Skip to content

Deprecate Media Overlays embedded media#2402

Merged
mattgarrish merged 7 commits intow3c:mainfrom
marisademeglio:fix-issue-2397-mo-embed
Aug 30, 2022
Merged

Deprecate Media Overlays embedded media#2402
mattgarrish merged 7 commits intow3c:mainfrom
marisademeglio:fix-issue-2397-mo-embed

Conversation

@marisademeglio
Copy link
Contributor

@marisademeglio marisademeglio commented Aug 16, 2022

Fixes #2397

In the core spec:

Removed 9.3.2.4 entirely, added note to MO text element about how referring to embedded timed media with the MO text element is possible but unspecified.

In the RS spec:

Removed 9.3.2 and added a note about deprecation to the top of 9.3.

See:


Preview | Diff

In the core spec:

Removed 9.3.2.4 entirely, added note to MO text element about how referring to embedded timed media with the MO text element is possible but unspecified.

In the RS spec:

Removed 9.3.2 and added a note about deprecation to the top of 9.3.
@marisademeglio marisademeglio changed the title Deprecated Media Overlays embedded media Deprecate Media Overlays embedded media Aug 16, 2022
@mattgarrish
Copy link
Member

I wonder if we should give this the TTS treatment and instead of making it like it never existed, move the authoring and reading system requirements to a note? (Maybe including the TTS section.)

It might never get implemented, but publishing as a note leaves it as an option.

@marisademeglio
Copy link
Contributor Author

I wonder if we should give this the TTS treatment and instead of making it like it never existed, move the authoring and reading system requirements to a note? (Maybe including the TTS section.)

It might never get implemented, but publishing as a note leaves it as an option.

I honestly don't think it's worth the effort. I don't think it will ever get implemented and to specify it properly, even as a note, would be a not-insignificant time investment.

@iherman iherman requested review from iherman and mattgarrish August 17, 2022 04:03
Copy link
Member

@iherman iherman left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@marisademeglio (also answering your question) the entries in the change sections are added by hand. Copy-paste-modify is your friend...

@iherman
Copy link
Member

iherman commented Aug 17, 2022

I wonder if we should give this the TTS treatment and instead of making it like it never existed, move the authoring and reading system requirements to a note? (Maybe including the TTS section.)
It might never get implemented, but publishing as a note leaves it as an option.

I honestly don't think it's worth the effort. I don't think it will ever get implemented and to specify it properly, even as a note, would be a not-insignificant time investment.

I agree. The TTS seems to be on much more stable foundations, using external and, mainly, more widely known standards. Although MO is based on SMIL, the latter is, alas!, a completely unknown standard and probably the only reason it has not been deprecated is EPUB's usage of (a small subset of) SMIL. A WG note, to be made worthy of publication, would require quite some work.

@mattgarrish
Copy link
Member

The TTS seems to be on much more stable foundations

Look out for multiple renditions, in that case...

At any rate, if we don't, then this introduces ambiguity onto the authoring side, as it doesn't say what's supposed to happen now if you do reference an audio or video element. It might be helpful to add a note where media elements are mentioned that previous version of the specification defined auto-playback of these elements, but it's no longer the case (if audio is not provided, then TTS kicks in). Authors should consider alerting users that these elements exist and they need to go back and find them after listening to the work.

Comment on lines +1869 to +1870
<p class="note">Earlier versions of this specification included some information about embedded audio and
video [epubmediaoverlays-32]. This feature has been deprecated.</p>
Copy link
Member

@mattgarrish mattgarrish Aug 17, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We can't deprecate in a note, otherwise the deprecation doesn't have normative standing. I'd recreate 9.3.2 rather than completely deleting the section so it matches other deprecated features. I can't suggest a fix through the github interface, but put back something like:

<section id="sec-embedded-media">
	<h4>Embedded audio and video (deprecated)</h4>
	<p>Automatic playback of embedded audio and video is now <a data-cite="epub-33#deprecated">deprecated</a>.</p>
	<p>Refer to its definition in [[epubmediaoverlays-32]] for more information.</p>
</section>

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This change should also be integrated before merging.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done in 9d5bb34

@mattgarrish
Copy link
Member

mentioned that previous version of the specification defined auto-playback of these elements

Actually, re-reading the note that's there I can probably live with it as-is. I was trying to read the line in the diff and it looked like it was only referring to video.

Co-authored-by: Ivan Herman <ivan@ivan-herman.net>
@iherman
Copy link
Member

iherman commented Aug 18, 2022

The TTS seems to be on much more stable foundations

Look out for multiple renditions, in that case...

At any rate, if we don't, then this introduces ambiguity onto the authoring side, as it doesn't say what's supposed to happen now if you do reference an audio or video element. It might be helpful to add a note where media elements are mentioned that previous version of the specification defined auto-playback of these elements, but it's no longer the case (if audio is not provided, then TTS kicks in). Authors should consider alerting users that these elements exist and they need to go back and find them after listening to the work.

There is such a mention in the RS spec. Do you mean to add something similar in the core spec as well?

@iherman
Copy link
Member

iherman commented Aug 26, 2022

The issue was discussed in a meeting on 2022-08-26

List of resolutions:

  • Resolution No. 1: Once the edits are finalized, merge PR 2402 and close issue 2397.
View the transcript

1. Deprecate embedded audio or video as targets for Media Overlay <text> references.

See github issue epub-specs#2397.

