- From: Nigel Megitt <nigel.megitt@bbc.co.uk>
- Date: Thu, 25 May 2017 16:16:14 +0000
- To: Timed Text Working Group <public-tt@w3.org>
- Message-ID: <D54CC174.40508%nigel.megitt@bbc.co.uk>
Thanks all for attending today's TTWG meeting. Minutes can be found in HTML format at https://www.w3.org/2017/05/25-tt-minutes.html
There was a request to draw your attention specifically to Handling of LWSP in tts:fontFamily #248<https://github.com/w3c/ttml1/issues/248> – please look at this and if you have a preferred option add a thumbs-up +1 reaction to it.
In text format:
[1]W3C
[1] http://www.w3.org/
Timed Text Working Group Teleconference
25 May 2017
See also: [2]IRC log
[2] http://www.w3.org/2017/05/25-tt-irc
Attendees
Present
Nigel, Dae, Pierre, Nigel, Glenn
Regrets
Andreas, Thierry
Chair
Nigel
Scribe
nigel
Contents
* [3]Topics
1. [4]This Meeting
2. [5]TTML2 - Audio Description
3. [6]LWSP between non-terminal components of style and
parameter attributes
4. [7]IMSC Pull Request 229
5. [8]PQ HDR in PNG
6. [9]Logical vs Physical pixels
7. [10]TTML issues
8. [11]TTML progress tracking
* [12]Summary of Action Items
* [13]Summary of Resolutions
__________________________________________________________
<scribe> scribe: nigel
This Meeting
Nigel: I don't think we have anything to discuss on TPAC today;
... on TTML2 there's a request to discuss Audio Description
(AD) at the head of the meeting
... and we have a bunch of TTML issues and pull requests plus
some activity on IMSC.
Pierre: We have a blocker for pull request #230 so we can not
close that until we have
... resolved the bigger question in TTML1.
Nigel: The HDR in PNG is currently paused awaiting feedback so
we should cover that
... briefly.
... Any other business or points that anyone wants to make sure
we cover today?
group: [silence]
Nigel: OK that's our agenda.
Pierre: I'm not opposed to AD today but without Mike and
Andreas being present it might
... need to be repeated later.
Dae: I see that Mike has views on this too.
Nigel: I've had private conversations with Mike about AD and I
think he has reservations
... but is willing to accept it, probably.
TTML2 - Audio Description
Nigel: In terms of a walk-through, the first thing is the
Requirements:
[14]Audio Description Requirements
[14] https://github.com/w3c/ttml2/wiki/Audio-Description-Requirements
Nigel: These were circulated in September and revised following
feedback in October to
... establish the alignment with the MAUR requirements covering
the same space.
... It is envisaged that, workflow-wise, small additions to
TTML2 would satisfy any subset
... of the workflow described.
Pierre: [concerns about the ability for TTWG to influence
distribution formats]
Nigel: [commitment from BBC to produce an open source
implementation showing how
... client side mixing could work in a browser] No intention to
require or request that
... TTML2 is used as a distribution format for AD.
Glenn: Doing integration work to add audio into TTML2 it became
clear that the semantics
... for audio were incomplete. TTML has requirements for text
to speech.
... Adding further functionality that completes the audio
semantics and also happens to
... satisfy the ability to provide the audio description
service requirements is something
... we should do to make audio work in TTML. This is the wrong
time to be making a
... decision about whether to include these new features - if
it goes into the spec then
... we have a process for figuring out if it stays in beyond
CR, based on implementations.
Nigel: Thanks both. So that's a quick summary of the
requirements, then in the issue:
[15]Add support for Audio Description requirements
[15] https://github.com/w3c/ttml2/issues/195
Nigel: I describe some specific semantic requirements for
TTML2.
... [describes the audio processing model being proposed]
... Map from a TTML2 element structure into a Web Audio graph
... Additionally I see that text to speech semantics were added
by Glenn subsequently,
... which is fine.
Glenn: My audio model for TTML is quite simple. Right now TTML
produces a series of
... raster images typically on visual presentation medium. It
does not yet have an audio output,
... but it certainly can. All of the equipment that I'm looking
at in terms of supporting the
... audio feature set basically ends up producing a single
audio output stream similarly
... to how the visual rendering part of TTML produces a single
motion picture output.
... That would be the more generalised output. As for the model
for processing audio, I'm
... currently assuming that it's possible to do it all in
software but that implementations
... could optionally use hardware features to optimise and
improve efficiency. For example
... mixing, panning, text to speech synthesis, I'm assuming it
is possible to do all these in
... software in a presentation engine.
Nigel: There's good evidence for that by the way.
Glenn: On some devices that may be impractical, just as it may
be impractical to do
... visual compositing and HDR color on some devices.
Nigel: Thanks Glenn, that's the summary. Any other questions?
Dae: It seems like it won't be in IMSC2. If AD is threatening
the timescale of WR, are we
... agreed to remove it?
