See also: IRC log
<scribe> scribe: nigel
Glenn: [needs to leave after 45 minutes]
Nigel: Today we have a bit on
TPAC/CSS, IMSC, TTML, HDR in PNG, WebVTT
... Any other business or specific items to cover?
group: [no other business]
action-497?
<trackbot> action-497 -- Nigel Megitt to Invite csswg to joint meeting at tpac 2017, with list of topics. -- due 2017-06-15 -- OPEN
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/497
Nigel: I've begun to pull together the list of CSS features, at:
Nigel: That's a work in progress
- input very much welcome.
... I've lifted the existing points - I don't know all the
information. For example with
... fontShear do we need it in CSS?
Glenn: fontShear is frequently
used in Japanese subtitling, with more than one angle
being
... used in the same document. CSS does not support that at the
moment, though you
... could support it in SVG with 2D transforms.
Nigel: Let's keep this scoped to CSS, since that's this document's purpose.
Dae: "Ignore" means that's the fallback, in this case CSS can't do it at all.
Nigel: Thanks for the
explanation.
... Please could I ask everyone with an interest in this to add
detail where they know it, and
... if there are any missing features, add them.
... Some more context here - when TAG reviewed TTML2 they
offered to facilitate a
... joint meeting with CSSWG, and we invited CSSWG to our
meeting at TPAC. In advance of
... that TAG has invited me to discuss these issues with them
around 26th July so there's a
... deadline for sorting this document out so we can share it
with TAG. I suggest getting
... this to a shareable state by Friday 21st July.
Pierre: We need to add multiRowAlign too.
Nigel: Thanks I wasn't sure about
that. Also you have some codepens too? It would be
... useful to add those in please.
Pierre: Ok
Nigel: Regarding TPAC, Thierry created a wiki page for it
Thierry: The dates are 6-10 November, and we are meeting on 9th and 10th.
Nigel: Plus we are meeting jointly with the Media and Entertainment IG on Monday 6th.
Thierry: We have published the
IMSC 1.0.1 CR today
... This has triggered the Call for Exclusion of 60 days until
Sep 11.
... I think we should target, if we have enough
implementations, a PR publication by Sep 15.
Nigel: Can we do a PR earlier and just delay Rec until the end of the CfE?
Thierry: If we want to go fast
track we could ask for this - I have to check in the
Process
... how to do that. When do you think we can do that? In terms
of implementation.
Nigel: Now we have published the
CR I will go through the BBC processes to add our
... implementation.
Glenn: I could add the new features to TTV as a second implementation.
Thierry: I'll send you the exact
procedure if we want to go into PR review earlier than
the
... 60 days in case we meet the exit criteria earlier. Then
it's another 4 week to Rec, so we
... can envision a Rec before end of October if we keep to that
schedule.
Nigel: Well done getting here, thank you everyone!
Thierry: We've been using the
2017 Process and therefore this allows AC review starting
... after CR publication (it used to be PR) but now your AC Rep
can from today review the
... document and agree to Rec publication.
... If we publish a new CR then it might trigger a new review.
Your AC Rep will have
... received a review form.
Glenn: I did a bit of
administrative work after last week's meeting to organise the
TTML1
... issues into milestones, with the test related ones in a
separate milestone.
... I made sure that every issue had a milestone.
Nigel: On Tuesday I found 2 open issues with no milestone.
Glenn: I'll check that.
... In TTML2 I'm soon going to ramp up work to resolve existing
issues, hopefully in a week
... or so.
Nigel: Last week I opened a pull
request to address some formatting and a "silly" mistake
... about the HDR thing.
Glenn: I'll look at that and merge it today.
Nigel: I also raised an issue for
supporting font variation settings, with the use case of
... number presentation.
https://github.com/w3c/ttml2/issues/400
Glenn: Now that you've raised
that I'm keen to go ahead with it using the syntax in the
... CSS fonts module. One of the criticisms of this feature is
that it is tied to TrueType and
... OpenType, but practically pretty much all, or anyway most,
fonts in use fit this category. If an
... implementation provides a semantic mapping for other font
formats then it can still be
... used so I'm not concerned about that.
... Just in case we get similar comments in the future.
... One more thing: since it is a kind of open-ended
extensibility mechanism for directing
... the font and layout system to do something special we may
want to create a registry
... to give guidance to authors and implementors about which
features are supported,
... otherwise there will be no expectations for specific
support, unless we define profile
... feature designator.
Nigel: So in my case I would want
the tnum value to be variable, so there would be a
... feature designator for that?
