Web and TV Interest Group Teleconference

11 Jun 2014

See also: IRC log


Yosuke Funahashi, Mark Vickers, Mark Sadecki, Janina Sajka, Clarke Stevens, Paul Higgs, Glenn Deen, Bin Hu, Wu Wei, Cyril, Daniel Davis


<trackbot> Date: 11 June 2014

<Bin_Hu> azkim, aaaa is me

<scribe> scribenick: ddavis

<scribe> scribe: Daniel

<jcverdie> Regret+ JC_Verdie

yosuke: We've just finished the AC meeting and today we have the AB meeting so a relatively small number of people on this call.
... Now we're in the second round of gathering use cases and requirements.
... Firstly, we should share some information from the AC meeting that's relevant to this IG

AC meeting update

yosuke: We reported the outcome of the Munich TV workshop
... There was a good response from the audience. Some of them were interested in the actual service of using HTML5 and interactive TV programs.
... We also had an update from W3C Team about testing activities.
... W3C Team reported the first round of the testing activity - Test the Web Forward is working well but funding is not working well.
... Now they'd like to continue to work on crowd sourcing.

ddavis: Today I'm going to speak to W3C local offices about TV activities in W3C

Use-case gathering

<yosuke> http://www.w3.org/2011/webtv/track/actions/open?sort=due

yosuke: Let's look through the issue tracker

<scribe> ACTION: 194 to Clarify use case 1 description [recorded in http://www.w3.org/2014/06/11-webtv-minutes.html#action01]

<trackbot> Error finding '194'. You can review and register nicknames at <http://www.w3.org/2011/webtv/track/users>.

<yosuke> http://www.w3.org/2011/webtv/wiki/New_Ideas

Use case 1 - https://www.w3.org/2011/webtv/wiki/New_Ideas#Audio_Fingerprinting_.28watermarking.29

ddavis: I added a bit more explanation to clarify the use case
... I added this text: " In other words, the user device could extract information in the broadcast feed to retrieve additional data related to the content being played."
... Maybe it's better to have two use cases

yosuke: I think so

PaulHiggs: It's possible that the web app could be running on the TV itself rather than a second device.
... I thought we would say the ability of a web app to detect the TV program that's being viewed.
... Then the second device could extract info from the broadcast feed.

Glenn_d: What you could extract from the audio stream is a registry identifier or ad ID (inside the watermark).
... Once you have that, you can resolve it to get the show and episode identifier.
... It won't tell you where you're watching but what you're watching.

PaulHiggs: The example is correct but the second sentence of the description doesn't seem to be correct.
... You could add something about listening to its audio information to get further information from the internet.

Glenn_d: There are two ways of doing this - either there's a watermark you extract or you're going to take the audio and match it against an existing fingerprint.

PaulHiggs: That's why it needs to be two use cases.
... Watermarking is non-audible tones that the tablet can hear. Fingerprinting is listening to a clip and comparing it to a database.

Glenn_d: There's also a third case - the option that the TV device may have some fingerprinting tech built-in. It can notify the tablet of what's being watched.

?: Samsung TVs have a watermarking feature that can do this.

PaulHiggs: Is that audio or something else?

<Glenn_d> i'm not on the call

<Glenn_d> np :)

PaulHiggs: So in that third case you're saying the TV is doing the detection but not using audio.

Glenn_dd: Correct.

ddavis: Is just two use cases enough?

PaulHiggs: I had thought maybe we don't need the third one - it's mostly about getting the information on the companion tablet.

Glenn_dd: There is one difference - the industry uses a 6-second clip for fingerprinting and can tell where in the show I'm watching.
... Whereas with watermarking is just an identifier for the show.

Clarke: Fingerprinting is taking information that's already there are characterizing it.
... Watermarking is added inaudible info to the stream to identify the source or add e.g. copyright information.

PaulHiggs: The watermarking technology could do more that just identify the show.

yosuke: The benefit of fingerprinting is you don't need to add info to the stream beforehand.
... If you add more computing resources you can get more precise info of where you're watching.
... For the use case, when you're listening to the stream, without the original provider adding metadata or audio watermarking, you can use a fingerprinting service to get program info.
... For watermarking, the video creator can add anything they like within the stream and the receiver can decode that info.
... The service provider can deliver additional info to applications.

