Re: Proposal from HbbTV

On 10/15/14, 4:25 AM, "Alexander Adolf"
<alexander.adolf@condition-alpha.com> wrote:

>On 2014-10-14, at 17:13 , Bob Lund <B.Lund@CableLabs.com> wrote:
>
>>> [...]
>>> The original_network_id is buried in data structures in MPEG tables. So
>>> you would need to make a reasonable guess about which format the table
>>> payload has, before being able to parse out the original_network_id.
>>> Catch 22.
>> 
>> Why is guessing involved? Each section/table starts with a table ID
>>field
>> which identifies the type of table. It should be deterministic to find
>>the
>> original_network_id: the PAT identifies the PID carrying the NIT which,
>> according to DVB-SI, carries the network information section with the
>> original_network_id for each TS.
>
>Since both, DVB and ISDB use the network PID mechanism, you would need to
>feed the table you receive through both the DVB NIT parser, and the ISDB
>NIT parser, and then decide which one delivers more credible results
>(e.g. based on number and/or severity of parsing errors). I.e. a prior
>knowledge of whether the TS originates from a DVB or an ISDB system is
>required.

So I guess DVB and/or ISDB needs to specify something in the TS.


>
>>> [...]
>>> Unfortunately there is nothing that the UA can count on.
>>> 
>>> With that said, it doesn't mean that it couldn't relatively easily be
>>>put
>>> there. My suggestion would be for W3C to be in touch with ATSC, DVB,
>>>and
>>> ISDB and propose to them making the transmission of the TSDT mandatory
>>>in
>>> their specifications, and put a transport_stream_descriptor in it to
>>> allow receivers to unambiguously identify the system type. I shall be
>>> happy to help you crafting the message.
>> 
>> How about HbbTV making a recommendation to DVB, citing the Web use cases
>> we¹ve been discussing. I would be happy to work with you on that.
>> [...]
>
>What about making it a joint message between HbbTV and the Inband Tracks
>Community Group then? ;-)

Do you have a specific proposal in mind?

>
>  --alexander