W3C

Timed Text Working Group Teleconference

17 Jan 2019

Agenda

Attendees

Present
Cyril, Nigel, Glenn, Andreas, Gary, Matt, Pierre
Regrets
Chair
nigel
Scribe
Cyril

Contents


<scribe> scribe: Cyril

This meeting

<glenn> back

nigel: we can discuss TTML Profile registry if needed

<glenn> no

nigel: seems not
... we can discuss feature requirements
... I added a label to the repository
... then we can discuss the joint F2F
... it does not have an agenda yet
... there has been activity this week but nothing specific marked on agenda for TTML2 or WebVTT?
... I have some AoB items
... group homepage and weekly agendas

glenn: I see that Matt from Ericsson has joined us

Matt: Matt Simpson, Red bee media, provider of captioning, audio description and dubbed media
... involved in EBU nearly 8 years

TTML Profile Registry Actions, Pull Requests and Issues

nigel: nothing to discuss

TTWG Future requirements

nigel: I've added a label for "has a business test"
... last week we had a reminder that we want to have tests for new features
... the equivalent for requirements is "has a business test"

<nigel> https://github.com/w3c/tt-reqs/issues/15#issuecomment-455116509

nigel: I've added an example to issue #15
... the point is to make sure once we have the feature, is that it passes the business test

glenn: the phrase "non-unicode image" is a strange construction
... a character or a symbol that has no unicode character, e.g. gaijin characters can be found on Wikipedia
... when we introduced images in TTML2, this use case was mentioned as a possible driving force
... so I'm supporting your statement with a tweak

pal: I added a comment because I think there is confusion between what you had in mind and the issue on IMSC,
... I think there are 2 use cases, with different requirements, even if they end up with the same solution
... I don't know what you want #15 to be
... re images in IMSC, I will ping the commenter because I need to understand what they have in mind

nigel: point well made regarding merging the 2 points
... I'll take an action to separate the 2 requirements: text profile vs image profile

pal: I would say one is transmitting complete programs and the other one supplementing graphics

cyril: also, to me there is a difference between displaying an image and displaying an image inline

pal: Digital Cinema supports images, but for inline graphics, it works by embedding a font file
... it integrates well with the existing rendering

nigel: the task of converting SVG into a font file is not to be underestimated (perhaps I don't know enough about it)

cyril: you can embed SVG in font files

https://helpx.adobe.com/fonts/using/ot-svg-color-fonts.html

glenn: inline images in IMSC is not supported because the profiles are distinct
... but the idea of using SVG to store a glyph outline is certainly feasible
... there was the SVG font format
... but not implemented and people moved away
... I prefer using the idea of embedding a font than images
... the embedded font can be extremely small

nigel: one of the reason I think it did not meet the requirements
... is because you need to map your input image format into a font
... it's not what I've been asked for as a requirement

glenn: if you put an image in, it does not scale well
... it's inadvisable to use raster images for that

pal: in addition to resolution scaling issues, if it's a character, there is no special case
... the twitter character could be rotated, etc.

nigel: SVG is preferred because it scales well

cyril: this is very similar to emoji fonts

nigel: no because they have assigned unicode characters

glenn: nothing in TTML prevents the use of non-BMP character codes
... we should have access to all characters

pal: including private areas

glenn: private use area

cyril: it would be good to clarify if the images are multi-colored or single-colored

nigel: the private area would work but that's not standard
... we envisage fully colored images

ack

scribe: I just wanted to raise the presence of that label and encourage people to write a testable business statement, verifiable once it's been implemented

Joint f2f meeting with EBU Timed Text, Feb 1 2019

glenn: since we are working closely with EBU, I presume you posed the questions about their requirements

nigel: yes, EBU has visibility that the window was open for new requirements
... also because of the shared membership
... the agenda is looking light at the moment
... thanks Andreas for organizing the 2 diners
... I take the action to sketch up the agenda

TTML2

nigel: nothing to discuss

WebVTT

nigel: nothing to discuss

AOB

nigel: if people have topics to discuss at the F2F, they should raise them

glenn: I would like to have a discussion on TTML2 2nd Ed vs TTML 3
... we need to have guidelines or understanding on how it will go
... I would like to bring the idea of modularization
... to make TTML3 lightweight version to introduce modularization

nigel: regarding AOB, first one is the home page
... the source for the group home page is now served with github pages, just for that home page
... we have direct control of it in the TTML repository
... I've done some cleaning this week

<nigel> TTWG Home page

nigel: the Emmy has been pulled off
... I've added the code of conduct and put a link to github, meetings and the pulse of github
... everything for WebVTT tests

atai2: why do we still need the issue and action trer?
... is it for past issues?
... we should mark it as deprecated

nigel: good point, it should go, we don't need it, because we can do everything with GitHub
... about our meetings and the agendas
... I switched to a light email on the basis that less is more
... but it may still be too much
... I'm open to feed
... I used to have upcoming future meetings, I've taken that out
... each meeting is an issue so I've put a link to the project board

cyril: I could live with a very simple email confirming that the meeting happens with a link to GitHub

nigel: [group silence]

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/01/17 16:49:28 $