Glenn_dd: Some places have discussed a hybrid system - a watermark to obtain the show and a fingerprint to sync where you're watching.

<scribe> ACTION: ddavis to split up use case 1 into watermarking and fingerprinting. [recorded in http://www.w3.org/2014/06/11-webtv-minutes.html#action02]

<trackbot> Created ACTION-200 - Split up use case 1 into watermarking and fingerprinting. [on Daniel Davis - due 2014-06-18].

Action - ddavis to Check current ability to change playback rate for html media element.

<trackbot> Error finding '-'. You can review and register nicknames at <http://www.w3.org/2011/webtv/track/users>.


ddavis: This was already in the spec so I removed the use case.

janina_: This is very useful, e.g. for immigrants, and can help a lot of people.

Action 196 - Split synchronisation use case (#4)

<trackbot> Error finding '196'. You can review and register nicknames at <http://www.w3.org/2011/webtv/track/users>.


ddavis: There are now two use cases - one where the media content is identical and played on multiple devices, the other where the media is different (but related) and synced across multiple devices.

yosuke: Any comments?

Action 197 - Classify the accessibility requirements into general ones and specific ones.

<trackbot> Error finding '197'. You can review and register nicknames at <http://www.w3.org/2011/webtv/track/users>.

yosuke: This is mine. We had a talk with the HTML accessibility sub group during the AC meeting.
... Mark, Janina, Daniel, Kaz and I joined.
... We exchanged some thoughts about how we can work together.
... There are two actions we can take.
... One is improving new standards or specs, e.g. TV Control API, that are related to this Interest Group, adding new accessibility features.
... We can contribute to this.
... The other approach is that there are lots of external orgs using HTML5 and related standards.
... We can improve awareness for people who use W3C standards of what is best for accessibility.
... We could create guidelines for other SDOs to use or reference.
... Accessibility is regulated and raising awareness is important.

ddavis: There are already guidelines being worked on so we would not rewrite them - just add to them or somehow make it easier for TV-related organisations and companies to find and digest the existing guidelines.

MarkS: I can see the benefit for tailoring guidelines for a particular industry.
... We'd have to spend time learning the needs of that industry. We could collaborate on this.

janina_: I'm all for it if the group thinks it's worth doing.

ddavis: What is goal for you?

janina_: First goal is that all the material we have is fully supported. This material, e.g. captions which are enshrined in law, exist from before the web was born.
... Captions are already there and easy to use.
... We think the specs already support this.
... There may be cases where different devices display different things.
... Whatever the viewing setup, we want this work supported.
... The third goal is that people creating the user agents have guidance on how to use our resources.
... If we can have a example video that has every accessibility component that we suggest, that would be good.
... We've had some good response from education because they're under US law to provide accessible materials including videos.
... I think it's achievable.

Clarke: I'm doing some work on broadcast and cable side on descriptive video. I should sync up with your group at some point.

<MarkS> Clarke, ping me at mark@w3.org and I will send you the details

yosuke: We could talk about the video in the IG, but we need significant contributions.
... For now I'm not sure how many members can contribute.

MarkS: This part was a broader goal.

yosuke: One idea is that currently hybrid TV is an important topic. There are lots of different delivery methods but one single viewing experience.
... That requires different standards from different stakeholders.
... That's also a challenge because we need to compile several different things. There's no comprehensive study about how to maintain accessibility in such an environment.
... The media accessibility guidelines (to be published) could cover this.
... In this IG and other SDOs, people are creating new hybrid standards. If we can check what problems the hybrid systems have, that could help a lot of people.
... Any further comments?
... Let's firstly create some notes to clarify the goals of the clarification between the TV IG and the media accessibility sub group.
... We can refine this on the mailing list and then work together efficiently.

Action 198 - Ask timed text wg about 4k affecting captioning.

<trackbot> Error finding '198'. You can review and register nicknames at <http://www.w3.org/2011/webtv/track/users>.

ddavis: Sorry I haven't done this yet.

