{Minutes} TTWG Meeting 2020-12-17

Thanks to all those who joined our last meeting of 2020. Minutes can be found in HTML format at https://www.w3.org/2020/12/17-tt-minutes.html

Our next call will be on 2021-01-07.

In text format:

   [1]W3C

      [1] https://www.w3.org/

                Timed Text Working Group Teleconference

17 December 2020

   [2]Previous meeting. [3]Agenda. [4]IRC log.

      [2] https://www.w3.org/2020/12/03-tt-minutes.html
      [3] https://github.com/w3c/ttwg/issues/163
      [4] https://www.w3.org/2020/12/17-tt-irc

Attendees

   Present
          Andreas, Atsushi, cyril, Gary, Mike, Nigel, Pierre

   Regrets
          -

   Chair
          Gary, Nigel

   Scribe
          nigel

Contents

    1. [5]This meeting
    2. [6]Recharter: requirement to rejoin group
    3. [7]MPEG Liaison #167
    4. [8]Review of 2020, and look-ahead to 2021
    5. [9]Presentation customisation requirements
    6. [10]AOB
    7. [11]Meeting close

Meeting minutes

  This meeting

   Nigel: Agenda today: reminder to rejoin the group, MPEG
   Liaison, Review and look-ahead, Presentation customisation
   requirements
   … AOB?

   atsushi: Wondering whether to remove the old iCalendar files,
   and to inform everyone that the URL of spec-timeline has been
   changed.

   Pierre: Need to drop off at half past

  Recharter: requirement to rejoin group

   Nigel: After rechartering, which we just did to adopt Patent
   Policy 2020, if you haven't rejoined the group, please do.

   [12]Link to rejoin the group

     [12] https://www.w3.org/2004/01/pp-impl/34314/join

   [13]New Charter

     [13] https://www.w3.org/2020/12/timed-text-wg-charter.html

   Pierre: Only your AC rep can do it for you. Your AC rep has to
   nominate your organisation to join, then you can nominate
   people to join the group.
   … It's not as simple as just clicking that link.

   Nigel: I get thrown by this every time it happens. I wish it
   was simpler and clearer about what is going on!
   … I raised #168 for updating the Charter link on the home page.
   I think that's done now.

   Atsushi: I believe so.

  MPEG Liaison #167

   Nigel: Just to let you know I sent this last night and closed
   the issue.
   … It's archived in member-tt, and we got an acknowledgement
   back that it has been received.
   … Thanks for the reviews. I hope it's useful.

   Cyril: Yes, I think it is useful.

  Review of 2020, and look-ahead to 2021

   Nigel: [goes through slide deck quickly]
   … [14]Pointer to slides
   … And end up with what are your priorities for 2021?
   … Not expecting an answer now, but I'd like everyone to reflect
   on what they can bring and what we can deliver next year.
   … It's been a tough year for everyone in 2020, we should be
   pleased by what we have delivered.

     [14] https://www.w3.org/AudioVideo/TT/wip/ttwg-in-2020.pdf

   Pierre: The holy grail is alignment with CSS. That's a real
   pain point, and it's entirely within W3C members' ability to
   do.
   … Especially in the case of some of the features that the
   broader web community (fillLineGap!) where users are looking
   for
   … that functionality for general use, not only subtitles.
   … If we manage to achieve that it would be great for the web
   community in general.
   … The other one that is a lot more complicated, and maybe not
   for W3C, is more interop testing of subtitle implementations.
   … We've created a lot of tests, but don't have a lot of interop
   testing, and many implementers are not W3C members.
   … It's something for the industry as a whole. Certainly solving
   the CSS issue would be my main goal.

   Andreas: On this point, do you think it would be beneficial to
   link interop testing, e.g. in HbbTV or ATSC, to W3C or is that
   out of scope
   … for our group?

   Pierre: I'm not sure. I think it's valuable but I'm not sure
   W3C is the right place.

   Nigel: I think it's worth bringing up, thinking especially
   about WPT and the way the same orgs do that and W3C specs.

   Mike: Worth bringing in DASH-IF and dash.js? It's an open
   platform and probably has, via imsc.js, the best conformance
   with IMSC as far as I know.
   … They also have a content validator, that could be integrated.
   … It's DASH-centric, but the world's going there anyway.

   Nigel: I'd be interested in completing the AD work, but also
   … seeing if others want to standardise the format of data for
   passing customisation information into players.

   Pierre: Sounds like a great topic to me.

   Gary: I'd like to finally get back on the horse with WebVTT.
   … I know for me it's a bit harder because I'm basically the
   only person working on it.
   … Without having other people working on it too it's harder to
   remember and justify looking at things.
   … As for working with CSS WG, that would be good. We tried to
   meet them at TPAC but didn't end up doing it so it
   … would be nice to schedule some time with them. We have things
   that it would be worth working on with them.
   … The player options for caption styling is also something
   interesting.

   Mike: By caption styling you mean like user preferences?

   Gary: Yes

   Mike: I don't know if you shared CEB35 with others but there's
   a CE document we did that maps the common user preferences
   … onto 708 and IMSC1.

   Nigel: I didn't, I've been trying to implement it for IMSC 1
   and would like to share as an end of year demo if that's okay?

  Presentation customisation requirements

   Nigel: Following on from that, I've been trying to implement
   some of the stuff in CEB35 as modifications to imsc.js.
   … [shares prototype renderer based on imsc.js with sliders for
   user preferences]

   Cyril: Why is this relevant to IMSC?