Nigel: I agree that it would not be in IMSC 2. I'm not happy to
agree with removal from WR at this stage.
Dae: [question about handling of WR comments and if they are
prioritised for audio
... relative to other features]
Nigel: No specific priority given for or against any feature
set including audio.
Glenn: I think we should not talk about this in terms of AD but
just in terms of audio
... features. We should just be focusing on generic audio.
Nigel: +1 to that - in terms of profiles, we may later want an
AD in TTML2 profile based on
... the audio semantics in TTML2, so that is the point when we
should discuss AD more
... specifically.
Dae: [question about removing features, and when in the process
to do this]
Nigel: Typically we would not remove features unless they are
at risk in CR and they
... prevent us meeting the CR exit criteria, in which case we
would remove them to move
... to PR.
... Conversely during WR we need to agree a disposition of all
received comments, and if
... in extremis we get feedback explaining that we have a big
technical problem with a
... feature then we could agree to remove that feature prior to
CR.
Dae: [does not want to extend the publication period for WR]
Nigel: I have committed to put effort into getting the audio
features in within the agreed
... timescales for WR.
Pierre: [queries expertise in group for dealing with audio and
AD
... and ability to specify this and influence industry]
Nigel: [does not consider there to be a specific requirement
for vendors to be members
... of the group]
Dae: [concerns that a large number of audio-related comments
during WR could delay
... onward progress]
Nigel: We've discussed this for about an hour now, let's take a
break and reconvene in 5
... minutes and begin with the TTML issue that's also blocking
IMSC.
LWSP between non-terminal components of style and parameter
attributes
[16]Required spaces between non-terminal components of styling
and parameter attributes (issue #221)
[16] https://github.com/w3c/imsc/pull/230
[17]LWSP in rgba expressions?
[17] https://github.com/w3c/ttml2/issues/315
Nigel: Interesting there's no TTML1 issue.
Pierre: It looks like we have to solve this in TTML1 to make
progress in IMSC.
Glenn: I think there's a fair consensus at this point (to
verify) that white space around
... COMMA in tts:fontFamily is probably justified in TTML1, but
not in RGB expressions.
... The justification is based on actual usage; we have plenty
of fontFamily examples where
... there is white space but I cannot find any in RGB
expressions in TTML, though there are
... plenty in CSS. My suggestion is we add a note to TTML1
clarifying that LWSP is permitted
... around commas in fontFamily and another emphasising if
necessary that it is not
... permitted in RGB expressions. I have a different answer for
TTML2 however.
Nigel: Would that work in TTML1 Pierre?
Pierre: I'm not sure I'm there yet. There's evidence for spaces
in fontFamily so we have to
... deal with this. What's not clear in my mind is if the
revised TTML1 text should say that
... spaces shall not be there but by the way there's evidence
of documents with spaces,
... therefore processors should be tolerant. Or option 2 is to
say spaces are permitted but
... some processors may reject them. I'm not sure it's clear to
everyone. We need an issue
... in TTML1 and reference it.
... [creates issue]
[18]Handling of LWSP in tts:fontFamily #248
[18] https://github.com/w3c/ttml1/issues/248
Nigel: The next step is to solicit feedback?
Pierre: Yes, or to make a call if no feedback is received.
Nigel: As a strawman I would propose to adopt the second option
as a default, i.e. to
... permit LWSP in document instances.
Pierre: I don't have a strong opinion - I need to discuss it
with others.
Glenn: I do have a strong opinion, which is the second option,
to permit LWSP in document instances.
Nigel: I will add each option as a separate comment in the
issue so people can add a +1
... reaction to the option they prefer.
... They're now added on #248.
... Is there anything else we can do to unblock that issue now?
Pierre: No, if you know vendors or users I would encourage them
to weigh in on that issue.
Glenn: Skynav will be able to point to 4 independent
implementations that the second
... option is the better one, going back to 2004 timeframe.
IMSC Pull Request 229
[19]Remove tabs from examples (issue #225) #229
[19] https://github.com/w3c/imsc/pull/229
Pierre: I'm waiting for your input Nigel.
Nigel: Apologies, thank you for the reminder.
Glenn: I think you should remove the white space at the ends of
the lines.
Nigel: Is this example supposed to test functionality in the
presence of white spaces at
... the ends of lines?
Pierre: No.
Nigel: In that case we should make them as simple as possible
so we should remove the
... whitespace at the ends of the lines. I will update the PR.
Glenn: I agree that extraneous content should be removed from
examples, more generally.
Nigel: I've added a comment to the pull request.
PQ HDR in PNG
Nigel: I raise this topic just to note that there is a strong
objection to the current approach
[20]ICC profile contents are ignored, magic string used for
labelling #3
[20] https://github.com/w3c/png-hdr-pq/issues/3
Nigel: Please look at the issue and the linked thread in the
Color on the Web CG reflector.