Action 199 - And yosuke to create questionnaire for 4k stakeholders about web standards issues.

<trackbot> Error finding '199'. You can review and register nicknames at <http://www.w3.org/2011/webtv/track/users>.

yosuke: I wrote an email to Daniel suggesting we start within W3C members, and after that ask externally.
... 4K is a compression technology so the work is also done in MPEG and other SDOs.
... We can later ask external SDOs about their use cases and concerns.
... What groups should we ask about 4K?
... We already talked about asking the Timed Text Working Group. Should we ask e.g. HTML WG and its media task force?

ddavis: I think so.

yosuke: Maybe CSS WG - they're dealing with positioning and responsive media.
... Maybe SVG WG?
... And Media Accessibility Sub Group
... Any others?

PaulHiggs: Is anyone looking at accessibility and subtitles on companion devices?
... I'm thinking of something like a teleprompter mode where a transcript scrolls on my tablet.

Mark_Vickers: [inaudible]

MarkS: I think you're saying current web tech already supports this functionality.
... We currently have the ability to sync video with a transcript.

yosuke: Our situation with synchronisation is we need to define better standards for this.
... For example, if the main video is delivered using broadcasting, there is almost no buffer.
... But we may need additional synchronisation method. That's also the hybrid situation.
... The standard for this is not defined in W3C.
... It's important for accessibility but we don't have anything for this level of synchronisation.

ddavis: So add it as a new use case?

yosuke: Yes, but we already have lots of existing second screen use cases so maybe we can improve what we already have.
... Any other comments?

ddavis: I think Janina has just sent a new use case.

janina_: I think there are a couple of unique things.
... The use case is e.g. someone growing older who loses their hearing.
... They lose their hearing so that viewing movies/tv content, they can't follow the dialogue because of additional sound (gunshot, car noises, etc.)
... They could rely on captions but ideally they'd like to follow the audio soundtrack.
... In the UK there's a thing called clean audio.
... The notion is an alternative audio track which is primarily dialogue.
... This is made available to the viewer.
... So for example, the viewer has an app on their mobile device which tells them when a clean audio track is available for them to listen to separately.
... This audio app is important because a hearing therapist can tune the app to the frequencies that are best for that particular viewer.
... This is a very important strategy.
... It could also be possible to use the same app in the cinema watching movies.

yosuke: I think NHK also has similar research.
... They're keen to develop accessible technologies.

janina_: I think we'd like to know about things like that. You could help us with footnotes and who in the industry is doing what.
... You could really help with us compiling the guidelines.

yosuke: The industry would also like to improve society and get more recognition.

janina_: I heard a US senator saying you can be in business and still do good - we should reward them.

ddavis: Should we add that use case to our wiki page?

PaulHiggs: To clarify, the clean audio is just the commentary. Is there some kind of mixing that could bring in some sound effects?

janina_: Yes, I think there is.
... The additional audio should not get in the way of the dialogue.

PaulHiggs: And the app can adjust the frequency for the hearing band?

janina_: Yes, exactly.

yosuke: Any other topics to discuss?

<scribe> ACTION: ddavis to add clean audio use case to wiki page [recorded in http://www.w3.org/2014/06/11-webtv-minutes.html#action03]

<trackbot> Created ACTION-201 - Add clean audio use case to wiki page [on Daniel Davis - due 2014-06-18].

yosuke: Some work is going on in the mailing list. Let's clarify our schedule.
... We'll can think of milestones - when we finalise things and do gap analysis.
... On the mailing list, IG members can make comments.

ddavis: Sounds good to me.

yosuke: Thank you for joining us. Let's talk in two weeks.

Meeting adjourned.

Summary of Action Items

[NEW] ACTION: 194 to Clarify use case 1 description [recorded in http://www.w3.org/2014/06/11-webtv-minutes.html#action01]
[NEW] ACTION: ddavis to add clean audio use case to wiki page [recorded in http://www.w3.org/2014/06/11-webtv-minutes.html#action03]
[NEW] ACTION: ddavis to split up use case 1 into watermarking and fingerprinting. [recorded in http://www.w3.org/2014/06/11-webtv-minutes.html#action02]
[End of minutes]