See github pull request epub-specs#2402.

Ivan Herman: See Description of the issue background.

Ivan Herman: on a high level, what you can do with MO is associate a piece of time-stamped audio with a piece of text.
… it is used in a bunch of a11y situations.
… it is pretty well specified.
… based on the tests I have made so far it is well implemented too.
… however the specifications also try to allow association of an audio clip with a video (or several videos).
… now you have two time related media side-by-side.
… for that to work, the spec needs to define what is happening, considering that videos themselves often have UI controls themselves.
… the spec tries to make some sense of this use-case, but this is really underspecified.
… also, even for the cases which are relatively clear, we haven't found any implementations.
… so we have a complicated case, underspecified in spec, and not implemented.
… in side conversation with Avneesh, we did not find any important proper use case in a11y area.
… the tests I created were trying to take a video and add a piece of background audio. But this is an artificial use case, as adding background audio to a video can be done much more easily in editing software rather than via MO.
… so this lead to suggestion that we should just deprecate this type of video/audio MO from spec.
… Marisa, mgarrish, and I opened #2397 to discuss.
… based on this Marisa created PR #2402.
… but we wanted to discuss with WG first.

Dave Cramer: the fact that we don't have implementations of this weighs heavily to me.
… reluctant to keep this in the spec for that reason.

John Foliot: is anyone aware of popcorn.js?.

Ivan Herman: is there any use-cases for this in the a11y area which requires this?.
… MO has been included in the spec primarily for a11y reasons, although it is useful for other things.

George Kerscher: i agree generally. Audio descriptions for the blind is implemented in TV as a second audio channel that can be turned on.
… don't know the technology behind it (mixer behind the scenes that combines audio channels?).
… don't know that we can break new ground here by adding that feature, but its something people would want.
… what happens with descriptive audio is that the on-screen action is described for the blind person.

John Foliot: http://ncamftp.wgbh.org/rd/hgp/hgp.html.

John Foliot: watch the example at: http://ncamftp.wgbh.org/rd/hgp/hgp.html.

John Foliot: there is a package called popcorn.js. Work done a while back by WGBH, where they were using this js library to use time-stamp markers to activate functions.
… this is a demonstrated use-case, but i'm not aware of anyone who has taken this and run with it.
… not sure that we need this in our spec, as long as we're not frustrating the use of it in epub.

Avneesh Singh: this is removal of 9.3.2 from the RS spec. No one is blocking the creation of such MO files.
… say you have a video going on, and you need description for it, MO doesn't seem like a good solution for this.

Ivan Herman: the largest part of the MO concentrates on associating audio track with text.
… this is not being touched.
… the section that is affected is "embedded-media overlays", i.e. where HTML includes a video track.
… to what John said, if there are js libraries that can do things in a website, they can do it in an epub as well. No problem..
… what you described is a nice tool that can be used to get additional feature while a video is running. This can work in an epub as well.

Wendy Reid: agreed. I think we're just removing something that is creating confusion/opening ourselves up to adding another unimplemented feature.
… i haven't seen this in an epub or in any RS.
… description of video of visually impaired is not done via MO.
… RS would have to implement accessible media-player or epub would have to come with one, larger problem.

John Foliot: More info about accessible media requirements: https://www.w3.org/TR/media-accessibility-reqs/.

Ivan Herman: can we pass a resolution that once editors are ready, that we can merge PR #2402?.

Proposed resolution: Once the edits are finalized, merge PR 2402 and close issue 2397. (Wendy Reid)

Matthew Chan: +1.

Wendy Reid: +1.

Dave Cramer: +1.

GeorgeK: +1.

Ivan Herman: +1.

Brady Duga: +1.

Masakazu Kitahara: +1.

Avneesh Singh: +1.

John Foliot: +1.

Charles LaPierre: +1.

Resolution #1: Once the edits are finalized, merge PR 2402 and close issue 2397.

@iherman
Copy link
Member

iherman commented Aug 26, 2022

@marisademeglio @mattgarrish we have now a WG resolution to merge this after we all agree on the textual details. From my point of view, it is a go; I do not have any strong feeling about #2402 (comment)

@marisademeglio
Copy link
Contributor Author

@marisademeglio @mattgarrish we have now a WG resolution to merge this after we all agree on the textual details. From my point of view, it is a go; I do not have any strong feeling about #2402 (comment)

Great, no more changes needed then ?

@iherman
Copy link
Member

iherman commented Aug 27, 2022

@marisademeglio @mattgarrish we have now a WG resolution to merge this after we all agree on the textual details. From my point of view, it is a go; I do not have any strong feeling about #2402 (comment)

Great, no more changes needed then ?

Well, you should wait for @mattgarrish : he seems to have proposed some changes and those seem to be open still. Once that is settled it is a go...

@mattgarrish
Copy link
Member

This looks fine to me now but looks like we need to add a biblio entry for MO 3.2 to the RS file. I'm going to merge and manually correct that now.

@mattgarrish mattgarrish merged commit a3cd243 into w3c:main Aug 30, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Deprecate embedded audio or video as targets for Media Overlay <text> reference

3 participants