Timed Text Working Group Teleconference -- 19 Jun 2014

<trackbot> Date: 19 June 2014

<scribe> scribeNick: nigel

Actions

action-297?

<trackbot> action-297 -- Glenn Adams to Soften language about 'willful violation' regarding treatment of duration zero -- due 2014-06-12 -- PENDINGREVIEW

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/297

close action-297

<trackbot> Closed action-297.

action-300?

<trackbot> action-300 -- Glenn Adams to Add note to issue 263 explaining how ttml2 additions for @{extends,restricts} (partially) addresses this issue -- due 2014-06-19 -- PENDINGREVIEW

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/300

issue-263?

<trackbot> issue-263 -- profile feature set may not match intended feature constraints -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/263

glenn: I added the note.

close action-300

<trackbot> Closed action-300.

Issues

issue-263?

<trackbot> issue-263 -- profile feature set may not match intended feature constraints -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/263

glenn: Implemented since May 21. Initially raised by Mike Dolan so we may need him to review it.
... There's a question: should we update SDP-US. We don't normally update notes, so this is a process issue.
... What I implemented potentially resolves the issue raised.
... We should probably check with Mike and call for consensus to the group requesting input and close if no objections otherwise we'll see if someone wants to file an issue to update SDP-US in some way.

ack

nigel: I understood that though this is a problem surfaced by SDP-US it's really something for use in future specifications.
... I'm happy to leave this pending review and seek input as proposed.

pal: Glenn, if you could crystallise your thoughts on how this would work that would be helpful.

glenn: There's no formal way in TTML1 so the only option is to write it into prose.
... However if you write a profile that uses an existing feature and then says in the prose that it's a restriction or extension without defining a new feature that would be problematic.
... Right now the way people have been writing profiles has not included defining new features or extensions. For example CFF-TT and EBU-TT they just say 'use this feature' but we're restricting it semantically or syntactically.

pal: That's true in SDP-US too?

glenn: That's right. This leads to an ambiguity because the formal definition of the features is not restricted.

pal: In the context of TTML1 how important/tolerable is this?

glenn: The consequence, for TTML1 processor profiles, is that the profile asks for more support than is required by the content. Color is an example. The profile definition in SDP-US
... requires all colors to be supported by the processor but the prose only requires support for a limited subset.
... In TTML2 we're handling this differently but that's the discrepancy in TTML1.

issue-295?

<trackbot> issue-295 -- Remove code point restrictions from IMSC -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/295

pal: The two actions have been closed so if there's no new information this can be closed.

close issue-295

<trackbot> Closed issue-295.

issue-296?

<trackbot> issue-296 -- Remove xml:lang placement restrictions from IMSC -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/296

nigel: notes that Pal has posted some edits.
... This is related to issue-237 which is closed.

pal: The proposal was accepted and implemented in November.

close issue-296

<trackbot> Closed issue-296.

issue-308?

<trackbot> issue-308 -- It is unclear how a document is associated with a related video object -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/308

nigel: I still have an outstanding issue not resolved in §4.4 because the intention to display and remove ISDs on specific video frames remains.
... I think I'm happy with the time expressions part though.

pal: There's also issue-317
... Which is about frame alignment.
... Issue-308 doesn't make the link between Related Video Object and Related Media Object and that's been fixed. Also 308 implies that IMSC documents have to be authored for a particular frame rate,
... which is not true unless frame rate is used.

nigel: agrees that Issue-317 is separate and that this edit fixes Issue-308.

close issue-308

<trackbot> Closed issue-308.

issue-309?

<trackbot> issue-309 -- Image profile needs to permit text equivalent -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/309

nigel: There's ongoing discussion of this issue on the email reflector, including participants not present here.

pal: How will we resolve this?

nigel: John Birch is working on an edit proposal.

pal: I believe IMSC already deals with text equivalence.

nigel: I don't think IMSC deals with this strongly enough at present.
... I believe we must allow alternative text representations alongside images, for content providers to optionally use.

pal: Must is a strong statement - we need to get understand why this level of requirement applies here given that content providers can generate whole alternate documents.
... This depends on the application. The technology supports all the applications.

nigel: The specification only defines the contents of a single document, and if that document has image representations there's no accessible equivalent.

