W3C

Timed Text Working Group Teleconference

21 May 2026

Attendees

Present
Andreas, Atsushi, Chris_Needham, Cyril, Dana, Gary, Nigel, Pierre
Regrets
-
Chair
Gary, Nigel
Scribe
nigel, cpn, cpn3

Meeting minutes

This meeting

Nigel: One DAPT issue, IMSC 1.3 final details and Rec published, TTML Media Type and Profile Registry, WebVTT license
… Impact of AI technologies on our mission, and TPAC planning
… Anything else?

(Nothing)

IMSC 1.3

Latest publication REC 2026-05-21

Nigel: We published Rec today. Congratulations and thanks everybody for your work on it

News item

Pierre: Thanks everybody for helping push it along
… Could discuss IMSC Image Profile today?

unversioned IMSC link

Nigel: This link takes you to IMSC 1.3, but that doesn't refer you to 1.2 for the Image Profile. Propose adding to the Abstract to point people there

Pierre: I prefer either to change nothing - as IMSC Image Profile is a specialist tool, and our survey showed very few people use it, and we don't really want people adopting it going forward. So it's ok not to link to it
… Also, not a great idea to bake a link to 1.2 in the abstract in case there needs to be a new version of Image Profile in future
… May be better to acknowledge there are two profiles of IMSC, and have two canonical links to track each of the profiles

Nigel: The canonical link we have is derived from the shortname

Pierre: I realise the second option is ambitious, so I actually prefer to do nothing

Nigel: Any other views?

Andreas: Unless we hear complaints, do nothing

Cyril: IMSC versions are usually supersets of previous versions, and the Image Profile in 1.3 so it's not a superset

Pierre: Each version of a profile is a superset of the profile

Cyril: Otherwise my suggestion would have been not to mention the Image Profile, but make a blanket statement to refer to the previous version

Pierre: We have that statement, just not in the Abstract

Nigel: Do nothing seems to have enough consensus for now
… Atsushi, there's some complexity with errata and release tags. Anything else we need to do?

Pull request to revert to ED from Rec

Atsushi: We need to get the github.io version to ED per request from the Guide. Streamlined publication for IMSC 1.3 has Rec as final state. I commented out the publication status from the GitHub action

Nigel: So if we start working on a new version, we get a new token

Atsushi: Yes

Nigel: So the action is to create the release tag based on what's in the repo now. Pierre, do you want to do that?

Atsushi: You can create a tag from the current HEAD

Nigel: Yes, especially since there's no change to the Abstract

Pierre: Why did the CI fail?

Atsushi: It tried to update /TR

Pierre: I'll compare the HEAD of the main branch with the actual Rec, check there's no differences, then create the tag

Nigel: Sounds good. Thank you

TTML Profile Registry entry for IMSC 1.3

PR, needs an approve

Nigel: There's a PR for this issue. It just needs approving, now IMSC 1.3 is at Rec

Pierre: I'll approve that

Nigel: Thank you

Atsushi: The publication token is tied to the shortname, so if you change that you need a new token
… If the WG wants to update the Abstract, it should update in place of the Recommendation, so there's a checklist. There'd be no ReSpec source for that

Nigel: For now, we won't change the Abstract
… Anything else to cover on IMSC 1.3?

(Nothing)

Nigel: Thanks again Pierre for all your effort

TTML Media Type Definition and Profile Registry

Nigel: I think we don't have auto-publication for this?

Atsushi: We can configure it for any publication

Nigel: That would be good

Atsushi: Once we've resolved to use streamlined publication for the Registry. In the past we decided one by one, so just need a resolution

PROPOSAL: Auto-publish the TTML Profile Registry as a Note on pull request merge, based on the usual TTWG Decision policy applying to pull requests

Chris: Sounds good to me

<atsushi> +1

<atai> +1

RESOLUTION: Auto-publish the TTML Profile Registry as a Note on pull request merge, based on the usual TTWG Decision policy applying to pull requests

Nigel: There are some PRs open. Please review

DAPT

Profile identifier and TTML registry w3c/dapt#248

github: w3c/dapt#248

Nigel: When I raised a PR to add the DAPT profile identifier to the profile registry, I realised DAPT is different as it has two profiles shown, which are one for the content profile, and one for the processor profile
… I didn't find where this originated. The profile registry should have the processor profile
… And the DAPT documents ttp:contentProfile attributes will have the content profile
… That might be confusing for people. Maybe we change it so there's one profile designator
… Processor profiles have mime type parameters, whereas content profiles says what profile the content conforms to

Cyril: I looked at the DAPT spec, the only use we have for Processor Profile, besides its formal definition, is in a Note that says that normally you shouldn't use it, unless you have particular processor requirements you need to express
… We have no tests exercising that
… It's a bit annoying to have to issue a new CR for this, but...

