[mythtv] [mythtv-commits] Ticket #1866: DVB Scanning - Remaining Issues

[Yeasah Pell]

Stuart Auchterlonie wrote:
> On Mon, Aug 21, 2006 at 08:19:12PM +0200, Rudy Zijlstra wrote:
>   
>>>> And please, please, do ignore NIT while you are doing this?? Because 
>>>> many cable operators do not provide a valid NIT-actual. They provide
>>>> many NIT-Other, but no, or no valid NIT-actual.....
>>>>    
>>>>
>>>>         
>>> This is something I personally don't get. I can understand that a NIT 
>>> becomes invalid after time if nobody maintains it. But I don't see the 
>>> point in transmitting a NIT_other in cable natworks. It isn't a useful 
>>> information at all.
>>>  
>>>
>>>       
>> It is, if you have a different frequency layout in different area's, and 
>> only 1 HE. Or clearly less Head-Ends than area's. In that case the 
>> NIT-Other describes the area frequency layout *for a specific area*, and 
>> NIT-Actual is worthless. You use the network ID to select the correct 
>> NIT-Other for your area. If you are willing to use the NIT-Other 
>> information, better algo's are possible. But you need to know a bit 
>> about the network you are installing on. Often the NIT-Other will also 
>> tell you the area it is valid for, in the network description.
>>
>>     
>
> Okay, this is doable, but is going to require some thought on
> the best way of implementing it.
>
> I suspect the cleanest way of implementing this will be to
> expand the list of locations in the country list, and then
> behind the scenes encode in the exceptions to what we are
> looking for.
>
> In this case
> German Cable { require SDT = yes, require NIT = no }
>
> plus whatever we need in terms of the frequency list to
> scan....
>
>   
While you're considering that, keep in mind that one of the NA 
commercial providers (dish network) screws up their SDTs by broadcasting 
the exact same SDT stream (which includes information on every transport 
on their network) on all transports -- in addition to having a 
~30-second SDT loop time, this also means that the desired SDT is almost 
always a SDT-other (and SDT-actual is for a different tp), and the only 
way to get the SDT on that network is to match by transport ID instead 
of matching by actual/other type (and you have to listen for SDT-other 
too of course) The NITs are similarly all-inclusive, but they are 
properly marked as other vs. actual (and the loop time is just a few 
seconds since the NIT data is considerably smaller)

So of course scanning and tuning do not currently work very well on this 
provider. You can hardly blame mythtv for that given the flagrantly 
noncompliant stream, but it would be nice if it could be made to work 
nevertheless. (Tuning is another matter entirely -- even if you knew 
enough to pick out the SDT-other you wouldn't want to wait up to 30 
seconds for it to come around on the stream, so tuning would have to be 
SDT-less to give reasonable performance -- Daniel was talking about 
allowing DVB PAT/PMT tuning with a tsid match a while back, and that 
would certainly work in this case)

I'm only mentioning that again because it could possibly be another 
thing handled by an exception list...

-y