W3C

- DRAFT -

Immersive Web CG Call

21 May 2019

Attendees

Present
trevorfsmith, cwilso, Leonard, atai2, bertf, Vlad
Regrets
Chair
Trevor Smith
Scribe
Chris Wilson

Contents


<joshmarinacci> @trevorfsmith good morning

<scribe> scribe: Chris Wilson

<scribe> scribenick: cwilso

Trevor: main topic will be https://github.com/immersive-web/proposals/issues/39 and https://github.com/immersive-web/proposals/issues/40, if you haven't seen them take a quick look.

*general envy of Ada's mask-face from new delivery*

Rik: Google Hangouts on Magic Leap works, and shows your camera feed

trevor: welcome all

Subtitles in 360 videos: Issues #39 and #40

atai: two issues have been opened. First issue is how to deliver subtitles in 360 videos, use cases are described in issues. Problems to solve: positioning subtitles in 360 video, identifying speakers.
... requirements have also been brought up in future of TTML (in Timed Text WG)
... original we wanted to solve this this year, but may not happen.
... first maybe we should discuss the requirements and use cases, but I'd like to understand where the appropriate venue is to try to solve this problem space.

trevor: one thing that has been discussed in the issue is the level of abstraction.
... we're using IRC queue, so please Q+ yourself

<bajones> +present

trevor: One ? - if a web developer is using WebXR, WebXR is kind of a black box for rendering. Doesn't really go into how things are rendering, there's no obvious place to manipulate the scene or insert subtitles, e.g.

klaus: this could be done at different levels - at app level, which doesn't help developers much, or could be purely at the user agent level, but as trevor mentioned is not really the way WebXR works today; or maybe it's a separate layer?
... I'm not sure there's a "right" solution to this. It's not really clear what the proposal is for what one would want to do.

<Zakim> klausw, you wanted to say is this app-level, UA level, or a mix?

vlad: first: regarding XR and advanced applications. in typical subtitle scenario, user may not hear audio at all.
... in 3D audio, you may get this cue, but the subtitle may need to give this directional cue also.
... you'll need to provide this information somehow.
... in general XR applications, text may be labels or the like, not just subtitles.
... i.e., it might be better to think of #40 as "labeling" not "subtitling"

<bajones> +q

atai: thanks for feedback. Klaus, I think what you are saying is that it's unclear where to solve the requirements. I think there's not a single ecosystem to solve it. The requirement from the user is "I want subtitles" - so we need to figure out which ecosystem or ecosystems can solve it.
... size and position of subtitles are also important - like in 2D - so TTML is attractive. But where do you signal there are subtitles at all?
... it's not clear if this is scope of your group?

trevor: I think it's clearly in the purview of the WG... It would be interesting to hear from Nell.
... (about how they've considered this problem)

brandon: fascinating example of how we can augment data on the web for immersive. A high-quality user-space library could step in and take over some of this. This would still be a good home for the eventual need.
... the big question for me is are there aspects of the accessibility angle that would best interact with the UA level?
... there are a wide variety of accessibility systems out there. The most compelling reason for a browser standard may be that avenue to assistive tools.

<johnpallett> :q

<trevorfsmith> heh

cwilso: It's clear that TTML should be expanded to capture the 360 positioning of subtitles. But the challenge is we don't have a 360 video primitive in WebXR
... in the short term, if TTML can be added to for this case, a user library can take over this space for the moment when it does the video stretching, etc.
... once we add a primitive it should be able to do this, but we need a general 3d overlay layer.

<Vlad> To offer some clarity on things I mentioned: 3D audio provides directional cues so we need to consider them, and the 3D audio source itself can be a moving object. Coordinate based positioning of subtitles could be sufficient in certain cases, but we also need to accommodate object-based attachment points (also true for labeling use cases).

atai: the scope of this gets pretty big quickly, so not sure where to start. 360 video seemed the most obvious to start.
... actually about rendering of text, though, of course.
... not limited to 360 video environment. Is it better to look at something else?

johnpallett: there are a number of things here - the transmission of subtitles beyond 2D subtitles and broadcast transmission standards. There is prior art in broadcast standards for using web standards to enhance broadcast standards.
... probably best to start with "can this use case be enabled in webxr today" - exploring what can the browser do vs what can the standards support?
... broadcast standards might benefit from the coordinate system.
... it would help to know the lay of hte land from the transmission perspective.

atai: specific delivery over the web, this is probably the best place to start.

johnpallett: it's worth exploring what the browser can do today, separate from the conversation about how can webxr be used in other standards (E.g. the coordinate system in transmission standards)(
... e.g. atai mentioned some MPEG work.

trevor: maybe Timed Text could help get some context where this fits.

atai: my question is really, some parts of this need more work on your side.
... extending TTML ...
... the question is, on the list of topics you currently work on, is this going to see some attention?
... is this something you're going to work on in the near future?

trevor: in general, yes, but most people are head-down on trying to ship first rev of WebXR.
... until we have a base on which to ship something else, some of these conversations are general.
... we can use the proposals issue to track.

cwilso: Is this a work item in Timed Text WG? That's probably the top thing to solve. Then we need to do overlays in WebXR (on the plate), and a JS lib can be built
... prior to generating a primitive.

trevor: will follow up with Andreas and the TTWG.

Face-to-face session topics

trevor: if you are a repo manager, please use the form: https://forms.gle/yeXkH4cfrVAX4pt16 to suggest topics in the next day or so, we're about to make the schedule.

<Leonard> This one? https://github.com/immersive-web/proposals/issues/50

New DOM overlay feature incubation repo (which I am going to create now)

trevor: John will be the repo lead; please take a look.
... any other topics?
... hearing no further requests... Andreas and I will follow up

<bertf> thanks!

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/05/21 17:50:48 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.154  of Date: 2018/09/25 16:35:56  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Present: trevorfsmith cwilso Leonard atai2 bertf Vlad
Found Scribe: Chris Wilson
Found ScribeNick: cwilso

WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

People with action items: 

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)


[End of scribe.perl diagnostic output]