13:57:10 <RRSAgent> RRSAgent has joined #me
13:57:14 <RRSAgent> logging to https://www.w3.org/2024/07/02-me-irc
13:57:14 <Zakim> Zakim has joined #me
13:57:23 <cpn> Meeting: Media & Entertainment IG meeting
13:58:22 <ohmata> ohmata has joined #me
13:58:57 <kaz> present+ Kaz_Ashimura, Alcia_Boya, Hisayuki_Ohmata
13:59:05 <kaz> agenda: https://lists.w3.org/Archives/Public/public-web-and-tv/2024Jun/0005.html
14:00:33 <kaz> present+ Hiroki_Endo, Chris_Needham
14:01:08 <alicia> alicia has joined #me
14:01:29 <HirokiEndo> HirokiEndo has joined #me
14:01:47 <tidoust> tidoust has joined #me
14:02:18 <piers> piers has joined #me
14:02:21 <cpn> scribe+ cpn
14:04:17 <cpn> present+ Tatsuya_Igarashi
14:05:07 <Igarashi> Igarashi has joined #me
14:05:07 <cpn> Slideset: https://www.w3.org/2011/webtv/wiki/images/0/03/2024-07-02-MEIG-Meeting.pdf
14:05:16 <kaz> q+
14:05:24 <Igarashi> present+
14:05:30 <kaz> q-
14:06:02 <kaz> present+ Francois_Daoust, Kinji_Matsumura, Piers_O'Hanlon, Ryo_Yasuoka
14:07:13 <kaz> topic: Agenda
14:07:21 <kaz> cpn: (goes through the agenda)
14:07:43 <kaz> ... CTA WAVE liaison, Media WG/HDR, IETF SCONEPRO, DASH-IF, TPAC
14:07:47 <kaz> ... any other topics?
14:07:53 <kaz> (none)
14:08:43 <kaz> topic: CTA WAVE liaison
14:08:53 <kaz> cpn: got a liaison letter from CTA WAVE
14:09:31 <kaz> -> https://lists.w3.org/Archives/Public/public-web-and-tv/2024Jun/0003.html liaison letter
14:09:46 <kaz> cpn: interoperability of media contents
14:10:09 <kaz> ... demo at media web symposium 2024 as well
14:10:17 <emilio> emilio has left #me
14:10:22 <kaz> ... if any views, would hear from you
14:10:41 <kaz> ... would say we're very interested and would invite CTA WAVE to present and discuss
14:10:50 <kaz> ... need to identify the date
14:10:56 <Igarashi> +1
14:11:07 <kaz> ... will negotiate a session to discuss the detail
14:11:29 <kaz> ... various links in the liaison letter
14:11:39 <kaz> ab: have to be something separate
14:11:58 <kaz> ... the goal is to make sure browser support for different contents
14:12:17 <kaz> ... could be web platform test by different browser vendors
14:12:40 <kaz> cpn: any specific points?
14:12:54 <kaz> ab: maybe resource intensive for resources
14:13:18 <kaz> ... may use automatic image validation
14:13:43 <kaz> cpn: the test set to be clarified too
14:13:57 <kaz> ... QR codes here and time codes
14:14:23 <kaz> ... image processing by the smartphone's video camera
14:14:46 <kaz> ab: it's an interesting approach definitely
14:14:59 <kaz> poh: The code and media for the CTA-WAVE DPC Test suite is here: https://cta-wave.github.io/dpc-test-suite/
14:15:10 <kaz> present+ Martin_Lasak
14:15:15 <kaz> rrsagent, make log public
14:15:27 <kaz> cpn: not sure how to test video element
14:15:36 <kaz> q?
14:15:42 <kaz> q+ martin
14:15:50 <kaz> ml: we hosted that demo
14:16:15 <kaz> ... right side QR code includes encoded information for time synchronization
14:16:43 <kaz> ... one thing important is video streaming's refresh rate
14:16:52 <kaz> present+ Dan_Druta
14:17:20 <kaz> ab: many videos allow to use 50fps
14:17:38 <kaz> ml: need to consider all the options, 50fps, 60fps and 120fps
14:17:45 <kaz> ab: flicker also
14:18:01 <kaz> ml: system measuring MSE
14:18:17 <kaz> ... measurement might be distracted by the device itself
14:18:29 <kaz> ab: often have manual components
14:18:45 <kaz> ... in the case of Web platform test
14:19:02 <kaz> ... automatically tested more or less based on browsers
14:19:20 <kaz> ml: tests vary what you want to test
14:19:42 <kaz> ... long-running tests lasts every after another
14:19:54 <kaz> ... a full-test costs much
14:20:22 <kaz> cpn: Martin, have you seen the discussion by CTA WAVE guys around future meeting?
