Timed Text Working Group Teleconference

plh: I can't stay for long but want to note that the IMSC 1 Rec was published today
... https://www.w3.org/TR/ttml-imsc1/ is REC
... https://lists.w3.org/Archives/Member/w3c-ac-members/2016AprJun/0012.html
... Karen told me this morning she still wants to do the press release but unfortunately
... it won't happen this morning due to other events.
... I did not put last week's stuff in, because I received it from Pierre too late yesterday,
... so it got published as is today.

nigel: Fantastic, well done everyone, great news!

glenn: An announcement message did go out shortly before the meeting.

plh: Congratulations again, apologies I cannot stay for the rest of the meeting.

This meeting

nigel: I don't think there's much to discuss on Charter, but there's stuff to discuss re TTML1 and TTML2

<scribe> scribe: nigel

nigel: AOB?

group: no AOB

Action Items

glenn: I need to write up some of these - some have been taken care of one way or another.
... No specific progress to report on them this week.

action-459?

<trackbot> action-459 -- Thierry Michel to Process the PR at https://github.com/w3c/tt-profile-registry/pull/3 and post for review to the reflector -- due 2016-04-21 -- OPEN

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/459

tmichel: Thanks to plh the pull request has been merged while I was away.
... The follow up is for the group to review the draft.
... Plh told me you want to split the table?

nigel: Yes, I don't mind who does it but the table is a bit unwieldy now so I think it should be split.

tmichel: Shall I do that?

nigel: I'm happy for you to do that.

tmichel: Okay I'll do that after the call and send around a call for review probably tomorrow.

action-459: [TTWG meeting 2016-04-21] agreed to split the table prior to sending for review.

<trackbot> Notes added to action-459 Process the PR at https://github.com/w3c/tt-profile-registry/pull/3 and post for review to the reflector.

mike: There's still an action to replicate the media type information from Annex C of
... TTML1 and add the codecs parameter and update the IANA registration. Is that still
... needed?

nigel: Yes please.

mike: So the process is to make a Pull Request and then request review?

glenn: Which document?

nigel: https://w3c.github.io/tt-profile-registry/index.html
... Yes, we can be flexible about this, but for example making a pull request and discussing
... it in a meeting prior to merging would be fine.

mike: Okay I'll do a pull request against the profile registry and then we can all look at it.

nigel: Fantastic, thank you Mike.
... I also propose to merge Cyril's TTML versions document into this too, as a separate exercise.
... That material is more descriptive.
... https://dvcs.w3.org/hg/ttml/raw-file/5ee90ec0d897/ttml-status/index.html
... Looking at that, there's so little content in there that we maybe just add new
... descriptive sections and leave Cyril's where it is.

IMSC

nigel: Following on from the Rec publication noted at the top, are there any other points?

tmichel: I noticed yesterday that we don't have an acknowledgments section. It's nice
... to have so maybe we should put a placeholder for the next version of IMSC.

glenn: I suggest we create an errata document to start with and add an ack section to
... that as a placeholder. I think this was probably an editorial oversight.

<scribe> ACTION: tmichel Add an acknowledgment section to the IMSC 1 errata section. [recorded in http://www.w3.org/2016/04/21-tt-minutes.html#action01]

<trackbot> Created ACTION-461 - Add an acknowledgment section to the imsc 1 errata section. [on Thierry Michel - due 2016-04-28].

pal: Philippe put the github copy of the errata in the wrong place on github, so I need to
... move it.

<plh> wrong place?

tmichel: Okay, move it and then I'll make a pull request.

pal: Can we put an issue on github to add the acknowledgments?

tmichel: Who should be acknowledged - just group members?

glenn: Only people who have made some kind of contribution to the spec should be
... acknowledged.

nigel: I've created https://github.com/w3c/imsc/issues/177

glenn: You could also list group participants, but not random people who are not
... engaged or have not contributed something.

tmichel: If we want other people, e.g. from the community group, let me know. We
... could also acknowledge people for filing implementation reports, again, not mandatory.

glenn: If people have contributed in some way to the process or the spec then I think
... they should be acknowledged.

nigel: +1

pal: This work began from a DECE contribution.

glenn: I'd acknowledge the group, unless you choose to pick named individuals. That would be fine too.
... Take a look at what we did in TTML1 - I think it's instructive.
... I made special mention to a few people who contributed more than the average share.

nigel: Who's going to start work on this?

tmichel: I will.

nigel: Thanks.

pal: Based on our discussion last week I sent plh too late yesterday a document that's
... a roadmap to the IMSC family of specifications, addressing what we discussed last week,
... that "latest version" might point to one of several possibilities. I created that document
... and a pull request. It did not make it to this release of IMSC 1, but my suggestion is
... still to accept the pull request and leave that document there. It's useful, and if we
... ever do a 2nd edition then it will be there ready to go.
... It's a really short document, that lists the current IMSC documents. Right now there's
... only one.

glenn: Looking at TTML1, the Latest version points to TTML1, but it could be updated

nigel: We discussed this last week and decided to use a separate document to dereference
... all the different possible branches that a spec can take, from 1st ed to 2nd ed, or v2 etc.

