- From: Nigel Megitt <nigel.megitt@bbc.co.uk>
- Date: Wed, 1 May 2019 13:49:17 +0000
- To: "public-audio-description@w3.org" <public-audio-description@w3.org>
- Message-ID: <D8EF625B.42818%nigel.megitt@bbc.co.uk>
All,
Attendance was low today but Matt and I went ahead with an issue and pull request review session regardless. I recorded minutes which are available in HTML format at https://www.w3.org/2019/05/01-audio-description-minutes.html
Please do look at the open issues and pull requests on the repository https://github.com/w3c/adpt and add any thoughts, comments, suggestions. After reviewing the two open issues without proposed changes, and seeing no reason not to proceed, I will open pull requests to address them, for further review.
In summary, I think we will target as small a specification as we can at first, which means no support for SMPTE timebase and only one description active at any one moment in time. This should make implementation easier. If strong use cases arise for adding features we can assess those and decide what to do – I would strongly encourage deferring adoption until later versions of the specification unless the specification is strongly broken otherwise.
Those minutes in text format:
[1]W3C
[1] http://www.w3.org/
Audio Description Community Group
01 May 2019
[2]Agenda
[2] https://lists.w3.org/Archives/Public/public-audio-description/2019Apr/0002.html
See also: [3]IRC log
[3] https://www.w3.org/2019/05/01-audio-description-irc
Attendees
Present
Nigel, Matt
Regrets
Chair
Nigel
Scribe
nigel
Contents
* [4]Topics
1. [5]Review this agenda
2. [6]Open pull requests
3. [7]Issue 8 Constrain to one leaf element being active
for audio at any one time?
4. [8]Issue 10 Do we need to support `#contentProfiles`
or `#processorProfiles`?
5. [9]Future Meetings
6. [10]AOB
7. [11]Meeting close
* [12]Summary of Action Items
* [13]Summary of Resolutions
__________________________________________________________
<scribe> scribe: nigel
Review this agenda
group: agenda looks fine
Nigel: Attendance is low today but we'll use the time to make
what progress we can
Open pull requests
Nigel: Looking at #6 [14]https://github.com/w3c/adpt/pull/6
... Editorial fixes. Closes #4. Also makes any fragment ids
valid.
... First to note is that Cyril raised the issue but hasn't
managed to review the pull request yet.
... In the diff there are some changed dispositions - this is
because I based the branch for this on the branch for pull
... request #9 so the disposition changes are included in
those.
[14] https://github.com/w3c/adpt/pull/6
Matt: This looks fine to me. [approves pull request]
Nigel: Thank you!
...
... The next one is pull request 7, Issue 0005 restructure #7
[15]https://github.com/w3c/adpt/pull/7
[15] https://github.com/w3c/adpt/pull/7
Matt: Got that. Yes, this is a bigger change.
... This is a reordering of the sections rather than the
content?
Nigel: Yes, mainly.
Matt: What was the driver for this change?
Nigel: This was the feedback that we generated from the
previous meeting where we tried to think about the best
... structure for the document especially for people coming to
it for the first time.
Matt: I remember
... You've added a real world example?
Nigel: The new appendix D is renamed, the other examples were
in other appendices, but I moved them into the Introduction.
... Section 2.1
... The diff has done something strange to the formatting of
the example includes, the preview does a better job.
Matt: Yes, that makes sense
Nigel: One of the goals we had was to move the more important
stuff higher up the document.
Matt: That makes sense.
Nigel: One effect is to move the requirements to the appendix
Matt: This makes sense to me - I'm impressive we've kept it
this small a document!
... [approves]
Nigel: Thank you!
...
... Next is Permit only media timebase #9
[16]https://github.com/w3c/adpt/pull/9
[16] https://github.com/w3c/adpt/pull/9
Matt: I said in the issue this fundamentally makes sense.
... It potentially makes it more of a distribution format than
an archive format.
... Implementers may have a strong view.
Nigel: I was hoping to elicit that view!
Matt: My gut feeling, and talking with other broadcasters
about, say, EBU-TT Part 2 as a way into EBU-TT, is that
... the ability to support the backwards compatibility remains
very strong. Anything that forces a move away from that
... purity of form may present implementation challenges in the
future.
Nigel: I think I understand.
Matt: If we try to chase something technically pure here then
real world implementers may have problems if they
... cannot access the out of band data to recalculate any
offsets.
... They may try to introduce their own fields or data to
handle that within the file which may have unintended
... consequences or proliferation in varied practices.
Nigel: That's true.
Matt: If this is used for the share of data, and everything we
have built so far is built on SMPTE timecode, I don't think
... we ever put any content in before the start of media. In
all practical purposes it doesn't make a difference but it does
... require knowledge or an assumption about the start of media
and how it relates.