14:20:29 <kaz> ml: need to talk to Louay
14:20:35 <kaz> cpn: ok
14:20:41 <kaz> ... would like to invite you two
14:20:52 <kaz> present+ Andre_Paul
14:21:18 <kaz> ... will follow up to invite you
14:21:31 <cpn> q?
14:23:13 <kaz> topic: Media WG update
14:23:20 <kaz> cpn: EME updates
14:23:33 <kaz> ... EVE v2 first public WD in progress
14:23:42 <kaz> ...publish HDCP version registry
14:23:57 <kaz> ... also publishing all EME/MSE registries to /TR
14:24:12 <kaz> ... may need maintenance updates
14:25:08 <kaz> -> https://datatracker.ietf.org/wg/sconepro/about/ SCONEPRO resource (Dan)
14:25:43 <kaz> cpn: Registries include:
14:25:53 <kaz> ... MSE Byte Stream Format Registry
14:26:00 <kaz> ... EME Initialization Data Registry
14:26:07 <kaz> ... EME Stream Format Registry
14:26:14 <kaz> ... then
14:26:31 <kaz> ... WebCodecs VideoFrame Metadata Registry
14:26:51 <kaz> ... human face detection metadata and background segmentation
14:27:03 <kaz> ... various topics MSE needs
14:27:12 <kaz> ... manage to media source
14:27:22 <kaz> ... it's ongoing working progress
14:27:33 <kaz> ... comments/questions?
14:27:36 <kaz> (none)
14:27:59 <kaz> topic: HDR update
14:28:21 <kaz> cpn: PNG WG working on PNG v3 with HDR support
14:28:23 <kaz> ... at CR
14:28:39 <kaz> ... plan is finalize it as a REC
14:28:54 <kaz> ... then
14:29:07 <kaz> ... Color on the Web CG next meeting on July 9
14:29:21 <kaz> ... TAG feed back on current HDR Canvas proposal
14:29:56 <kaz> -> https://github.com/w3c/ColorWeb-CG/blob/main/hdr_html_canvas_element.md HDR Canvas proposal
14:30:12 <kaz> cpn: anything you're aware?
14:30:30 <kaz> (nothing specific)
14:30:38 <kaz> topic: IETF SCONEPRO
14:30:52 <kaz> -> https://datatracker.ietf.org/wg/sconepro/about/ SCONEPRO resource (Dan)
14:31:00 <kaz> cpn: would invite Dan to talk about that
14:31:12 <kaz> dd: tx for summarizing the slides
14:31:40 <kaz> ... SCONEPRO stands for “Secure Communication of Network Properties”
14:32:05 <kaz> ... the actual scope is a lot narrower than the title
14:32:18 <kaz> [[
14:32:19 <kaz> Goal: In adaptive-bitrate video streaming, enable the network can signal to the client the bitrate it
14:32:19 <kaz> should be using
14:32:19 <kaz> ○ Maintain throughput at acceptable video quality
14:32:19 <kaz> ○ Apply network operator policies and network conditions
14:32:21 <kaz> ]]
14:32:44 <kaz> dd: particularly important for mobile careers
14:32:57 <kaz> ... video becoming dominant for media
14:33:59 <kaz> ... not always easy to detect specific video on time
14:34:24 <kaz> ... due to network operator polices and network conditions
14:35:21 <kaz> ... what SCONEPRO really aims is application providers detect the situation
14:35:31 <kaz> ... majority of the video traffic
14:35:42 <kaz> ... at least large video providers and social media
14:35:53 <kaz> ... that's how it works and starts
14:36:07 <kaz> ... early stage discussion at IETF in March 2020
14:36:28 <kaz> ... now July in Vancouver BoF meeting
14:36:42 <kaz> ... it is a one-way communication for now
14:37:11 <kaz> ... why is this relevant to this group is
14:37:24 <kaz> ... as you know, QUIC is app-layer protocol
14:37:41 <kaz> ... implementations on both the server side and the client side
14:38:16 <kaz> ... it is most likely protocol-side activity
14:38:28 <kaz> ... how that would be materialized?
14:38:43 <kaz> ... need clear requirements
14:38:53 <kaz> ... corresponding to a WG
14:39:10 <kaz> ... the agenda of July meeting is not published yet
14:39:19 <kaz> ... but hopefully available soon
14:39:29 <kaz> cpn: tx for your information
14:39:40 <kaz> ... any thoughts/questions?
