Timed Text Working Group Teleconference

<trackbot> Date: 24 September 2015

This meeting

nigel: Proposes we look at Action Items, IMSC, TTML and WebVTT mapping and codecs registry.
... AOB?

group: no other business.

Action Items

action-420?

<trackbot> action-420 -- Pierre-Anthony Lemieux to Follow up with glenn on issue-406 -- due 2015-09-10 -- OPEN

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/420

pal: Glenn has provided the input needed to make progress, that #uax14 feature is unique in
... TTML in that it controls whether the processor uses the uax14 algorithm to determine line
... breaks but it is not communicated in the document. It has to be specified by the profile.
... I'm trying to find out if in general implementors follow the uax14 recommendations so we
... can determine if it should be mandated in IMSC 1. So I have enough information to proceed.
... If you know of implementations and if they follow uax14 recommendations let me know.

nigel: We should take HTML/CSS implementations as a data point.

pal: CSS3 seems to mandate it.

close action-420

<trackbot> Closed action-420.

issue-406?

<trackbot> issue-406 -- #lineBreak-uax14 is never 'used' by a document? -- open

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/406

issue-406: [meeting 2015-09-24] IMSC 1 should mandate that processors shall/shall not implement uax14 or say nothing about it.

<trackbot> Notes added to issue-406 #lineBreak-uax14 is never 'used' by a document?.

pal: Here's a link to the CSS3 spec: http://www.w3.org/TR/css-text-3/

IMSC stuff

pal: I've received a render from an implementor for one of the test suite tests, probably
... the most complex one. It's an MP4 file for track6-frag0-sample1-subs0.ttml
... I stepped through the render and it seems good to me. We should add it to the results and
... use it to encourage others. That implementor now plans to continue with the other tests.
... tmichel, can we post that and add a link to the implementation report?

tmichel: Yes, I've missed that somehow - if you can resend it to me?

pal: I'll upload it to somewhere we can all see it, so you can pick it up from there.

<scribe> ACTION: tmichel to download and upload to implementation report page the test render MP4 [recorded in http://www.w3.org/2015/09/24-tt-minutes.html#action01]

<trackbot> Created ACTION-424 - Download and upload to implementation report page the test render mp4 [on Thierry Michel - due 2015-10-01].

<pal> http://sandflow.com/public/GIC_IMSC_DEMO_H264_30s_20150829_1.mp4

nigel: Last week we said we'd discuss how to proceed with encouraging implementors.

pal: I think a prompt update email from tmichel wouldn't harm. I have another name to add also.

nigel: Is that the best thing to do? Would it help to have a message from the chair too?

pal: We can kick around some ideas offline about this.

nigel: Okay, but for the time being tmichel should send a prompt email?

pal: Yes, and we should also thank the contributor of this test result.

<scribe> ACTION: tmichel Send a prompt/thank you to the potential test result contributors for IMSC. [recorded in http://www.w3.org/2015/09/24-tt-minutes.html#action02]

<trackbot> Created ACTION-425 - Send a prompt/thank you to the potential test result contributors for imsc. [on Thierry Michel - due 2015-10-01].

nigel: And pal will send tmichel a new name/address?

pal: yes.

issue-417?

<trackbot> issue-417 -- #writingMode constraint inconsistent in image profile -- raised

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/417

pal: This is about features and dependent/contained features. #writingMode 'may' be used in
... both profiles, but image profile doesn't actually permit vertical writing mode, so it's just a
... matter of moving it to the individual profiles from the common set.

reopen issue-417

<trackbot> Re-opened issue-417.

issue-429?

<trackbot> issue-429 -- Glyph vs code point in HRM -- raised

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/429

pal: This is a tricky one. The HRM today is written in terms of glyph but what the document
... talks about is code points. You don't know how code points get turned into glyphs until
... the text is rendered because it depends on the font and the context. It's complex.
... I'm trying to find out if the intended term was code points or glyph.

reopen issue-429

<trackbot> Re-opened issue-429.

issue-431?

<trackbot> issue-431 -- Example of documents that fail the HRM -- raised

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/431

pal: This should be pretty straightforward to put together.

reopen issue-431

<trackbot> Re-opened issue-431.

nigel: There are a whole load of pending review issues on imsc 1. Are there any to call out?
... Otherwise we can give a 2 week review period and if there are no queries/problems then
... close them.

pal: I've sent Glenn a summary and really he should review the resolutions.

nigel: I've checked all the check-ins and most of the issues, so they're looking good to me
... at the moment, but I need to double check.

pal: Okay, then I'll look at the now open ones.

TTML and WebVTT mapping document

action-422?

<trackbot> action-422 -- Courtney Kennedy to Update the mapping document status of this document to include feedback mechanism. -- due 2015-09-24 -- PENDINGREVIEW

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/422

courtney: I've done that.

https://dvcs.w3.org/hg/ttml/raw-file/tip/ttml-webvtt-mapping/mappingbetweenTTMLandWebVTTW3C.html

close action-422

<trackbot> Closed action-422.

Courtney: There's another action and an issue that I have proposed resolutions for and then
... there's a whole bunch of others I'm still thinking about, and want to discuss with Andreas.

action-423?

<trackbot> action-423 -- Thierry Michel to Check the license on the ttml to webvtt mapping document -- due 2015-09-24 -- PENDINGREVIEW

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/423

courtney: I figured out how to put the right licence in there.

tmichel: It's true that the right copyright is in there, which is fine. Now the document has
... changed from Editor's draft to Unofficial draft. Is that because you couldn't make it an ED?

courtney: I'm using respec as you suggested and it's hard to control. When I make it an ED it
... gets broken links. Making it a UD fixed the links but broke the licence. The respec guys
... helped me to fix the link and the copyright but the effect is it's a UD.
... If I use the template from tmichel then it puts a wrong link in.

tmichel: Let's discuss offline. In the meantime there's no problem having an Unofficial Draft.

courtney: I'm not sure it matters at the moment. The right copyright statement and no bad links is more important.

nigel: I've also noticed that it says 'Latest Recommendation' which is a bit misleading.

