Meeting minutes
This meeting
Nigel: Today we have IMSC-HRM and DAPT on the agenda.
… Specifically with IMSC-HRM, some test suite changes and implementation work
… And for DAPT the main thing to cover is the languages PR
… Any other business, or points to make sure we cover within those agenda topics?
no other business
IMSC-HRM
Add note on conversion from CTA 608 signals w3c/imsc-hrm#72
github: w3c/
Nigel: There's been some debate on the wording for this editorial note, would
… be good to complete it today if we can.
… Pierre, any specific questions to be able to resolve this?
Pierre: The general disagreement is if we should make a more general statement and
… I'm not excited by that for 2 reasons.
… 1 is, there's a case I'm comfortable talking about, which is the 608 line 21 example at 59.94fps
… I've seen a lot of streams like that.
… 2nd is, this is just a note and we should try to avoid designing a system in the note.
… The general question of conversion to IMSC is worth its own document.
… I'd like to keep it to just a specific example highlighting the challenges.
… That's my preference.
… I'm happy with the group decision ultimately.
… I can try to make another pass explaining that more clearly.
Nigel: Worth noting that this note is right after an "e.g." pointing to 608 in the text immediately above it.
… Can we just remove the phrase about 59.94 Hz? The issue would remain at other frame rates
… such as 29.97.
Pierre: If that resolves both your and Cyril's concerns I'm happy to do it.
Cyril: I don't think adding the sample rate helps, and I don't want to add a note that is not totally accurate.
Nigel: Maybe say "at the video field rate" to clarify that this isn't happening once every few seconds,
… but many many times per second.
Pierre: There's a general misunderstanding - 608 is sampled at a field rate, but sometimes the packets
… are combined for other transport mechanisms. But 608 has no option other than sampling at the field rate,
… at least classic 608, like line 21.
Nigel: Is this the change that we can accept then?
Pierre: I'll put this on screen. [shares screen]
… Now says "... for every CEA-608 packet, which are sampled at the video field rate. ..."
Cyril: I like that we removed the 59.94. Field rate is related to interlace. Should we say field or frame?
… I'm fine with this. A field is a frame when there's no interlace.
… And maybe we will have to change it again if we have Teletext issues.
Nigel: Teletext doesn't have this "build up a buffer character by character" approach, so it doesn't arise.
Pierre: That's modified, I'll ask for a re-review.
SUMMARY: Live edits made, needs re-review
IMSC-HRM Tests and call for implementations
Nigel: Exciting news: I completed a second validation implementation of IMSC HRM,
… based on the EBU-TT Live Interoperability Toolkit, as a starting model.
… That means it can process all the files in the test suite that can conform to EBU-TT-D as well as IMSC Text.
… As part of this exercise I updated, and Pierre reviewed, the tests to be EBU-TT-D conformant.
… We had some other fun along the way, identifying some tests that weren't quite right,
… and found bugs.
Pierre: The bugs we found were in the implementations and we are fixing them right now.
Nigel: We are also in the process of adding tests for region presentation.
… That came out of a fun discussion about the handling of tts:showBackground on region elements
… that set a backgroundColor that is not entirely transparent.
… There's also an issue about one of the test pairs, dur005, which we're in the process of resolving.
… So it's been a useful exercise.
… Next steps: finish sorting out the tests, and update the implementation report to add the additional implementation.
Pierre: Have you checked the implementation report?
Nigel: Not yet.
Atsushi: I have not either.
… The main question is we may need to list the mapping between HRM features and the test suite.
Pierre: In each of the test files there is a description.
Atsushi: In any case we may need a script to list those things easily.
Pierre: Why?
Atsushi: There are a bunch of test files.
Pierre: They're organised into two directories, which are sufficient for automatic testing.
Atsushi: In any case I need to set some time aside to check it.
Pierre: Ok.
Nigel: Edits needed at the moment are: adding the new tests, and the 2nd validating implementation
Nigel: If at some stage we need to demonstrate the truth of the emails to the Chair listed under
… Content producing implementations, I'm sure we can find a way to do that.
… Anything else on IMSC-HRM?
Pierre: Not for me, thanks for everyone's help and input on this. We're getting super close
… and hopefully can close this very early next year if not this year.
DAPT
Rework audio description original and translation languages w3c/dapt#179
github: w3c/
Nigel: This has been open since July 31 and has been through an amazing number of iterations,
… each one making it better than before.
… I think we're probably about done with it, and would like to know if we can merge it today
… or if anyone wants more review time.
… Since the last review the main change is to model the daptm:langSrc attribute better and to
… define the semantics and constraints better.
… Those changes are in line with the issue and pull request discussion, I think,
… but I am aware that Andreas had some reservations.
… I did try to address those.
Andreas: Thanks first for all your work on that pull request. It was an amazing amount of editing
… work to resolve the different comments. I think it gets better and better.
… As I signalled, maybe offline, I'm happy with the way forward as described in the comments.
… It's not just one line, you have to check the whole document.
… I can review it in the next couple of weeks but if you want to merge it now that's okay.
Nigel: I could merge it and then if you find issues with it then open new issues.
Andreas: That would be okay.
Cyril: I concur with Andreas. I would also like to give it another review, and I can do it tomorrow.
Nigel: OK then I will leave it open and then react based on the result of your review.
Cyril: After my review I will try to implement the solution in my prototype tool to see if there is any
… problem in our source content. I can do the review first. If I see any problem I will open a new issue.
Nigel: Great, thank you.
SUMMARY: @cconcolato to review, @nigelmegitt to respond accordingly. @andreastai to read through when time allows.
Next steps and Wide Reviews
Cyril: Do we have all the responses?
Nigel: Not quite.
Cyril: We have TAG and i18n.
Nigel: We haven't had the formal response from APA yet.
… We've been told not to expect a response on Security unless they identified a problem.
Atsushi: If we can resolve all of the issues raised against DAPT then we can state that
… all HR is finished.
Nigel: Please could you check on the status of those, Atsushi?
Atsushi: Of course, I will do it during winter holiday season.
Nigel: My other question is about the feedback from DVW - do we have a way to complete
… the conversation with them after the detailed discussion responses?
Cyril: I'm hoping to get a response from them, yes.
Nigel: In terms of tasks, we have some issues that need decisions to be made,
… at some point we need test suites (not formally until after CR),
… and there are a couple of PRs open from June that need something doing with them.
Cyril: Those PRs are still on my list!
Meeting close
Nigel: We have one more meeting this year, on 21st December.
… We've completed our agenda today. Thank you everyone. [adjourns meeting]