Timed Text Working Group Teleconference

<scribe> scribeNick: nigel

nigel: Today here at the BBC, alongside Ericsson, we managed a world first, which was an entire TTML-based live subtitling solution from authoring to audience, using EBU-TT Live, EBU-TT-D and DASH.

group: [discussion about sequence numbering and identification] resulting in proposal
... to create a TTML extension spec at a later date for sequence processing.

glenn: I'll take the action to close the sequence numbering and identification issues on TTML2
... and mark them as for TTML.next or something like that.

nigel: Okay, I can't see any reason not to do that; we would need an update to the Charter
... to create a new Rec track spec.

This meeting

nigel: Agenda for today is F2F Meeting, IMSC, TTML. Any other business?

group: No other business.

F2F Meeting.

glenn: Will there be any WebVTT topics on the agenda for the F2F?

nigel: I'll check back with David on that.

action-482?

<trackbot> action-482 -- Nigel Megitt to Propose 12 and 13 jan to ttwg for a f2f meeting in london -- due 2016-10-20 -- PENDINGREVIEW

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/482

close action-482

<trackbot> Closed action-482.

action-483?

<trackbot> action-483 -- Thierry Michel to Create a template wiki page for the f2f details -- due 2016-10-27 -- CLOSED

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/483

nigel: In case anyone has the original wiki page bookmarked please update your bookmark
... to http://www.w3.org/wiki/TimedText/F2F-jan-2017 - the year was wrong originally.

IMSC

action-479?

<trackbot> action-479 -- Pierre-Anthony Lemieux to Refactor the imsc repository in preparation for future versions of imsc. -- due 2016-09-26 -- PENDINGREVIEW

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/479

Pierre: I don't think I'm waiting on anyone for that.

action-484?

<trackbot> action-484 -- Pierre-Anthony Lemieux to Draft a response liaison to dvb thanking them for incoming and requesting more details re timeline -- due 2016-10-27 -- OPEN

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/484

Pierre: I'm working on my proposal for that liaison to DVB.

nigel: Is there anything else to cover on IMSC?

Pierre: No, but I'm working towards a concrete plan for discussion in the F2F, as a good arbitrary milestone. The timing seems right.

nigel: Yes it does.

Pierre: As part of the plan we have to figure out the sequence of events to keep implementers
... and other organisations in the loop so they can review and have their say.
... I think having a good plan by mid January makes sense for where we're going.

TTML1 & TTML2 issues, actions, PRs, editorial actions etc

nigel: There's been a lot of activity this week. What's most important?

glenn: On the font size issue ( https://github.com/w3c/ttml2/issues/200 ) I think there
... was a proposal to make concrete that relative value calculations are in relation to
... computed values, by improving the language. Also to improve the language around
... relative value resolution in the style processing sections.

nigel: I would also like to add the surprise result examples that I constructed.

glenn: I agree.
... Some of this debate led back to a comment in TTML1 that I think we should now
... make normative, and take it out of the note. The last note in ttp:cellResolution is where this
... is elaborated. We might also want to make a statement under tts:fontSize.

nigel: So the relative value discussion is to ensure that computed values can be expressed
... in terms of rw and rh?

glenn: Yes, correct, I'll add that as a note to the issue 200.
... posted as https://github.com/w3c/ttml2/issues/200#issuecomment-256665979
... By the way I think of em as another way of writing % basically.

nigel: In the case of anamorphic scaling you would still say that em is always related to
... the height, right?

glenn: Yes, if you use em in the horizontal part of a two part font size that refers to the height
... of the font never the width.

nigel: We haven't discussed use of rw and rh as measures for fontSize. I don't think it
... causes any particular difficulties but we can now make the font height proportional to
... the width of the root container region. Just noting it.

glenn: I agree.

nigel: I think that concludes on issue 200. Is there anything else we can discuss to assist
... with your editing work Glenn?

Glenn: At the moment I'm processing issues as quickly as I can, so if anyone has any
... questions on technical issues to raise then please do so. I plan to work initially on
... syntactical changes and then on semantic definitions.
... I haven't forgotten about Audio Description by the way Nigel.

nigel: And I haven't forgotten about examples. Perhaps I'll do tts:textShadow next.

Glenn: That would be a good one.
... I have not yet dealt with the background area merging, which I think is an issue on
... tts:border with <border-radii>. Have you proposed some syntax for this Nigel?

nigel: I did not think that syntax was required.

Glenn: I would like to see it as a parameter on border-radii.

Pierre: The proposal is really aesthetically pleasing. Can it be implemented in CSS?

Glenn: The web platform supports SVG so you can do anything that TTML asks for.

Pierre: I'm sure you've been following this - SVG is not supported uniformly across all browsers.

Glenn: The core SVG for TTML rendering is supported fairly uniformly across all browsers.

Pierre: [expresses desire to constrain features to what can be supported on the web platform in CSS]

Glenn: [argues that this is not a scope limit for us]

Pierre: The SVG point sounds like it answers my question: it is implementable in SVG.

Glenn: Yes. I try to rely only on a subset of SVG 1.0 not even any later versions.

Pierre: OK. [discussion of this being a feature of TTPE and testing]
... So should all the presentation semantics be implementable in SVG 1.0?

Glenn: Yes.

Nigel: There's one definite exception, which is the reference to the Web Audio spec; the
... Chair of the Web Audio Group has told me that they are expecting to move to CR before
... us on TTML2. Do we also depend on SMIL for continuous animations in SMIL?

Glenn: Actually no since SVG 1.0 has animation semantics that are based on but not
... dependent on SMIL.
... I don't see a problem with a guideline for the group being that we should avoid
... defining functionality in TTML that is known to be impossible or impractical to render
... using SVG when we are talking about visual presentation.

Pierre: I think that's a good place to start.

Glenn: Profiles might make decisions based on stricter guidelines.

Pierre: Right.

Nigel: I did have a look at the recently merged PRs and add my review comments, all positive
... if I recall correctly. I would encourage everyone to do similarly, and please add positive
... comments as well as negative ones.

Glenn: By the way on https://github.com/w3c/ttml2/issues/105 that you reopened Nigel
... I agree that we need to handle smpte discontinuous so I will add something for that.

Nigel: Great, thank you.
... Having a look at the pulse, I see https://github.com/w3c/ttml2/issues/141

Glenn: The question is if there is a requirement for alt text to be enforced on images.

Pierre: That was a long discussion on IMSC 1 so if you are going to include something
... similar in TTML2 then reuse the same text.

Glenn: I probably would not make it a Shall.

Pierre: It does not say that today.

Glenn: Okay I have no problem with making it a recommendation.
... A quick question: in TTML2 we have both foreground and background images. Can we
... safely assume that any use of background image in SMPTE-TT or IMSC 1 is as a foreground
... image with no additional information presented in front?

Pierre: Yes, certainly in IMSC 1.
... I would recommend not putting information in images that might require some kind of
... video description.

nigel: I would be careful to ensure that if there is a need to use any textual description for
... an image, for example for translating into a braille display or some other accessible
... technology, then we do not prevent that.
... We're at the end of today's meeting. Thanks everyone. [adjourns meeting]