<scribe> ACTION: tmichel tmichel to work with courtney to get the TTML WebVTT mapping to look right with respec [recorded in http://www.w3.org/2015/09/24-tt-minutes.html#action03]

<trackbot> Created ACTION-426 - Tmichel to work with courtney to get the ttml webvtt mapping to look right with respec [on Thierry Michel - due 2015-10-01].

close action-423

<trackbot> Closed action-423.

issue-415?

<trackbot> issue-415 -- TTML2 is confusing -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/415

courtney: I'm a bit confused by this but what I think he's asking for is not to use '2' in the T2VP
... profile for mapping. He also said the constraints are against IMSC 1 which is not correct.
... I changed the name of the profile to TVTT as a response to the '2' issue.

issue-415: [meeting 2015-09-24] Numeral 2 removed from name of mapping profile to avoid confusion with TTML2. Profile is not based on IMSC 1 however note Issue-446 requests this.

<trackbot> Notes added to issue-415 TTML2 is confusing.

nigel: Let's point mike back at this to review.

courtney: Thanks to Mike and Pierre for the detailed review - I really appreciate it.
... Andreas is on vacation and I really want to discuss some of these with him before deciding
... on how to resolve them. Our idea in defining a new profile was just to structure a mapping
... strategy. At least in my mind it wasn't something that you'd publish and author to. It was
... just a way to structure the mapping. The feedback is really interesting though. I want to think about this some more though.

nigel: I had the same take as you Courtney but we should recognise the feedback - if we
... define a profile then people might use it! We should think about how formal we make the profile
... document.

courtney: The goal of splitting the mapping into two parts was 1) to reduce the semantics
... and 2) to convert the syntax. The benefit that Andreas pointed out was that then you always
... have a valid conformance-checkable document at every stage. I thought that was a nice
... benefit of dividing things up this way. Because that was our intent, the profile isn't related
... to IMSC or other TTML profiles so much. It's just about defining a subset of TTML that is
... the same as WebVTT functionality.

pal: It makes perfect sense and is a great process. From a user's perspective, the end result
... is that if a document meets the constraints defined in the document it can be converted
... losslessly into WebVTT, to the temptation will be to author to that.

courtney: That's interesting - it's true, but the converse it not true. Just because you don't
... start with this profile doesn't necessarily mean that the mapping will be lossy.

pal: It will look like a very tempting option to author for lossless mapping to WebVTT.
... If that happens to be a subset of IMSC 1 then that's great because implementors can just
... use a subset of IMSC 1 to target mapping to WebVTT. If it's not then we should see if IMSC1
... needs to change.

courtney: I think we should do that exercise and we may find that we're very close to IMSC 1
... already. Then we should think if it makes more sense to use IMSC 1 and restructure the
... document to discuss the timing conversions that are allowed in IMSC 1. I agree that this is
... a pragmatic approach that needs more discussion and I'd like Andreas to be part of it.

pal: Absolutely. It may be worth just noting that the TVTT profile is a subset of IMSC 1.
... It doesn't necessarily require a major document restructure.

nigel: It may be that we defer this discussion until TPAC to manage everyone's availability.
... Looking at all the issues raised, is there any reason not to open them all?

courtney: I haven't looked in depth at all of them yet. I think the forced display one is a bit different.
... I think that will result in some additional text in our document describing how to mark
... WebVTT as forced. I think it depends on the carriage/wrapper too e.g. HLS, BMFF, HTML web page.

pal: I think we could mention that you may need two or more separate WebVTT tracks.

courtney: In my experience content creators like to deliver the forced subtitles as a separate
... track so they can adjust the timing relative to non-forced subtitles.

pal: That's interesting. So the forced/non forced subtitles are delivered as two separate files?

courtney: Not always but often.
... That's a beyond-the-format question. But we could put in the mapping document how to
... approach that problem in different containers.

nigel: A related point is that WebVTT documents can be constructed for chapters by extracting
... metadata from the TTML div structure for example.

<scribe> ACTION: nigel write up an issue for WebVTT for chapterisation etc. [recorded in http://www.w3.org/2015/09/24-tt-minutes.html#action04]

<trackbot> Created ACTION-427 - Write up an issue for webvtt for chapterisation etc. [on Nigel Megitt - due 2015-10-01].

TTML profile registry.

nigel: Mike has done some work on this, discussions are ongoing on the reflector.
... He's moved it to https://www.w3.org/wiki/TTML/ProfileRegistry
... The main direction to note is that we're merely hosting the short codes and separating out their usage.
... There's also a question arising, perhaps to tackle offline, whether we can create a standalone
... Note to update the Media Registration with IANA.

tmichel: Let me check on that.

<scribe> ACTION: tmichel Check if we can update the TTML IANA Media Registration in a standalone Note or if we need to do it in a Rec [recorded in http://www.w3.org/2015/09/24-tt-minutes.html#action05]

<trackbot> Created ACTION-428 - Check if we can update the ttml iana media registration in a standalone note or if we need to do it in a rec [on Thierry Michel - due 2015-10-01].

nigel: Thanks everyone - we got through a lot, and skimmed past a lot in that meeting, but we're
... 3 minutes over so I'll adjourn for now. See you all next week. [adjourns meeting]

<scribe> scribe: nigel