Glenn: Right, though we would
likely not publish a new designator for every individual
... feature. It could be that we put them in the extensions
space for designators. It is
... something to think about.
Nigel: Any other TTML comments right now?
Glenn: There's been some
reflector conversation initiated by Viacom.
... I would suggest Nigel points the commenter towards your
work on live contribution of TTML.
Nigel: Good idea.
Glenn: It doesn't make much sense
to do roll-up for pre-canned captions.
... There are two features that I've considered adding in the
past but never done so:
... end="indefinite" and begin="now" for the live use case.
Nigel: The EBU-TT Live spec works around their absence okay.
Glenn: mediaOffset could be used for the begin="now" case.
Nigel: Actually I think it can't,
but that is very similar to the reason I wanted to
introduce
... a tt element attribute "clipBegin" to carry the same
meaning.
Thierry: The Wide Review of TTML2
started when the publication was released around 10 days
ago.
... I have sent calls for review to all the W3C groups and the
external bodies, with one
... remaining for the Japanese IPTV Forum. I need to send that
and also I need to work on the
... TAG questionnaire. So now I will be tracking those comments
also for TTML2.
Nigel: Fantastic, thank you for doing that (and reminding me about it!).
Thierry: The review ends Sep 30.
Thierry: Silvia has done a lot of
work on editing the WebVTT draft of the CG mainly fixing
... broken links and linking to W3C HTML spec instead of
WHATWG.
... The draft was published today on TR as an ordinary draft.
My understanding is we should
... trigger another draft soon for wide review.
Nigel: It's also worth noting
that Silvia has now joined the WG and has effectively taken
on
... shared Editing role with Simon.
Pierre: Who does Silvia represent?
Nigel: It's CSIRO in
Australia.
... Any progress on looking at the Dispositions for the
previous Wide Review?
group: [silence]
Pierre: All the comments filed on
GitHub have been resolved. The only outstanding question
... I know of is related to the supported bit depth of PNG. PNG
supports only 8 and 16 bits,
... and nothing in between. Some folk have pointed out that 8
is too little and 16 is too much.
... 8 is too little for general use, but it's okay for
subtitles and captions. Possibly we should
... extend the note on bit depth to say that. I'm reluctant to
state that until someone has
... a concrete issue filed. I think this would impact only the
informative note. I think this
... is ready to be published.
Feedback on HDR compositing #400
Nigel: I forwarded some feedback
to TTML2 that may be relevant here.
... I think the overall comment is that in PQ both narrow and
full range are in use so the
... spec should support that. I am not sure if that also
applies back to the PQ in PNG note.
Pierre: The note only deals with
full range quantisation. We could conceivably add a row
... for narrow range - technically that is easy to add to the
document. I suggest that we
... keep this a living document, and address that point if and
when it is raised.
Nigel: OK - the point raised on
TTML2 was about the range of the video rather than the
... range of the image.
Pierre: Yes, though in a simple
implementation a copy or blit of full range subtitle
pixels
... to narrow range video you would not get what you expected.
That is less of an issue for
... subtitles and captions.
Nigel: Why is that?
Pierre: If you get the luminance a bit off then it is not orders of magnitude difference.
Nigel: Great, shall we do one
final review pass of the document and the comments before
... proposing to publish it next week?
Pierre: Sure. And Thierry if you
could look at it and check its suitability for publishing
then
... that would be useful.
Thierry: Okay I will look at it today or tomorrow.
Nigel: Thank you.
Pierre: I've been updating the
tests and have just pushed a pull request to address the
... spaces before line breaks issue we discussed last
week.
... There are 15 issues against the IMSC tests, all but that
one being cosmetic, but that one
... is important to get right. It affects only a few tests, but
as you know Chrome, which is
... used to generate the renders, seems not to suppress spaces
before explicit line breaks
... if those spaces are within a span. That's also something
I've added to the wiki about CSS
... and TTML compatibility. We really should understand if it
is a Chrome bug or a divergence
... between CSS and XSL-FO behaviour. To me it looks like a
Chrome bug.
Nigel: Me too.
Pierre: It means that any implementation using CSS needs to include some extra manual work.
Nigel: Will you add the IMSC 1.0.1 exit criteria tests into the imsc-tests?
Pierre: Yes when I have
confirmation from implementations that they're right.
... It would be just a copy and paste since the format of the
tests is the same in both
... imsc-tests and the IMSC 1.0.1 exit criteria tests.
Nigel: We seem to have completed our agenda in an hour, which was secretly my plan all along, so thanks everyone! [adjourns meeting]