HTML-A11Y Media Subteam Minutes for 12 January

Minutes from today's HTML-A11Y Task Force Media Subteam teleconference
are provided below in text. They are available as html at:


NOTE: New Actions assigned on Shawn and Janina.

                                                           - DRAFT -

                                                       HTML-A11Y telecon

12 Jan 2011

   See also: IRC log


          Janina, John_Foliot, Judy, Geoff_Freed, Sean_Hayes, Eric_Carlson




     * Topics
         1. Identify Scribe
         2. Actions Review
         3. Continued Itemized Bugs Review -- See itemization below
     * Summary of Action Items

   <JF> JF waves

   Oh no, Silvia! Sorry to hear this!

   <scribe> agenda: this

   <scribe> scribe: janina

Identify Scribe

   scribenic: janina

Actions Review

   Silvia, we're wondering if we can reassign action-8 to you?

   <silvia> do we have a link to action-8?

   <JF> Silvia, are you able/willing to take on action item 88? Review Media Fragment URI 1.0

   <silvia> ah! well, I am part of the team that wrote the Media Fragment URI 1.0 spec, so it would be kinda useless for
   me to review it

   <silvia> you'll notice I am one of the authors

   OK, Silvia. That makes sense. We'll kick this action back to PF (where it came from).

   Next, Bug 1195 -- Qualify success criteria

   We'll follow up on this at the main TF call tomorrow.

Continued Itemized Bugs Review -- See itemization below

   Bug 8657 -- Eric?

   Allow UA to reload fallback content ...

   Q: Should a video start playing if additional "fallback" content not fully loaded?

   We thought not last week.

   EC: Sean is correct. External track considered "ready" even if not loading.
   ... So, what should the behavior be if caption track loads more lowly than main resource?
   ... Not sure that was the initial intent of this bug, though.

   SH: If text track is hidden, can still be considered "ready." Also concerning.
   ... Not sure "ready" should be allowed if something fails to load.

   <silvia> I think we are confusing the ready state of the media with the ready state of the caption track

   <silvia> there are two independent ready states

   EC: There's "have metadata" to indicate when enough of a file is available to know what it is, and enough to render a
   single frame

   SH: Need some kind of buffering mechanism

   EC: There are two states beyond the single-frame, finally a state that goes to eof
   ... State can go up and down.
   ... Agreeing with Silvia, thought Sean asking about how this fits with main resource.
   ... We may be getting caption data, but not know ...

   <silvia> I assume if the main resource has an error state, there will be no attempt to load the text tracks

   SH: But captions for an hour's movie is still pretty small data

   <silvia> if the main resource loads, then an attempt to load the text tracks will happen, too

   <silvia> this can succeed (i.e. LOADING or LOADED state) or fail (ERROR state)

   Silvia, I think the concern is that the video not play if the attempt to load the text track fails, or is delayed

   <silvia> if it succeeds, the tracks are available for display - if not, they are not

   SH: Think we can close this bug as it's about fallback, but we're discussing a different bug.

   <silvia> should a broken text track really stop a perfectly fine video from playing?

   Silvia, are you saying that the movie doesn't play if the attempt to load text track has not yet finished?

   <silvia> if a text track is to be loaded and thus the browser goes from NONE to LOADING state, it probably makes sense
   for the playback to wait a bit

   GF: If user said, want captions, need to indicate they didn't load, proceed anyway?

   <silvia> if after an acceptable waiting time the captions haven't loaded, the caption track should probably transition
   to ERROR state and an indication being shown on screen, but the video start playback

   SH: Perhaps ua not required to play video if indicated resource unavailable

   <silvia> then it is up to the user to decide if they want to accept the video without captions or try a different track

   <silvia> I'm just considering that a video that hasn't got any captions available will also play and not be stopped by
   an accessibility requirement of a user

   <silvia> so if a caption load fails, this is the same fallback situation

   GF: Ideal to let the user know that error has made my default requirements unavailable this time.

   EC: Yes, but that's not html spec, it's ua

   <silvia> a message such as "english caption track failed to load" would certainly be helpful

   SH: It's 'should' so there's wiggle room for the ua

   <silvia> similarly a message to the screen reader about "english text descriptions failed to load" would be good, too

   Silvia, won't be a screen reader with captions, though. Probably only the browser

   JF: Close 8657, not what we're discussing here

   <silvia> janina: <track> also supports text descriptions

   Yes, in that case could be screen reader

   JF: Hand this to UAWG?

   RESOLUTION: We will suggest UAWG describes proper behavior when alternate content representation not loaded and an
   error is generated.

   <silvia> it could be useful to add a note to the spec that UAs should indicate (visually or aurally) to the user if a
   text track that was deemed to load failed

   Bug 8673 -- not yet ready ...

   <silvia> it's clearly a UA thing though, so I am happy with the resolution

   <Judy> sean: we need to capture best practices on captions

   <JF> Group decision to report back - concur with Editor on bug #9774 - not a valid issue - pass

   <JF> JF - asking Sean if we can capture potential authoring issue as he identified?

   scribenic: janina

   <JF> Sean: what happens if caption is position such that the captions are position outside of a readable region -
   possible that an author can do so

   SH: It might be possible to set 0 picsel size, can't prevent authors from doing something stupid

   <JF> not something that can be mandated in a spec, but it is important author guidance

   SH: wcag doesn't explain captions, though, so best practices not provided that way
   ... perhaps existing knwledge should be captured as wcag techniques

   <JF> suspicion that this should be formulated as a WCAG success technique

   GF: No single source of agreed caption guidelines, a neutral place like wcag a good spot
   ... Happy to help with this

   JB: This is important info, esp that people are approaching in parallel, but not coordinated in a neutral repository
   ... Would like to se this go forward

   GF: Should also consider descriptions

   <Judy> JB: absolutely

   JF: Suggesting that SH and GF create a statement to be forwarded to WCAG

   action sean to create a statement with geoff to forward need for caption and description techniques for wcag

   <trackbot> Created ACTION-98 - Create a statement with geoff to forward need for caption and description techniques for
   wcag [on Sean Hayes - due 2011-01-19].