pal: IMSC satisfies that.

nigel: I disagree that IMSC satisfies that.

pal: If you multiplex the documents that becomes a single document. All applications I know of support multiple tracks already.

nigel: We aren't defining the application just the document format. IMSC does not support multiplexing at the moment.

pal: All applications I know of for IMSC already support this and other things too.
... I'm not opposed to informative metadata as John has pointed out, that can be useful for debugging. That's probably a good idea. I see potential harm but no requirement in specifying a less flexible and rich profile than text profile.

nigel: I agree and propose that documents should be able to contain both text profile and image profile and a selectivity mechanism for the processor to choose which representation to use.
... It's broken now because it doesn't satisfy the W3C's own guidelines for accessibility.

glenn: Historically it has never required a separate document to satisfy the requirement for alternate text representation.

pal: I believe the guidelines are intended to be more general but I did not see in the recommendations anything that points to a single document even if that's how it's been implemented so far.

glenn: Most 'lay people' would say that these requirements do apply. It may be worth getting a sense from someone e.g. involved in WCAG.
... Technically you could just put metadata in, which would be a non-standard solution and probably not adequate, but it would be useful to get input to ward off future objections.

pal: Why would metadata not satisfy the recommendations?

glenn: it might, but I can't answer this.

plh: We're talking about backgroundImage. Why would you require alternate content?

glenn: This is using backgroundImage for the content.

nigel: This originated in SMPTE-TT.

plh: The use of backgroundImage in CSS wasn't intended for this. In CSS it wasn't supposed to carry content that needs to be made accessible.
... If it's really part of the content then we should address the naming and we do need a text alternative for that.

pal: There's at least one digital cinema application when a lot of the time subtitles are delivered as image. They can also be delivered as text. These are displayed on screen for hard of hearing.
... In parallel there's another document for captions that uses the same format with a different profile, and it's always text, and is used for captioning devices.
... This is done because you typically want more control over the display of the text for captioning devices than you could by a simple alt tag on an image. This is how cinema meets accessibility requirements.
... Sometimes there are two subtitle tracks, one for language translation, another for description of sounds.

plh: I think glenn is right - we should ask. I'm wondering if it's even meaningful.
... Here those persons can hear what is there, so if they can hear the video then will they still need to have the subtitling.
... It's not just about text alternatives, but a lot more, so it would need a dedicated track.

pal: This is the only way to provide a good experience.

plh: In that case we should put some guidelines in that section that says by the way we recommend that this is not used for non-sighted applications.

pal: I really want to understand the use cases and am happy to follow up with W3C folk.

<scribe> ACTION: plh Contact W3C Accessibility expert re Issue-309 [recorded in http://www.w3.org/2014/06/19-tt-minutes.html#action01]

<trackbot> Created ACTION-301 - Contact w3c accessibility expert re issue-309 [on Philippe Le Hégaret - due 2014-06-26].

pal: I'm happy to help, as this is a different use case than HTML.

nigel: There are use cases that aren't just for non-seeing and non-hearing folk, e.g. touch-based. I've also mentioned the issue of the attribute name showBackground in the email thread.

glenn: It's possible that we can meet these genuine requirements with separate documents. We should also consider the use cases of search and information processing.

nigel: adjourn for 2 minutes until 16:03 please

issue-312?

<trackbot> issue-312 -- forcedDisplay attribute is metadata so should be in ttm not tts namespace -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/312

nigel: For me the edit in §4.5.3 has not addressed the issue since all content is expected to be displayed anyway.

glenn: We should change the name of this issue as it's not the path we want to go down.

pal: It isn't a metadata element, and is not in the ttml namespace, but is in the imsc namespace as discussed before.

nigel: I will change it.

pal: Please propose an alternative and see the example

nigel: I see no example in 4.5.3

glenn: Rather than feature designator it should be an extension designator.

pal: §4.5.3 references an example.

nigel: That example describes the use case but not the semantics of the attribute.
... This is very different from the current system model in which a document is being processed only if the content is to be shown.

