{Minutes} TTWG Meeting 2020-11-05

Thanks for attending today's meeting. Minutes can be found in HTML format
at https://www.w3.org/2020/11/05-tt-minutes.html.

In text format:

W3C Timed Text Working Group 05 November 2020 Agenda [1]. IRC log[2].

[1] https://github.com/w3c/ttwg/issues/157
[2] https://www.w3.org/2020/11/05-tt-irc

Attendees

Present
    atai, atsushi, cyril, gkatsev, pal
Regrets
    Nigel
Chair
    Gary
Scribe
    gkatsev

Contents

    This meeting
    Interaction between tts:writingMode and tts:direction on paragraph
element. w3c/ttml2#1211
    incoming liaison letter from MPEG regarding ISO/IEC 14496-30
    ending meeting

Meeting minutes

gkatsev: [goes over agenda]
Interaction between tts:writingMode and tts:direction on paragraph element.
w3c/ttml2#1211

gkatsev: we talked about this a month ago

pal: plan was going to file new issues
… a new issue was filed
… but there's a disagreement again
… someone needs to take this on and do it before end of the year

Remove application of tts:unicodeBidi to tt:p element

-> Improve interoperability of tts:direction on tt:p with CSS3
implementations

atai: a step forward would be to get a concrete proposal and then get
consensus that way

pal: I agree with title of issue 1212 but description isn't clear

pal: I'll create one or two pull requests

cyril: we can work on it together

atai: Nigel assumes no one is using a XSL-FO processor
… I think ttpe is using one
… does a joint meeting with CSSWG necessary to conclude this?

pal: I think a PR would be helpful to move forward regardless

gkatsev: a PR would be nice to make progress
incoming liaison letter from MPEG regarding ISO/IEC 14496-30

<atsushi> https://www.w3.org/2001/11/StdLiaison#I

<atsushi> > ISO JTC1/SC 32 (Class C Liaison)

<atsushi> > XML and Dublin Core Ivan Herman Jim Melton

cyril: going over liaison
… people had questions on how to stream TTML with multiple periods in DASH?
… MPEG is seeking clarification with TTWG around usage

atai: would be useful to have Mike Dolan on the call

cyril: request is what terminology should be used

pal: why mention timeline, we should use document time?

cyril: timeline is well understood in MPEG
… a TTML has a timeline and you place things on the timeline
… the mp4 has a timeline
… anchor TTML to mp4

pal: don't use timeline for TTML.
… Say here's how to map document time in TTML to MPEG time
… All it has is time offsets

cyril: still, there's some ambiguity. There's nested time containers in
TTML.

pal: the nesting of temporal expressions is just syntactic sugar
… everything translated to start and end in document time

document time in TTML2

cyril: the term Document Time doesn't seem to be defined in TTML2

pal: document times are expressed on the document temporal coordinate space
… times come out as offsets

atai: the document needs to say what the reference for those offsets are

cyril: [presents proposed changes]

atai: important to circulate outcome to implementers to see if they can
understand it

cyril: MPEG sent similar liaisons to all types of organizations
… so implementers can understand it

<atsushi> (I'm operating webex client on tablet, and it seems I've touched
wrong line.... sorry for mis-operation for muting noise.)

pal: a TTML decoder returns temporal coordinates where each ISD is
… and these temporal offsets are the interface
… How do you take these temporal coordinates and map them to ISOBMFF
timeline?
… This mapping is critical to sync up with audio and video

cyril: I see what you're saying, I'm not sure how to translate it to text
… that's understandable by implementors

pal: make a formula for mapping temporal offsets in seconds into the
ISOBMFF timeline
… it's the same formula for each TTML document
… regardless of the ordering in the TTML track

cyril: how do I refer to this time offset in TTML spec

pal: I think it's called time coordinates

gkatsev: writing up something and sending it to the group might be a good
way to progress
ending meeting

gkatsev: [wrapping up]
… [mentioned thinking about process2020, patent policy 2020, and default
branch name for next time]

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