   Bug 10693

   Silvia, can you join the call for a few minutes about this bug?

   <silvia> am in Germany and not near a phone, sorry

   What, no phones in Germany!!!

   Can you look at Bug 10693 and say whether you can follow up on it?

   Bug 10841

   <silvia> I will have to follow up indeed

   John filed this just ahead of last call cut off ...

   What to do when a descriptive track is longer than the primary resource time?

   GF: How to know is first question
   ... What to do should be pin preference; Stop movie automatically?
   ... If recorded audio, not too hard to tell
   ... But, what if text?

   SH: Can think of algo with audio
   ... Also ok in SAPI for TTS

   GF: agree on mechanisms

   SH: I just pause every description, very simple

   JF: Does this require discrete control?
   ... A pause button for main resource, another for description?

   SH: Not necessary
   ... Pause the main resource when playing the description; cycling mechanism
   ... Thought it was a cognitive disability question, someone wants a lot more time

   <gfreed> i have to run.

   Thanks, Geoff!

   <gfreed> will read the minutes for the rest of this thread.

   <silvia> we actually discussed this during a meeting at IBM Japan and we thought that a "extend timeline" marker would
   be helpful - then the author can signal to the browser to wait for the screen reader to finish

   <silvia> some more details at

   <silvia> I do agree that we haven't currently got a mechanism to deal with extended text descriptions

   <silvia> I believe experiments are necessary

   SH: Works because doesn't rely on screen reader. TTS is generated directly fot eh rendering

   JF: Can keep this open, as needing testing and more discussion. We will also note that non screen reader mechanisms
   need to be considered.

   EC: Not sure we can solve by the time the spec is done

   SH: Agree
   ... Main thing is to make sure there are enough hooks to make sure we can do the right thing

   EC: Agree

   SH: I already have demos that work with HTML5 browser

   EC: Silvia has also

   <silvia> Sean: are you doing extended descriptions? I am only doing synchronized descriptions

   <Sean> yes.

   <silvia> (apologies if I have seen it before - I don't remember)

   DRAFT RESOLUTION: We believe this is about UA behaviior, and the spec should only provide sufficient resources for UA
   to create compelling solutions.

   We believe those sufficient hooks already exist in the spec.

   Silvia: Are you OK?

   <silvia> with this resolution?


   <silvia> I am not 100% sure we already have everything we need - I believe some extra markup may be necessary in

   Silvia, Can you elaborate further on next week's call?

   <silvia> if Sean has an example that we can all look at and that doesn't need anything extra, that would be fine, too

   Silvia, Shawn will follow up with you.

   <silvia> thanks

   Bug 10842

   <JF> Bug 10842 - Support the isolation of speech from other background sounds in AV media

   The "Clear Audio" case

   It's about production, but if packaged in with main media there's currently no api to determine, expose, select, etc.

   Agreement that we need to solve this when we solve alternate lang, recorded audio description, sign lang translation,

   Silvia, isn't there another bug that better represents this problem?

   <silvia> yes, the multitrack API

   <JF> bug number?

   <silvia> it's not "better", but just a different aspect of the same bug

   <silvia> 9452

   So, we close 10842 in preference for 9452?

   SH: Link this one to 9452

   <silvia> I wanted to spend a day putting a proposal together which we can further discuss - haven't had the time yuet

   <silvia> we had some good email discussions

   Close this one, and add Clear Audio to 9452

   <silvia> I would recommend keeping both around

   <silvia> even if we have multitrack, we still need an API to separately control the volume of the tracks

   <silvia> is sort-of a next step issue

   <silvia> we can put a message in 10842 though that it relates to 9452

   <silvia> and maybe even waits for resolution for 9452

   Yes, but 10842 implies that creating clear audio is a ua technology, and that's something we need to dispose of. It

   <silvia> it needs more info, namely that the separate control of volume of different tracks does not require audio
   analysis, but only separate volume control through an API

   <silvia> if you want, I can add a sentence

   Silvia, we need to be closing bugs. We're proposing to capture the need for discovery, plus the need for independent
   control of multiple tracks with 9452.

   Independent control of multiple tracks is also in our reqs.

   <scribe> ACTION: Janina to annotate 9452 with clear audio discovery and selection, as well as independent control of
   multiple playback tracks [recorded in]

   <trackbot> Created ACTION-99 - Annotate 9452 with clear audio discovery and selection, as well as independent control
   of multiple playback tracks [on Janina Sajka - due 2011-01-19].

   <JF> new audio WG - see:

   <JF> Decsion: once janina completes Action-99, close bug 10842

   Discussion of whether the proposed Audio API WG can take our issues for controls. Feeling of most is that it isn't.

   It will be very late, and is not currently scoped for controls.

   JF: Remaining is 4 bugs for next week?

   EC: A question for the Chairs as to when we should resolve these?

Summary of Action Items

   [NEW] ACTION: Janina to annotate 9452 with clear audio discovery and selection, as well as independent control of
   multiple playback tracks [recorded in]

   [End of minutes]

Found Scribe: janina
People with action items: janina


Janina Sajka,	Phone:	+1.443.300.2200

Chair, Open Accessibility	
Linux Foundation

Chair, Protocols & Formats
Web Accessibility Initiative
World Wide Web Consortium (W3C)

Received on Wednesday, 12 January 2011 23:41:38 UTC