glenn: The second para in §4.5.3 that describes forcedDisplay talks about SHOULD be displayed regardless of etc. This disregards the semantics of the visibility property and timing and the display property, and the mapping to a temporally active region.
... Additionally the region itself could be made invisible. This doesn't factor in all of those aspects of display into the equation so I'm concerned about how this reaches down inside the content
... and defines 'display regardless of everything else'.

pal: It's more 'don't display the other stuff' unless this setting is true.

glenn: I'm more of the opinion that this maybe should be metadata, as it's a higher level protocol that's outside of the scope of the current semantics of TTML. I'd be a lot more comfortable
... if this were defined in a metadata namespace. The language needs to be characterised as a higher level protocol, and it would be useful to have examples of ignoring and not ignoring content tagged with
... specific languages, for example Spanish and English, where some higher level processor is selecting one of those languages to display.

reopen issue-312

<trackbot> Re-opened issue-312.

issue-313?

<trackbot> issue-313 -- Presented Region section is informative only -- pending review

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/313

glenn: showBackground and visibility also apply to region.

ACTION nigel to re-review issue-313

<trackbot> Created ACTION-302 - Re-review issue-313 [on Nigel Megitt - due 2014-06-26].

issue-306?

<trackbot> issue-306 -- Appendix N.2 omits fractional time expressions -- raised

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/306

glenn: My intention was that if there were a fraction component of seconds then it would be interpreted in the seconds part.

open issue-306

reopen issue-306

<trackbot> Re-opened issue-306.

issue-314?

<trackbot> issue-314 -- Temporally active is not defined for regions -- raised

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/314

reopen issue-314

<trackbot> Re-opened issue-314.

glenn: I'm not sure if I agree with this but we can consider it. Before I would agree with that I'd like a use case.

nigel: There's an error in the issue - it should say showBackground="always" rather than "true"

glenn: I guess in this case "always" is semantically qualified by temporal activity. I agree it should be made explicit. The ancillary question is should we make "always" unqualified.

nigel: A use case is if region backgrounds should be painted even when no text is selected into them.

reopen issue-314

<trackbot> Re-opened issue-314.

issue-321?

<trackbot> issue-321 -- typo - s/ttp:feature/ttp:extension/ in 5th paragraph of 6.1.5 -- raised

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/321

reopen issue-321

<trackbot> Re-opened issue-321.

glenn: This is a copy paste error.

issue-322?

<trackbot> issue-322 -- Formula for dropNTSC time expressions is incorrect. -- raised

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/issues/322

reopen issue-322

<trackbot> Re-opened issue-322.

reopen issue-315

<trackbot> Re-opened issue-315.

reopen issue-317

<trackbot> Re-opened issue-317.

reopen issue-319

<trackbot> Re-opened issue-319.

Change Proposals

nigel: change proposal 2 https://www.w3.org/wiki/TTML/changeProposal002 was discussed last week and amended
... I believe a further edit is in the queue from CP2

jdsmith: I think so.

nigel: There are Ed Notes in 9.3.2 and 9.3.3 but not in 9.3.5 to add an elaborated example.

glenn: expresses concern that over-complicating the examples makes them less useful.
... We also might consider a section in the wiki that has more examples rather than putting them all in the spec itself.

nigel: What CPs would folk like to work on in the future?

glenn: It would be useful to prioritise. An additional column to indicate priority 1, 2 and 3 might be useful. The owner can tweak the priority.

<scribe> ACTION: nigel Add priority column to Change Proposal Index [recorded in http://www.w3.org/2014/06/19-tt-minutes.html#action02]

<trackbot> Created ACTION-303 - Add priority column to change proposal index [on Nigel Megitt - due 2014-06-26].

glenn: We should maybe put a mark in the sand for the change proposals we want to finish before we go to FPWD.

nigel: agrees.

plh: We need to do a call for exclusions at FPWD.

nigel: We set a schedule in March to get to LC in August, and haven't got to FPWD yet.

glenn: Maybe August is not impractical for FPWD at this point.

nigel: Let's aim for that.

AOB

nigel: Regrets from me for next 2 weeks.
... As noone has volunteered to chair then I propose that we do not meet until our next on 10th July.

- DRAFT -

Timed Text Working Group Teleconference

19 Jun 2014

Attendees

Contents

Actions

Issues

Change Proposals

AOB

Summary of Action Items

Scribe.perl diagnostic output