Timed Text Working Group Teleconference

12 Mar 2015

See also: IRC log


mike, pal, tmichel, nigel, dsinger
glenn, Andreas, Frans, jdsmith


<trackbot> Date: 12 March 2015

<scribe> scribe: nigel

This Meeting

nigel: I think given those present we can best cover the SMPTE liaison, IMSC 1 test suite and the F2F agenda

tmichel: We also need to mention the Sapporo TPAC opportunity for a F2F

group: no other AOB

Action Items


<trackbot> action-378 -- Pierre-Anthony Lemieux to Draft smpte request including dispositions from action-369 and backgroundcolor initial value question -- due 2015-03-05 -- OPEN

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/378

pal: I've drafted this and sent it round.

mdolan: The initial values for the root container were neither defined nor specified as
... implementation dependent. So SMPTE went about setting initial values. It's unfortunate
... that they're not the same as what's in TTML1SE.
... Timing wise, SMPTE 2052 was originally drafted before TTML1 was published as a Rec.
... I don't recall what the reasoning was, but that's the history.
... More importantly, IMSC 1 came about from a request from DECE, based on TTML1 plus
... some constrained extensions from SMPTE. SMPTE also added for the image handling
... to be added to TTML2 and IMSC 1. So I think there could be a misunderstanding that either
... SMPTE or DECE wanted IMSC 1 to match identically. That may not have been the intent.

nigel: Are you saying that we could just make the IMSC 1 initial values the same as TTML1 and be done with it?

mdolan: That would be my proposal, to remove the initial values text from IMSC 1.
... I'd like to validate this back with DECE and report back next week.

nigel: Does this mean we don't need to bother sending this to SMPTE?

mdolan: Depending on the answer, yes. It may be better to go to DECE because SMPTE
... would be going back 10 years and the history is probably not well remembered, and is complex.
... Even if we got an answer it may not be helpful.
... I propose that I'll investigate and get back and we may need to ask DECE, or nobody.

pal: I like that plan, because I've certainly heard that not aligning with TTML1 is a bad idea.

nigel: +1
... I'll leave action-378 open in that case.

mdolan: If I get an answer back sooner I'll post it to the reflector.


nigel: There's been some offlist discussion. It's looking likely that there'll be an agenda item
... to talk about the plan for creating the TTML<-->WebVTT mapping document, rather than
... to talk about the substance of it.
... It's good news that there's enthusiasm for it, and we need to open that up to the right
... community.

pal: One quick request: I may not be able to be there all day Friday, so I'd appreciate it if
... we can cover IMSC 1 topics on the Thursday and TTML2 also.


nigel: Since last week Courtney has added her name to the attendees.
... She's likely to be interested in VTT topics and only be there on the Thursday too.
... Given that we're not talking about substance with the mapping, but planning, it should
... be maximum 30 minutes.
... I'm also wondering if we should explore inviting the TTCG to that agenda item, since
... they have a lot of expertise in WebVTT.

pal: Thanks for the prompt - I've just added my name to the attendee list.

nigel: Thierry, do you want to mention some details about Sapporo?

tmichel: It's worth noting the dates of TPAC, which will be in Sapporo, Japan at the end of October.
... We've been asked if we plan to meet there. The question is do we want to meet or not?

nigel: We have a deadline for responding to that survey too.
... The dates of TPAC are 26-30 October 2015.
... I need to respond on the questionnaire by 30th April.

tmichel: If we plan to meet, the earlier we plan it the better choice of dates we will have
... available.

IMSC 1 Test Suite

nigel: I eventually managed to complete my promised analysis and send it round.
... Pierre, you responded on the reflector that there were some points you thought we should discuss.

pal: Thank you nigel for putting that report together. I concentrated on the ones with No
... under 'In Test suite?'
... For the first row I created a test.
... For the second one, I'm not sure how a test can be written. You'd probably offer a TTML
... file and a video file and identify that it is not a legal combination but the spec does not
... say what processor behaviour is in that case. Also this made me look at how TTML1SE
... has been modified to say how to generate media time from timing expressions, and I'm
... not even sure this is really needed. I've been casually asking DECE folks about this.
... So I suggest doing nothing right now - there might be a suggestion to remove that clause.

nigel: So we may need to signal that this statement is 'at risk'?

pal: Give me a week - that could be one outcome.
... The fourth row in the table (the third red one) is something I'd like glenn's thoughts on.
... TTML1SE has some clear text on rounding and positioning. In the DECE member submission
... this text was there to try to create consistency across implementations for aligning
... regions on the pixel grid. That was specific text about pixel position on related video.
... In the scope of DECE that's a clearly defined thing. In TTML1SE there's no mention of
... pixels - it says that if a value of extent is not available to the implementation then what
... the fallback is. I'm not sure what the interaction is between the IMSC 1 statement and that
... in TTML1SE. It needs more study. I'll ask at DECE there too. Again it might be a situation
... where we remove this text or add a note or do some other change.
... That's my report.

nigel: There are also two actions to get/generate PNGs.

pal: They're already linked in via the EBU docs.

nigel: Are they clear enough to use as test criteria?

pal: I think there is enough precision there already. The EBU spec is pretty clear in what
... the output should look like.

<mike> need to drop. cheers until next week

nigel: I think it's not clear that the images in the EBU specs are generated by exactly these
... test case TTML documents.

<scribe> ACTION: nigel Obtain PNG images for #linePadding and #multiRowAlign test cases. [recorded in http://www.w3.org/2015/03/12-tt-minutes.html#action01]

<trackbot> Created ACTION-379 - Obtain png images for #linepadding and #multirowalign test cases. [on Nigel Megitt - due 2015-03-19].

nigel: I haven't seen any other comments on the tests. Have you?

pal: No.
... I think we're good, so I'll report hopefully next week on those two items.

<inserted> dsinger joins the call

nigel: I was earlier talking about inviting the CG for an agenda item at the F2F to talk about the TTML <--> WebVTT mapping.

dsinger: Sure, that's a good idea. I'll send an email to the CG saying that the TTWG invites them to join at the F2F for that.

nigel: Thanks. We're out of agenda for the folk present so I'll adjourn now. See you next week. Thanks all. [adjourns meeting]

Summary of Action Items

[NEW] ACTION: nigel Obtain PNG images for #linePadding and #multiRowAlign test cases. [recorded in http://www.w3.org/2015/03/12-tt-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015-03-12 14:48:32 $