Timed Text Working Group Teleconference

<scribe> scribe: nigel

This meeting

nigel: We have a new handy Board: http://www.w3.org/AudioVideo/TT/board/
... We now have a list of Pull Requests which are a nice thing for us to review.
... We should think about how long we leave them open for review - I think we said 2 weeks.

pal: We also need to consider PRs geting stale and avoid them needing to be rebased as
... other code changes around them, so leaving them open for a long time generates more work.
... So we should go through the PRs. Unless someone objects to merging the PRs maybe
... the editor should automatically merge after 2 weeks. Then further discussions can
... be handled with new filed issues.

nigel: I agree with that.
... Also you can clearly identify what went into each change in case you want to back it out later.
... So the proposal for today's meeting is to review the open PRs. Any other business?

group: No AOB.

Action Items

action-448?

<trackbot> action-448 -- Nigel Megitt to Put timing diagrams somewhere they can be reviews -- due 2015-11-06 -- PENDINGREVIEW

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/448

nigel: I did this as a PR which we'll be able to look at later in this meeting.

close action-448

<trackbot> Closed action-448.

action-449?

<trackbot> action-449 -- Philippe Le Hégaret to Create board for ttwg similar to that for webperf -- due 2015-11-19 -- PENDINGREVIEW

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/449

nigel: I already mentioned the board, which plh made, so this is done.

close action-449

<trackbot> Closed action-449.

IMSC

pal: Let's start with PR #88: https://github.com/w3c/imsc/pull/88

nigel: By the way, a quick way to see all the PRs is to look at: https://github.com/w3c/imsc/pulse
... How can we see the HTML natively, with the changes applied?

<pal> http://rawgit.com/w3c/imsc/issue-68/spec/ttml-ww-profiles.html

pal: The recipe is, when you click "View document" then click "Raw" and then replace raw.githubusercontent.com with rawgit.com
... The changes address the issue on the Note in §9.5. I removed the wording around
... complex scripts and simple scripts. My research showed that it is sometimes seen as
... contentious and also incorporates other areas of complexity.
... I also listed the scripts directly rather than referencing the appendix, and removed appendix H.
... Appendix H was only going to be referenced once, so there was no advantage in doing that.

nigel: I can't see Appendix 24 at the Unicode link.

pal: It's Annex 24.

<pal> http://www.unicode.org/reports/tr24/tr24-24.html

pal: I'll change the wording to point to the right place.

http://www.unicode.org/Public/UCD/latest/ucd/PropertyValueAliases.txt

nigel: Don't you want to include the "common" script in the set with GCpy=12?

pal: I haven't because I haven't gone through everything in common to check it yet.
... There's another page where you can see the list of scripts
... I'm looking at it now and there are a lot of symbols but they're not all in the recommended characters
... It's hard to tell if some of those would have a one to many mapping.
... I'd like to take an action item to dig deeper into that. Also recall that Glenn still
... owes us a list of simple vs complex scripts, so we can deal with that together.

<scribe> ACTION: pal Check character to glyph mappings for common script value [recorded in http://www.w3.org/2015/11/19-tt-minutes.html#action01]

<trackbot> Created ACTION-450 - Check character to glyph mappings for common script value [on Pierre-Anthony Lemieux - due 2015-11-26].

pal: Based on this, I recommend we merge this and then action further changes if we identify them.

nigel: Seems okay to me. Any other views?

Frans: okay by me

pal: I'll make that change right now and see how it works...

nigel: I can confirm that addresses all my issues in issue #68

pal: That's merged now.
... Next I propose we look at the aspectRatio one.
... It's issue #84 and PR #89

https://github.com/w3c/imsc/issues/84

https://github.com/w3c/imsc/pull/89

pal: I want to make sure that "Related Video Object frame" literally means each image frame of the related video object,
... regardless of how it is displayed. Because the goal is to capture authorial intent with respect to the positioning relative to elements on the video frame.

nigel: So the goal was not to address the case where not all of the video is shown?

pal: Exactly. That's a similar goal as with synchronisation: capture authorial intent and then
... let applications handle other scenarios as nicely as they can.
... It's complex as your presentation shows nicely.
... I can imagine lots of possibilities, with applications generating different versions.
... [lists lots of possibilities!]

nigel: We could signal a safe area within the root container, outside of which there's no content.

pal: You can do that by, say, having aspectRatio="4 3".

nigel: That doesn't work on 16:9 video when they're displayed on a 21:9 display - that can
... clip the container.

group: [discussion about what devices should do if the video frame is not fully shown]

pal: We could add an example for when the display device is not showing all of the
... related video object frame, to show a reasonable behaviour in that case, vs the one specified.

nigel: Okay

pal: I'll have a go at drafting one that we can then review.
... The drawback is you lose positioning, but that's a choice that maybe the system in
... conjunction with the user might correctly take.
... Issue #81 is more straightforward. Glenn pointed out that if somebody uses cellResolution on linePadding,
... the author should be encouraged to set ttp:cellResolution on tt. I looked at EBU-TT-D
... and it says "should" set cellResolution. The proposal here is to follow that in IMSC 1.
... I had an action item to go through the features in IMSC 1 and work out feature
... dependencies, so I've done that.

nigel: I saw that and was confused that lots of features were being prohibited.

pal: What I did was, for image profile, where parent features were prohibited, I listed the
... dependent features as being prohibited also.

nigel: That sounds great - we'll do the 2 week review, and if no adverse comments, the Editor can go ahead and Merge the PR.

pal: That will get us to Dec 3, so hopefully then we'll be able to get another CR published.

nigel: Yes, that would be good.

pal: I've started the list of substantive changes and will update further based on these closed issues.
... There have been a lot of cosmetic changes but I only found 3 substantive changes.
... For example the changes in the Recommended Characters don't affect conformance.
... Then there's the change to the HRM, and ttp:profile element being prohibited, the addition of Liberation font to the reference fonts.

nigel: We're out of time now.

pal: Shall we do 2 hours next week?

nigel: Okay, no problem.

TTML2

nigel: I just wanted quickly to mention the Timing work at https://github.com/w3c/ttml2/pull/143

pal: It looks good - it's a great help.

nigel: Great. The main thing is that the content is right.

pal: Yes, I think it matches my intuitive understanding. The second diagram is one where
... an external context starts some way through the media timeline, which shrinks the
... Root Temporal Extent. So the Root Temporal Extent is shrunk by the external context.

nigel: Yes, that's right.

pal: That makes it unambiguous so people have a clear thing to disagree with!

nigel: Great. It would be good for the record if you could add your review comment to the PR.

pal: Done!

nigel: Thanks everyone, we'll meet for 2 hours next week as requested. [adjourns meeting]