Timed Text Working Group Teleconference

<trackbot> Date: 23 July 2015

<scribe> scribe: nigel

This meeting

nigel: Runs through likely agenda. Any other business?

group: No AOB

Action Items

nigel: Glenn sent an update before the meeting - he has nothing to report on his actions.
... Andreas, how is Action-406 looking, re HTMLCue?

atai: It's on my list, but nothing to report now.

Action-409?

<trackbot> Action-409 -- Nigel Megitt to Discuss liaison with unicode with richard ishida -- due 2015-07-09 -- OPEN

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/409

nigel: I got a response from Richard Ishida.
... http://sandflow.com/public/cldr/imsc-codepoint-table.htm
... He asks why there are very few blue characters in some locales.

pal: The blue characters are the ones that are already in the main and punctuation and digit character
... sets for those locales. The non-blue ones show those that would effectively be added to the
... existing character sets for each locale. I'm happy to clarify that if it was not clear in the original
... communication.

nigel: He also suggests that Unicode would want us to propose how the characters would be
... integrated into CLDR, i.e. are we proposing new categories including all the characters or just
... new categories for the additional characters for subtitles and captions.

pal: I think for ease of review, just including the delta is going to be harder for the reviewer to look at
... instead of the full set of characters including those that are already in the main and punctuation
... sets. I'm interested in the feedback from this group. I found that when I just tried to generate
... the delta, i.e. just those characters that are not already present in Main and Punctuation that leads
... to a disjointed set. I think it would be easier for reviewers to see the whole set. Also perhaps in
... some languages there is a real difference between characters typically used in everyday print
... media or chat, etc, compared with those available and appropriate for subtitling and captions.

nigel: There's a trade-off between ease of review and ease of maintenance, if Unicode has to add
... any new code points to two sets for each locale instead of one.

pal: We could propose an algorithm. I am happy to propose a more detailed proposal if that's what
... you sense is what they want.

nigel: I do get the sense from Richard that Mark would like something quite detailed and specific
... as a proposal to review.

pal: Okay I'm happy to propose that with a set of rules.

action-409: [Meeting 2015-07-23] pal to generate a more detailed proposal of what we're asking Unicode to do.

<trackbot> Notes added to action-409 Discuss liaison with unicode with richard ishida.

Issues

issue-399?

<trackbot> issue-399 -- add errata for initial value fallback language -- open

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/399

pal: In seeking feedback on IMSC I didn't understand the style inheritance rules, but in answering my
... question on the topic Glenn found an issue that he's proposing to fix.

nigel: I'm not sure what the impact of that is.

pal: Me too. It's a complex algorithm, so I'm not sure what the impact is.

reopen issue-399

<trackbot> Re-opened issue-399.

nigel: The issue has the email thread attached, which is helpful.
... Is this to do with the inheritance of styles either from regions or from styles?

pal: Yes, exactly. There are lots of combinations of inheriting from parent elements, specifying styles inline and others.

nigel: I looked at that once a while back in assisting with an implementation, and it turned out to be
... surprisingly simple for most styles, in that you just ended up generating a list and searching for
... the first one of the required type, starting from the correct end of the list. It was computationally
... simple, though I suspect there are complexities that we may have missed, and things like fontSize
... expressed in % are of course relative to the parent.

TPAC 2015

nigel: I may have mentioned before, but in case I didn't, registration is open.
... http://www.w3.org/2015/10/TPAC/ for the details of registration.
... Also https://www.w3.org/wiki/TimedText/tpac2015 for our own registration and agenda.

WebVTT TTML mapping

Courtney: We're making progress on this, nothing to bring to the group yet.

pal: Have you run into any major issues, or is it merely a question of filling the holes in a framework.

Courtney: We're taking the approach of defining a profile in TTML that will map to WebVTT, which
... was Andreas's idea. We may want to extend it in the future. By taking a subset of TTML it makes
... the mapping pretty straightforward in some ways.

pal: That was the plan in Las Vegas I think. What's the relation between that and IMSC1?

atai: I think it should be a subset of IMSC1.

pal: Can you tell now which features from IMSC1 will be or will not be included? I'm happy to help by the way.

atai: I think there are some features absent from IMSC, like aspectRatio, progressivelyDecodable,
... multiRowAlign that are not mappable straight into WebVTT. Possibly the best thing to do is to
... look at the list when we have the first draft ready.

Courtney: Possibly we could send Pierre the profile to look at early, but we should have the profile
... ready relatively soon.

atai: The intention of the profile is to not include any features that cannot be mapped, but also
... to restrict the TTML structure to make it easier to map to WebVTT.

nigel: You mean algorithmically easy to map?

atai: Yes, for example the timing. You can make certain aspects simpler to make it easier to map.
... It needs to be seen in context rather than discussed on its own. There's a lot to be done, which
... could be done inside TTML, like mapping to a certain state that makes it comparable to WebVTT
... to make the mapping algorithm easier.

nigel: Is that coincident with Glenn's ISD plan?

atai: No I don't think so.

nigel: Rather than having two transformations it might be better to have one that's more useful.

pal: The reason I'm asking is I'm getting requests from IMSC users. When is the target date for a
... first draft?

Courtney: I don't know exactly - we're going to have a call after this to sync up. I think some time
... in August for a first draft.

nigel: That's great. Are there any other technical questions that have arisen from that work?

atai: Not from my side.

Courtney: I don't have any yet.

nigel: We've hit the end of our agenda, so I'll adjourn the meeting. Thanks very much everyone.
... [adjourns meeting]