Pierre: I've tried to get to the bottom of what was concerning
Chris because it was not
... clear and it sounds like the main concern was the
implication in the proposed Note that
... prohibits use of the embedded ICC profile, which is not
what it was meant to mean.
... I have offered to clarify it so that it does not imply that
at all.
... As far as I can tell that is the concern. It is possible
that an ICCMax profile might be
... better, and that could be added later. My main blocker is
getting an ICC profile that
... does not have an Adobe copyright, so that's what I'm
working on right now.
... When I get that updated profile I will commit it, and I've
encouraged Chris to revise his
... issue to be more specific. In the absence of that I will
create a pull request that hopefully
... will satisfy Chris.
Nigel: I think it would be reasonable for me to ping Chris with
a reminder tomorrow, given
... that I proposed a week's review 6 days ago.
Pierre: If he can't agree to that review period then we can say
that when we have resolved
... the copyright issue we plan to move forward, which sets a
new effective deadline for
... expressing his concerns.
Logical vs Physical pixels
Pierre: Have we got anything more to discuss on this?
Glenn: Last night I did a subsequent edit, and I believe it is
all resolved now.
... I took out unreferenced terminology for example everything
with Viewport.
Pierre: Ok great.
Glenn: I also took others out like dot pitch, storage units,
sample units and substituted
... logical pixels. I took out the word "physical" and used the
word "display" pixels and
... referred to hypothetical output devices as opposed to
actual output devices.
Nigel: Is this in a pull request?
Glenn: It has been merged already.
[21]Apply improvements to aspect ratio and pixel semantics.
#321
[21] https://github.com/w3c/ttml2/pull/321
scribe: There is one open issue on the definition of "contain"
.
... #30 has now been closed.
Pierre: It sounds like we ended up in a place where the root
container coordinate system
... is all logical, right?
Glenn: Right. We ended up with logical pixels and display
pixels and logical pixels have
... no aspect ratio, but that there's a logical transformation
via PAR that creates display
... pixels from logical pixels.
Pierre: Why have two terms?
Glenn: In my mind you author into logical pixels and there's a
transformation process that
... turns them into dimensioned pixels, and I feel that's an
important concept to retain.
... I am also trying to converge with the SVG model.
Pierre: I think we're very close to having logical coordinates.
Glenn: Please review the changes and raise issues.
Nigel: I've just found a broken link in 10.2.33 to presentation
context coordinate space.
Glenn: That's a link I removed so thanks, I'll resolve that
today.
Nigel: Ok that's an action on all to review.
[22]deprecate use of pixel units unless tts:extent on root
element is in pixels #330
[22] https://github.com/w3c/ttml2/issues/330
Glenn: Is there anyone who disagrees with this issue?
group: [silent]
Glenn: I'll take that as tentative support.
TTML issues
[23]revert deprecation of ttp:profile on root element #331
[23] https://github.com/w3c/ttml2/issues/331
Nigel: I've commented on that.
Glenn: Okay I'll look.
Nigel: I think there's evidence that people are trying
deliberately to target multiple
... processor profiles.
... Is there any possibility of extending the value space to
allow the short codes?
Glenn: I doubt that, due to the use of URI.
... Right now it is deprecated, we are just talking about
undeprecating it.
Nigel: To me undeprecating it could be improved maybe by
extending it too. I'm not hugely
... concerned.
TTML progress tracking
Glenn: We are getting near the end of May. We plan to have
everything at least in pull
... request form, so I will work this week to knock off the 24
items on my list. I have been
... taking things off my list that have no pull request.
Nigel: Please do not close issues just because you do not
intend to work on them. We will
... leave issues open so that people who want can pick them up
until we decide to
... assess the open issues and choose to close or defer those
we do not have work for.
... The reason for leaving issues open is to signal that work
is needed.
Dae: There are 18 features that need image example - I may need
help with those.
Nigel: Okay, feel free to send a message to the group reflector
listing those or send the
... list to me and I can include it in the agenda that I will
send on Tuesday.
... Do you also need the example documents from which to
generate the images?
Dae: That would help.
Nigel: We're out of time for today. Thanks all. [Adjourns
meeting]
Summary of Action Items
Summary of Resolutions
[End of minutes]
__________________________________________________________
Minutes formatted by David Booth's [24]scribe.perl version
1.152 ([25]CVS log)
$Date: 2017/05/25 16:11:18 $
[24] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
[25] http://dev.w3.org/cvsweb/2002/scribe/
----------------------------
http://www.bbc.co.uk
This e-mail (and any attachments) is confidential and may contain personal views which are not the views of the BBC unless specifically stated.
If you have received it in error, please delete it from your system.
Do not use, copy or disclose the information in any way nor act in reliance on it and notify the sender immediately.
Please note that the BBC monitors e-mails sent or received.
Further communication will signify your consent to this.
---------------------
Received on Thursday, 25 May 2017 16:16:47 UTC