... I don't know what role this group has in proffering
suggestions for good practice operationally.
... We could give guidance about how to deal with this
challenge, and if we don't then people might be creative
... and could come up with different solutions.
Nigel: That's true. I think this group is perfectly placed to
provide informative guidance of that sort.
... The other observation I would make is that we can start
small and if there's a need to introduce SMPTE later we
... can do that. It's easier than to remove unneeded features.
Matt: We just need to be sensitive to the trigger for that
process.
Nigel: Yes
Matt: We would need to stay on top of when that may need to be
revisited, rather than people coming up with creative
workarounds.
Nigel: Agreed.
... Given I've talked to 4 members so far who have been able to
live without SMPTE, and nobody has objected, I feel
... we have enough consensus to proceed with the change.
... That takes us to the pull request itself.
... I removed everything to do with clock and smpte timebase
and also noticed that region timing should be removed.
... It was already omitted, so there's no substantive change
for that feature; the prohibition is just made explicit.
Matt: Looks good to me, I will approve this. I will add a note
too.
Nigel: Great, thank you.
...
... There's one more: pull request 11 Change DFXP references to
TTML2 profiles [17]https://github.com/w3c/adpt/pull/11
[17] https://github.com/w3c/adpt/pull/11
Matt: Just looks like a note in section 4
Nigel: That's correct - the heritage is a bit of copy and paste
from IMSC, which was based in TTML1, which didn't
... define as many standard profiles as TTML2 does. This work
has to be based on TTML2 so I've changed it to match.
Matt: Looks good to me. [approves]
Nigel: Thank you, I'll merge those all later.
... Actually just to note I had a message from Peter Spoor
saying he will take a look, and Chris O'Brien also said he
would
... review the changes, so I'll give them a couple of days to
add any more comments before merging them.
Issue 8 Constrain to one leaf element being active for audio at any
one time?
[18]https://github.com/w3c/adpt/issues/8
[18] https://github.com/w3c/adpt/issues/8
Matt: The reasons for allowing this would need other changes in
the specification.
Nigel: For example?
Matt: Say if this is used for spoken subtitles with multiple
voices, then simply permitting multiple leaf nodes isn't
enough.
... We should work through the use case and see what is needed.
... It is not a current requirement for this document.
Nigel: Good point, we may add support for that later,
potentially.
... Just to note as well that Pilar raised an issue on the
w3c/tt-reqs repo asking for spoken subtitles support,
... and I think the requirement is all about signalling, rather
than content, and I have an action to try to summarise that
... and get her views on that summary.
...
... Right now, the two commenters are in agreement for this
issue.
... In the absence of any contrary views I will prepare a pull
request to make the change.
Issue 10 Do we need to support `#contentProfiles` or
`#processorProfiles`?
[19]https://github.com/w3c/adpt/issues/10
[19] https://github.com/w3c/adpt/issues/10
Nigel: Here I'm proposing we require support for the features
in processors but not for them to be used in documents,
... where usage would be optional.
... It might be a useful output of this group to describe a
best practice in how to use those features, at some point in
the future.
Matt: Agreed
Nigel: We have agreement from the two commenters so far and no
contrary views so I will prepare a pull request for this.
Future Meetings
Matt: I could support a later slot in the day
Nigel: That would probably be helpful for folk in Canada or the
US, say.
... I feel that we have had low attendance and input so far,
which may be a sign that everyone's happy, but I would
... be more confident if there were a higher volume of input.
... So I'll look to set up some calls, maybe just a couple, in
May, at different times, to try to give more opportunity
... for participation.
Matt: Makes sense.
Nigel: Thank you, that's an action on me!
AOB
Nigel: BBC put out a consultation about its iPlayer service
recently, and RNIB contributed feedback,
... which is public at:
[20]BBC Consultation
[20] https://www.bbc.com/aboutthebbc/reports/consultation/iplayerpit
[21]RNIB input (PDF)
[21] http://downloads.bbc.co.uk/aboutthebbc/reports/consultation/iplayerpit/rnib.pdf
scribe: I mention this because RNIB states support for a
standard format for audio description, which this work is
intended to create.
Meeting close
Nigel: Thank you for that - only two of us on the call today
but I think we made useful progress. [adjourns meeting]
Summary of Action Items
Summary of Resolutions
[End of minutes]
__________________________________________________________
Minutes manually created (not a transcript), formatted by
David Booth's [22]scribe.perl version 1.154 ([23]CVS log)
$Date: 2019/05/01 13:43:29 $
[22] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
[23] http://dev.w3.org/cvsweb/2002/scribe/
Received on Wednesday, 1 May 2019 13:49:42 UTC