Nigel: Formally, having different designators makes sense, but then balance that against potential for confusion. I think people will expect them to be the same

Cyril: There's a table in Annex F, explaining what features are required, optional, and how they're treated in the processor and content profiles
… If we remove the processor profile completely, how to deal with that?

Nigel: Not suggesting removing the profile, just use the same designator
… That's defined in 5.6 at the moment
… All we need to do is remove the word "content" and "processor" from each, and everything else stays the same

Cyril: Not sure why we introduced the processor profile concept, may be about audio that's supported by some tools
… But I remember the Note in 5.6.4 about not using the processor profile attribute
… I need to think about this, but my inclination is to remove mention of the processor profiles

Nigel: I want to make the smallest change possible, which is changing the designator, so both profiles use the same designator

Nigel: I can open a PR that makes this change for us to consider and make a decision based on that?

Cyril: OK

Andreas: How can we define two different profiles with a single designator?

Nigel: Both the processor profile and content profile each have a profile designator, but we make them the same value, instead of one being /content and one being /processor
… There are no content profiles in the profile registry now. Even TTML2 defines content profiles and processor profiles, with the same designators

Cyril: IMSC only talks about a profile designator

Andreas: You can still distinguish them by setting the type attribute?

Nigel: Yes, exactly

SUMMARY: @nigelmegitt to open pull request making the content and processor profile designators the same value

WebVTT

License

Nigel: Historically WebVTT had a different license to other TTWG Recs. I think we decided to maintain that state
… But at some point it was changed?

Atsushi: I checked the record, and decision to use Software and Document license for WebVTT in chartering in 2016
… From that point, WebVTT shall be published using Software and Document license
… From 2019, it used the Document license, in the publication, so I was confused

Gary: The Software and Document license is a little more open than the Document license and
… that's what we used in the CG. It makes sense to keep it.

Nigel: I don't know why it was changed. I think we discussed allowing TTWG to choose license on a spec by spec basis

Charter section on Licensing

Gary: I assume it was a mistake when we made the snapshot

Nigel: So we can revert it to the Software and Document license. That needs updating in the Bikeshed boilerplate PR?

Atsushi: Already done

Nigel: Anything else on WebVTT?

(Nothing)

TPAC Planning

Andreas: I'm interested to know if we have enough topics to discuss at TPAC, and whether people are attending, to help decide whether to travel

Nigel: I assume we would want to meet. Good to ask about topics. We're approaching completion on DAPT, might be done by then, there might be implementation things to discuss
… As Chair, I want to get TTML2 2nd Edition published. That's another topic
… WebVTT topics?

Gary: Maybe, but I may not be there in person

Nigel: Any interop topics to discuss?

Dana: I'd like to discuss that at TPAC. I've been working on the VTT interop investigation. I'm planning starting a monthly meeting soon, to keep people posted on what's happening

Gary: TPAC2026 is the last week of October in Dublin

Impact of AI technologies on Timed Text Working Group's mission

Nigel: We started discussing this in our last meeting
… Last week there was the MPTS show in London. A few stands there related to AI. A company from Germany, Phont, contacted me, putting more dimensions into captions, such as emotion
… They seemed to be amenable to the idea of standardisation
… I asked them about cost of authoring, and they said it took a long time to do manually, but their product uses AI to automate
… So although AI might not affect our mission directly, there's extra information we might want to put in caption files that's AI generated. Also AI processing to understand what's in media, so having a profile for that, DAPT might be good for that.

Cyril: There's another contact we had, Captions with Intention. Similar. Interesting to talk about that. I agree that authoring is a key part

Gary: I was thinking of the same thing

Andreas: Other organisations run study missions to formulate some use cases first. Then see if there's something to be done

Nigel: Chris, would this be a good idea for an MEIG study mission?

Chris: MEIG has been asked the same question, I'm expecting to have the same conversation there.
… Possibly on the agenda for the MEIG meeting the week after next.
… That would be a broader conversation, not just about captioning but media in general.
… I'm open to the idea of using MEIG as a place to cover that stuff.

Nigel: I'm thinking MEIG could note the direction of travel and identify any requirements.
… Those requirements could then come back into TTWG if appropriate.

Chris: It would still be the same people doing the work, but in a different context.

Nigel: It would be easier for non-members of W3C to participate in the MEIG work.

Nigel: We could fold other organisations into the work, easier to do there

Chris: And helps TTWG stay focused on its deliverables, and MEIG is more exploratory

Nigel: That's a topic for TPAC then, either in MEIG or TTWG

Next meeting

Nigel: 4th June, w3c/ttwg#340

Meeting close

Nigel: Thanks everyone, we're at time. [adjourns meeting]

Summary of resolutions

  1. Auto-publish the TTML Profile Registry as a Note on pull request merge, based on the usual TTWG Decision policy applying to pull requests
Minutes manually created (not a transcript), formatted by scribe.perl version 248 (Mon Oct 27 20:04:16 2025 UTC).