astra
>>>> something every couple of weeks. Mostly I don't notice until recordings
>>>> start failing, and then I don't have the necessary block of couple of
>>>> hours to do anything about it, and so I end up running mythtv-setup
>>>> perhaps every 18 months.
>>>>
>>>> Would it not be possible to achieve something like automatic channel
>>>> updates by having a job run, say once a week or once a month, which
>>>> performs a scan when the backend is not being used (e.g. in my case,
>>>> before it is shutdown), compares the results to the previous scan, and
>>>> then presents the differences to the user?
>>>>
>>>> e.g. 'New channel "XYZ" found on Input 1. Add/Ignore?'
>>>>       'Channel "XYZ" moved to frequency/sid/... Update?'
>>>>
>>>> Is this feasible?
>>>>
>>>> Does anyone else think this is a good idea?
>>>>
>>>> Regards
>>>>
>>>> Jeff
>>>>
>>> Yes, it is a good idea and is probably feasible with some changes to
>>> mythtv.
>>>
>>> mythtv-setup already has the capability to perform an automated scan and
>>> either apply immediately  or save it for later use, see
>>> https://www.mythtv.org/wiki/Command_Line_Channel_Scanner
>>> However, as far as I can tell this automated scan does not work for
>>> DVB-S/S2, hence changes to mythtv.
>>>
>>> I was thinking of something along the lines of comparing current channel
>>> setup with automated last scan and just identify something has changed,
>>> the user would then use mythtv-setup manually, to either rescan existing
>>> transports (with search for new transports selected) or import the
>>> automated scan results.
>>>
>>> Mike
>>>
>>
>> We used to do this automatically back around 0.19 iirc. The problem this
>> lead to was in some countries it resulted in valid channels being added
>> and removed on a daily basis (ie, channels that weren't actually being
>> added / removed) which played havoc with scheduling.
>>
>> The best theory i had on how to do this, was to collect the relevant
>> signalling tables and if the new / removed channel change has been there
>> for at least 24hrs, then "activate" that change.
>>
>> It is distinctly non trivial....
>>
> Amen to that. In the UK we have a number of channels that are
> effectively part-time, that is, only transmitted for portions of a
> 24-hour period. Example: BBC4, which runs from 7:00 pm to 4:00 am,
> sharing bandwidth with CBBC and CBeebies, which only run during the day.
> 
> If a scanner ran when these weren't operating, then they'd remove them,
> even though they were valid. They would also find other channels which
> wouldn't be there later. Hilarity ensues. Running once per 24 hours
> wouldn't be good enough to pick up all the variations.
> 

In the UK, we got the signalling right. ie. it's possible to signal that
a channel exists but is "not running".

Actually iirc, the UK "cheats", the channels never technically go off
air, they just transmit an extremely low bandwidth MHEG application
which says the channel is "off air".

> Hmm. This might be something that could be dealt with by using the EIT
> scanner, no? Even if it isn't being used for schedule data, it could log
> changes/variations in channel organization.
> 

The EIT scanner, or even whilst recording for those that don't use EIT.

The second case obviously requires sufficient recordings to be done
to "notice" the changes, whilst the EIT variant would be more complete
as it walks all the mplexes

Regards
Stuart