   Nigel: The options object passed to the player is something
   that captures the user preferences, and could be a candidate
   for standardisation.
   … You'd want different players to behave in interoperable ways
   given the same user preferences.

   Cyril: I like the idea. An aspect that could be interesting is
   the challenge of getting the content authored in a way that
   … works with the preferences. Brainstorming, I'm wondering if
   we could use this options parameter could be included in
   delivery requirements
   … saying "must look okay when options.value changes from x to
   y"

   Nigel: Good point, we had an exact example of that with
   different lines being delivered in different regions, which we
   had to post-fix.

   Mike: It's tricky to scale the region because you have to have
   heuristics. If you keep text in the same region
   … you should get consistent results.

   Cyril: Unfortunately people do different things!

   Mike: I've seen different behaviours. There may be some edge
   cases. Keeping the regions together may need explanation.
   … This is pretty cool - having some standard way to do this
   would be interesting.
   … Some people don't see colours well or have both hearing and
   visual impairments.
   … It's more of an accessibility requirement at least in the US.
   … When you scale you assume you want it to go towards the edge
   of the display region. That's not unreasonable,
   … but you've assumed a lot of interesting behaviour that varies
   from implementation.

   Nigel: Not so much. The region isn't moving, but displayAlign
   and textAlign are staying the same.

   Andreas: On the one side you have a UI that allows users to
   customise presentation. This is already done, as part of
   hardware or OS
   … like iOS or Android, that have the ability to change the
   subtitle presentation where they can detect it.
   … Then you have the format like IMSC that represents the
   author's intent, and in between a lot of things are happening.
   … As Cyril noted, when user preferences are applied it's not
   IMSC or the format any more, it's user preferences and the
   language
   … like HTML and CSS. This is already happening, for example in
   Germany using HbbTV there is a customisation feature where the
   input
   … format is a subset of IMSC, and it is well defined what you
   can do so it all works fine.
   … There needs to be some work done so the format and the UI
   work together to avoid destroying the authors intent.
   … You may need to author in a way that allows the preferences
   to be applied.
   … And it needs to be format agnostic. In iOS for example it
   needs to work on WebVTT, SRT and variations of TTML.

   Mike: What's your intent for this activity?

   Nigel: I'm using it as a test-bed at the moment, on the BBC
   fork. imsc.js is a reference player so if there's a
   standardised way to
   … pass user preferences in, then it would be a good place to
   put it.

   Mike: Colour customisation is important too.

   Nigel: Thinking about a colour map to allow users to define
   their own palette.

   Andreas: Some user preferences allow users to override
   settings, but at their own risk.
   … Even then you need to make sure that background and
   foreground colour are applied in a set, because if you change
   your foreground
   … colour preference and then there's a white background then
   you would see nothing.

   Mike: Preferences are set by the viewer, so they can fix it.

   Nigel: I tend to prefer giving tools to avoid users getting
   into that situation.

   Andreas: Others might be interested too, APA WG and some
   manufacturers.

   Mike: In the US this is driven by FCC in a regulatory manner.
   … At least in the US there is specific behaviour that is
   expected. We need to be conscious about how we do it as
   engineers,
   … and keep in mind regulatory requirements.

   Cyril: It's a BBC fork? I looked at GitHub, is it there?

   [15]BBC fork of imsc.JS

     [15] https://github.com/bbc/imscJS

   Mike: I'd to share this with some colleagues too

   Nigel: I'd be happy to talk with them if they're happy to do
   that. This work is on a branch in the BBC fork at the moment.

   Cyril: I'm interested in this activity - it would be great to
   have a video of this demo, if you can share that?

   Nigel: Interesting thought!

  AOB

   Nigel: You wanted to know about removing the old ics files?

   Atsushi: Yes just housekeeping. We have both composited and
   individual meeting files.
   … There are around 20 files in the directory. We could remove
   all the 2020 calendar invites. Is that fine for all?

   Nigel: Makes sense to me. No need to keep them for posterity.

   Atsushi: In any case they'll be there in an old commit.

   Nigel: Exactly.
   … I think just do it!

   Atsushi: Yes
   … Another point is just info to share. To include the spec
   timeline in the onboarding email new members receive when they
   join TTWG,
   … there's a customisation template. Nigel included a link to
   spec-timeline in that but we may forget to update the link year
   by year so now

   <atsushi> [16]https://www.w3.org/AudioVideo/TT/wip/
   TTWG-spec-timeline.html

     [16] https://www.w3.org/AudioVideo/TT/wip/TTWG-spec-timeline.html

   Atsushi: we have removed the 2019 year from the URL and used a
   non-dated link, so please be careful to check if you have it.
   … It is ^^ now

   Nigel: Good stuff, thank you!

  Meeting close

   Nigel: Thanks everyone, wishing you all a good break and new
   year if you're celebrating it. Our next meeting will be on 7th
   January.
   … I'm wondering about user presentation customisation as a
   potential topic for a future f2f by the way.

   group: [general mutual good wishes for the holiday period and
   the new year]

   Nigel: [adjourns meeting]


    Minutes manually created (not a transcript), formatted by
    [17]scribe.perl version 124 (Wed Oct 28 18:08:33 2020 UTC).

     [17] https://w3c.github.io/scribe2/scribedoc.html