- From: Bob Lund <B.Lund@CableLabs.com>
- Date: Mon, 6 Oct 2014 16:01:05 +0000
- To: Alexander Adolf <alexander.adolf@condition-alpha.com>
- CC: Jon Piesing <Jon.Piesing@tpvision.com>, Silvia Pfeiffer <silviapfeiffer1@gmail.com>, W3C Inband Tracks Reflector <public-inbandtracks@w3.org>, Nigel Megitt <nigel.megitt@bbc.co.uk>
On 10/6/14, 4:51 AM, "Alexander Adolf" <alexander.adolf@condition-alpha.com> wrote: >Dear Bob, > >On 2014-09-30, at 18:13 , Bob Lund <B.Lund@CableLabs.com> wrote: >>>>>> [...] >>>>>> >>>>>> For @id, why doesn©öt the use of PID work. >>>> >>>>> PID can't work in Europe as we do a lot of forwarding of broadcast >>>>> streams and service aggregation. A service on a cable trunk for >>>>>instance >>>>> may come from a terrestrial feed, which in turn is sourced from a >>>>> satellite feed, and the cable operator is mixing is some more >>>>>services >>>>> from another satellite or another cable. The probability that the PID >>>>> will be changed a couple of times along the way is quite high. For >>>>>this >>>>> reason DVB has invented the stream identifier descriptor, bearing a >>>>> component_tag. Even if the PID changes the tag is retained. >>> >>>> But why does it matter? The transport stream still requires that PIDs >>>>be >>>> unique. So, each HTML track will have a unique id if PID is used. What >>>> else does the track.id need to convey to the Web application? > >Ah that's a very good question! Apologies for not having provided more >explanation earlier on in. > >Beyond unambiguously distinguishing the tracks in the player - for which >PID would be sufficient - in HbbTV we are looking at applications >referencing tracks. Such an application will typically be authored and >provided by the content originator. If the original broadcaster in their >app were to use the PID under which they broadcast a particular track, >that PID would become altered after the signal going e.g. from satellite >into a cable network. But the cable provider downlinking the satellite >service will of course not be able to change the JS to accommodate the >new PID value. Hence, for HbbTV, we need and end-to-end unique identifier >that is independent of the underlying transport, so that a document or JS >script can unambiguously reference a particular track no matter what >sausage machines the TS has gone through. > >I was wondering whether OCAP might not have similar requirements on track >identification? Thanks for the explanation. eTV does have applications bound to specific content items but it is done using EBIF [1] for defining ¡°bound apps¡± and EISS [2] for signaling of those apps. The basic concept is that an EBIF app gets downloaded from a data carousel and EISS signaling embedded in an MPEG-2 TS (or conceivably ISOBMFF trak or DASH AdaptationSet) controls the launching and life cycle of the app. So, an app is ¡°bound¡± to content by virtue of the signaling being a track in the content. CableLabs has demonstrated how this can work in the Web world. The EBIF app is replaced by a Web page. EISS messages are delivered as metadata text track cues. The messages specify the app URL and control the lifecycle of the app. This approach removes the need for a globally unique content id to be associated , and instead, uses an app signaling channel associated with the content item. We¡¯ve got a demo and reference code if you¡¯re interested. > >>> For DVB systems, the spec might as well just require that the id be a >>> unique identifier arbitrarily assigned by the UA rather than requiring >>> that identifier to be the PID. > >I am afraid that I'm not fully with Nigel here. See my explanations above. > >> So the unique identifier assigned by the UA could be the PID, right? It >> would best for interoperability if the the UA set the MPEG-2 TS track id >> the same way in all cases. > >As opposed to the Web, the broadcast landscape is segmented into national >profiles since governments traditionally very closely regulate their >broadcast market. Hence, historically there is a different, national >broadcast profile in each and every country. Back in the analogue days, >there were eight different flavours of PAL, seven flavours of SECAM, and >three flavours of NTSC. Hence, there unfortunately still is no one, true, >global digital TV receiver implementation that fits all. > >But of course there is a one, true, global UA implementation that ought >to fit all. Apparently there is a structural gap to bridge here, and the >segmentation of the MPEG2-TS section as we are suggesting is a symptom of >this gap. > >I am fully with you that a single, global TV standard would have >benefited industry and consumers alike. But then - for numerous reasons - >we are stuck with the history we have. > >Looking forward to your comments, > > --alexander
Received on Monday, 6 October 2014 16:01:43 UTC