TTML

nigel: We have a number of things to cover: background image, time expressions and audio description.
... Let's begin with background image

https://github.com/w3c/ttml2/issues/157

glenn: I was in the process of implementing the semantics for the IMSC background image
... mechanism, based on SMPTE ST2052, and I realised that there was a discrepancy
... between what we ended up with in IMSC and the normal CSS/XSL-FO semantics.
... Here's the scenario: In CSS if you put a background image on a div element, let's say,
... and there's no content in that div, then you will not see any background image, as
... I demonstrated in the sample that I attached.
... If you manually size the content rectangle of the div or insert some content that causes
... it to be sufficiently big then the background image will become visible (possibly clipped).
... In SMPTE-TT it appears to have been dealt with by stating that the minimum width and
... height of the area of the div should be at least as large as the size of the background
... image. I didn't notice this difference to CSS before. The problem we have today is that
... now we're adding image support to TTML2 and we have IMSC 1 published we have
... to deal with the effects of that variation somehow and come up with something that
... makes sense in TTML2. That's the basis of the problem.

pal: In TTML2 images are an element, right?

glenn: There are 2 ways to use images. One is a content image using the image element,
... which definitely has a size that's equal to its intrinsic image size unless it's otherwise
... manually sized. If you were to use the image element to map the SMPTE backgroundImage
... attribute into TTML2 then this won't be an issue.
... In TTML2 we also added the backgroundImage style property, intending it to work like
... CSS/XSL-FO. We could solve this by putting some language somewhere saying that
... mapping from IMSC 1 image should be via the image element.

pal: Yes, why not?
... We have 2 choices. We could use exactly the same semantics in TTML2 as in SMPTE-TT;
... the other option is to map to the image element.

glenn: The other option is to use different semantics in TTML2 than in IMSC and adopt
... the CSS semantics, and require that the element is sized explicitly using ipd and bpd.
... That would work.

pal: Ideally using TTML2 it looks like the "right" way to do image as content is to use
... the image element?

glenn: Right, the backgroundImage isn't intended as being used for content.

pal: I'm thinking that we indicate that it should be mapped to the image element in TTML2.

glenn: I agree. We probably need to create a placeholder document to deal with such
... mappings.

pal: I'm happy to start work on that, and have that as the first item in it.

glenn: It could be maintained as a separate document or we could merge some of it
... into IMSC 2. We can make that choice later.

pal: Yes.

glenn: What I ended up deciding to do is to map smpte:backgroundImage to the image
... element because it wants to behave like that.

pal: Ok.

glenn: I'm tentatively using the same semantics as CSS for backgroundImage in TTML2.

pal: In CSS the name is background-image, right?

glenn: Correct.

pal: I guess we'll have to teach people!

glenn: Yes, it could be a source of confusion for authors.

pal: What's the name in XSL-FO?

glenn: backgroundImage. There's also an externalGraphic that works a bit like in SVG as
... an image. In SVG you can also apply a pattern as a fill, using an image, which works
... more like background-image.
... We'll continue to have both kinds of concept, a content image and a stylistic
... background image.

mike: The discussion came as a bit of a surprise to me because noone has looked at this
... for a while. I think we need to sort this out. The overloading on the name is unfortunate
... but it really was intended for content. I'm not really happy with any of the options but
... this is the least bad.

glenn: IMSC doesn't say anything about this right now, but the reference to SMPTE suggests
... that the normative language in SMPTE-TT applies.

pal: Yes.

glenn: You might want an errata for IMSC 1 that notes that applicability.

pal: Also note that it's not just the backgroundImage attribute but also the semantics
... that surround it.

glenn: Absolutely. One has to assume that that's true otherwise there's insufficient text to describe it.

pal: Do you want to file a quick issue?

glenn: Okay I'll do that.

pal: I think we're going to have to deal with this in the next version of IMSC. We have
... two possibilities. We can create a document where we track these - that's probably best.
... I'll refactor the repo to include an IMSC 2 draft, and note this in there.

glenn: Or do what we did in TTML and add a new repo.

pal: Then you can't easily move bugs from one to another. I'm thinking of having
... multiple subdirectories, one per big version.

glenn: You're the Editor!

nigel: Yes, that's fine.

pal: What about also adding a note to TTML2 for users of backgroundImage, warning
... about the mapping from smpte:backgroundImage?
... That might be helpful for IMSC and everyone else to reference.

glenn: Okay, it doesn't hurt to add that as a comment. Or add an informative mapping
... section.

pal: That would be great.

glenn: For mapping from other versions or profiles of TTML.

<scribe> ACTION: gadams create issue on TTML2 to add "Mapping from other Versions and Profiles of TTML" [recorded in http://www.w3.org/2016/04/21-tt-minutes.html#action02]

<trackbot> Created ACTION-462 - Create issue on ttml2 to add "mapping from other versions and profiles of ttml" [on Glenn Adams - due 2016-04-28].

nigel: I'd also like to raise https://github.com/w3c/ttml1/issues/210
... This came about from implementation work where it appears that leap second
... syntax is permitted even when it has no meaning.

glenn: I propose in TTML1 to add a section to §10.3.1 matching Nigel's proposal, and
... in TTML2 then we can strengthen that.

nigel: That sounds good.

glenn: When I process it on TTML1 I'll do it on TTML2.

nigel: I'll add an issue on TTML2 pointing to the TTML1 issue then.

<scribe> ACTION: nigel Add an issue on TTML2 regarding 60 second minutes to point to the TTML1 one. [recorded in http://www.w3.org/2016/04/21-tt-minutes.html#action03]

<trackbot> Created ACTION-463 - Add an issue on ttml2 regarding 60 second minutes to point to the ttml1 one. [on Nigel Megitt - due 2016-04-28].

nigel: I'd also like to signal my intent to specify and implement audio description in TTML2.
... The idea is to add some audio style attributes that could be subject to animation
... and use included or referenced audio files to generate an audio description

<glenn> https://www.w3.org/TR/REC-CSS2/aural.html

nigel: presentation experience that aligns with current practice. We also intend to
... implement a prototype of this using web audio.
... Thanks for that link Glenn, I'll check if that includes useful semantics I'll reference them.

glenn: That spec probably covers a lot of what you want to do already.

nigel: I may need to cross-check that with Web Audio as well.
... Apologies I've taken us 4 minutes over. Thanks all [adjourns meeting]