W3C home > Mailing lists > Public > public-html-a11y@w3.org > April 2011

Media Subteam Teleconference Minutes for Monday 11 April

From: Janina Sajka <janina@rednote.net>
Date: Mon, 11 Apr 2011 19:19:16 -0400
To: HTML Accessibility Task Force <public-html-a11y@w3.org>
Message-ID: <20110411231916.GZ2450@sonata.rednote.net>

Minutes for the HTML-A11Y Task Force Media Subteam teleconference from Monday 11 April are provided below in text. They are also available in hypertext at:


                                                           - DRAFT -


11 Apr 2011

   See also: IRC log


          Eric_Carlson, silvia, +44.154.558.aaaa, mark, Sean, Bob, janina, Frank, Judy




     * Topics
         1. Issue-152 Multitrack http://www.w3.org/WAI/PF/HTML/wiki/Media_Multitrack_Change_Proposals_Summary
     * Summary of Action Items

   People, I'm trying to get some W3C help about our telecon channel

   All, I'm trying to get our telecon channel unblocked

   <eric_carlson> what is the pass code for the Zakim bridge?

   <eric_carlson> never mind, I am in now

   I'm working on getting Zakim to let me in

Issue-152 Multitrack http://www.w3.org/WAI/PF/HTML/wiki/Media_Multitrack_Change_Proposals_Summary

   <scribe> scribe: janina

   Silvia: Believes Ian's proposal is best for moving forward; superior to proposals 2 & 3
   ... Makes all media elements
   ... Proposes to walk through the proposal
   ... Beginning look at Proposal #4 ...

   Bob, Question about exclusive track list, will that change?

   Silvia: Exclusive and multiple track list come in as attribs
   ... Ian seems to believe we can only have one video at a time, therefore exclusive

   Bob: So how to do a mosaic of multiple videos

   Silvia: So, you use separate video elements on the page and style through CSS, create a controller to get them to act
   ... Or, if they're all inband tracks, then you need IDs on each track and ref using media fragment URI

   Eric: It would work, but it makes it difficult to have generic scripts for unknown content

   Mark: If you create one, then the others are exposed and you can go forward

   Eric: For Apple, this would be a step backwards because we already support multiple tracks; would be a step back for us

   Silvia: So how does it now work?

   Eric: You display however it was authored
   ... QT has layers and each track is on video matrix that controls where it displays

   Silvia: So, if it's set up one way, PIP, for instance, how to display differently

   Eric: Either in script by disaling all video not to be displayed; then using as source

   Silvia: So we need to change from exclusive to multiple where we have default rendering, but could be changed as

   Eric: Think that's right
   ... Don't see the exclusivity as helpful

   Silvia: Think he's thinking about slides with someone addressing a meeting, but I agree I'd rather leave to a browser
   ... It would be nice to share a video that can render multiple video
   ... Suggest tracklist be replaced as multiple, not exclusive
   ... Not sure how some browsers will react, Opera ...

   Eric: WebM can only have a single video file
   ... Wonders Frank's opinion?

   Frank: We don't yet support multiple tracks
   ... We support that it's up to the UA to do this

   Eric: More than one video track in a single file is highly uncommon, but we shouldn't have it as a restriction in the

   Mark: In future, if pointing to a manifest, it will be more common; Should not require to download two videos when only
   one is to be displayed

   janina: We do have the user reqs which specify sign language video track
   ... Suggesting we check against user reqs when we think we have the engineering we like, just be be sure we got

   Silvia: I plan to do code examples

   Eric: No reason to restrict audio tracks

   Sean: Agrees

   Silvia: Our proposal always added kind label language and name ...
   ... Should we extend Ian's to also have gets for all four of these?
   ... Label is also important for creating menus

   Mark: Agree with both

   Eric: Also

   Silvia: Do we have use cases where we want looping?

   Eric: Absolutely, the same as looping on an individual file
   ... Suggest reset time to 0 when rendering reaches $DURATION

   Silvia: Good question, is it enough the group has loop state? Or each individual track?

   Sean: Feel like we're badly reinventing SMIL

   Silvia: SMIL does much more, though

   Sean: Not suggesting we adopt all of SMIL; simply saying SMIL has support for everything we're discussing

   Silvia: Think it would be difficult, but maybe worthy of the exercise?

   Sean: Best we make sure we don't do something incompatible with that model

   Silvia: Won't be used in the same context as SMIL, so why does it need to be compatible?

   Judy: We are hearing from multiple directions that alignment would be useful

   <sean> ACTION: Sean Explore reusing SMIL semantics for repeat [recorded in

   Silvia: Exploring metadata events ....

   Sean: I'm think about this a bit differently, splitting controller away from network parts
   ... That way controller need not be explicitly created, as there's always one by default

   Silvia: Is that basically just an IDL change?

   Sean: Yes.
   ... But the network states are separate

   Eric: Would you leave the existing interfaces?

   Sean: It's just rearranging them, it's just more tidy

   Eric: I find this interesting, but hesitate very much to remove anything we already have for compatability concerns

   Sean: You just say they map to the controller
   ... My problem, as currently we have things, is that the controller need always be explicitly created
   ... I'd rather it were implicit

   Eric: Actually, if you have elements in the same group, it is already implicit

   Sean: I don't like that it doesn't do that if it's not in a group

   Eric: Yes, that's an interesting approach
   ... One issue listed in Wiki we haven't yet discussed is default controls, think it's important to have UA default

   Frank: So, how would that work? A controller element positioned on the page?

   Sean: yes, perhaps a shadow DOM construct
   ... So, it might be something we could apply styling to
   ... And would be available by default

   janina: Asks if anyone is opposed?

   Eric: Not sure, but am conflicted

   Sean: So this is a thought expiriment -- there's the shadow DOM, but no representation

   Eric: Yes, a firewall between the shadow and the parent, so that js can't get to them

   Frank: We would have the same limitation in IE

   Eric: I haven't yet thought of a way to make it work

   Frank: Question about proposal 4
   ... How to get track list count --- Ah, I see it now!
   ... So, media controller maps to some amount of UI elements in background
   ... Is there a simple mapping here? A union of all the elements in the group?

   Eric: I'd say a union
   ... I personally believe we should expect all samples to be played to be loaded first

   Mark: Do we have a definitive list of track kinds that we want proposed for getkind?

   Sean: Think it's still open ended

   Mark: suggest we would want to close that at some point--a basic set?

   Sean: Yes, but extensible

   Frank: Audio, video and text?

   Mark: Thought it would tell you "texted video description," "sign language translation," etc

   Frank: What's the current thinking of displaying multiple text tracks simoultaneously

   Sean: We're not restricted on audio, so why on text?

   Frank: Yes, but not currently a spec req

   Sean: So, at minimum, text for each audio track

   general discussion of displaying multiple text tracks? spec reqs?

   Frank: There are three states, off, (no events and nothing displayed); showing (js generated and siaplayed); hidden
   (nothing displayed, but events fired)

   Eric: And if off, data not loaded

Summary of Action Items

   [NEW] ACTION: Sean Explore reusing SMIL semantics for repeat [recorded in

   [End of minutes]


Janina Sajka,	Phone:	+1.443.300.2200

Chair, Open Accessibility	janina@a11y.org	
Linux Foundation		http://a11y.org

Chair, Protocols & Formats
Web Accessibility Initiative	http://www.w3.org/wai/pf
World Wide Web Consortium (W3C)
Received on Monday, 11 April 2011 23:19:51 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:19 UTC