Re: chunked IMSC1

Perhaps we should schedule an agenda item to discuss this.

Some questions I have:

* What does backwards compatibility mean here? (we always have to check this!)

* Is there a TTML2 feature that's specific to inline animation (i.e. the set element as a child of a content element)?

* Is there an option to add support for #region-inline to a future version of IMSC, if it helps to meet LL use cases?

* Does Mike believe that _nobody_ has a requirement for some set of features of IMSC1, or just that the requirements are not local to his area of interest? Can we evidence that? Should we be pruning some feature support?

Nigel



On 21/03/2022, 16:53, "Michael Dolan" <mike@dolan.tv> wrote:

    I consider backwards compatibility to current segment coding standards to be a requirement. In that framework....

    Other than the regions being constant over the segment duration just discussed?

    Everything in <head> has to be constant.

    <set> would have to predict future content to work.

    I don't view the above as a real problem.  Live captioning today is already fundamentally constrained.

    I'd be delighted if there was a solution that enabled full features that is also backwards compatible with current segment encoding.  But I don't see it, and I don't believe there is a requirement to support everything possible in IMSC1 today, much less TTML2.

        Mike

    -----Original Message-----
    From: Nigel Megitt <nigel.megitt@bbc.co.uk>
    Sent: Monday, March 21, 2022 9:37 AM
    To: Michael Dolan <mike@dolan.tv>; Pierre-Anthony Lemieux <pal@sandflow.com>
    Cc: Cyril Concolato <cconcolato@netflix.com>; Glenn Adams <glenn@skynav.com>; public-tt@w3.org
    Subject: Re: chunked IMSC1

        There will necessarily be tradeoffs between LL live versus VoD scenarios; and full vocabulary or some subset.

    It is not obvious to me why such tradeoffs will be necessary - can you explain more?

Received on Monday, 21 March 2022 17:16:06 UTC