14:39:51 <kaz> ab: net neutrality?
14:39:56 <kaz> dd: good question
14:40:12 <kaz> ... Meta has just provided a document about neutrality
14:40:19 <kaz> ... one application vs another
14:40:30 <kaz> ... not discriminating the others
14:40:48 <kaz> ab: Web apps vs executables
14:40:54 <piers> Draft on net neutrality: https://datatracker.ietf.org/doc/draft-tan-sconepro-netneutrality/
14:40:56 <kaz> dd: all the applications to be considered
14:41:07 <kaz> ... would recommend you go thought the draft above
14:41:29 <kaz> ab: from my viewpoint, it's introducing dangerous mechanism...
14:41:40 <kaz> ... there is a big difference in general
14:42:13 <kaz> dd: just want to point you to the draft above
14:42:28 <kaz> ... actually written by the Facbook people
14:42:48 <kaz> ... some privacy issues associated with the subscriber's side
14:43:36 <kaz> ab: part of the QUIC issue is initiated by ISP
14:43:55 <kaz> dd: ISP eventually apply some policy
14:44:07 <kaz> ... it is not something squeezing money
14:44:30 <kaz> ... users should be aware since it's part of contract
14:44:45 <kaz> ... anyway
14:45:02 <kaz> ... just wanted to point you to the draft on net neutrality above for today
14:45:11 <kaz> ... further discussion could be done at TPAC
14:45:44 <kaz> cpn: wanted to talk about ManagedMediasSource API also
14:46:01 <kaz> -> https://github.com/w3c/media-source/issues/322 media-source Issue 322 - Proposal: Add quality attribute to ManagedMediaSource API
14:46:42 <kaz> cpn: it may be satisfies your use case?
14:46:50 <kaz> s/may be/maybe/
14:47:08 <kaz> dd: hopefully clarity about the protocol
14:47:19 <kaz> ... will just read it
14:47:36 <kaz> cpn: tx for joining and introduce the topic around SCONEPRO
14:48:12 <kaz> dd: most of the QUIC participants/contributors understand particular use cases
14:48:26 <kaz> ... application would convey some congestion
14:48:41 <kaz> ... interested in a re-prioritization
14:48:55 <kaz> ... what kind of prose is less important or so
14:49:11 <kaz> cpn: tx
14:49:38 <kaz> poh: presentation from YouTube also
14:49:51 <kaz> dd: it's API-based
14:50:03 <kaz> ... some of the benefit is very thoughtful
14:50:50 <kaz> poh: preferred quality to be specified by the quality attribute
14:51:34 <kaz> cpn: tx all for the discussion
14:51:53 <kaz> topic: DASH-IF special session on Content Authenticity
14:51:54 <kaz> and Provenance in DASH
14:52:05 <kaz> cpn: asked to ask you about your interest
14:52:32 <kaz> ... speaking rather as a participants from BBC
14:53:54 <kaz> ... (including C2PA technology)
14:54:06 <kaz> [slide 10]
14:54:48 <kaz> cpn: content credentials menu for deeper insight (on the left side of the UI)
14:55:23 <kaz> ... content credential icon (for valid and invalid content at the bottom left)
14:55:36 <kaz> [slide 9]
14:55:55 <kaz> s/10/7/
14:55:58 <kaz> s/9/8/
14:56:05 <kaz> cpn: (goes back to slide 7)
14:56:31 <kaz> ... discussion with people from Adobe presented this at June 7 DASH-IF meeting
14:56:59 <kaz> ... about C2PA, Dreato Assertions WG, Investigations with DASH.js and Browser integration via HTML <video> and MSE
14:57:27 <kaz> ... possible future meeting, but not planned yet
14:57:55 <kaz> ab: general question about content provenance...
14:58:11 <kaz> ... have some concern
14:58:53 <kaz> cpn: ok. interesting...
14:59:05 <kaz> ... if you have seen it, would suggest the recorded video too
14:59:32 <kaz> poh: one idea for authentication and provenance
14:59:39 <kaz> s/one/it's one/
14:59:45 <kaz> ab: wondering who is signing
15:00:27 <kaz> ... if I were the authentication platform provider, there should be some incentive
15:00:50 <kaz> topic: TPAC 2024
15:00:51 <kaz> [slide 9]
15:00:59 <kaz> cpn: we're getting out of time
15:01:02 <kaz> ... so just quickly
15:01:10 <kaz> ... open question to you is
15:01:25 <kaz> ... if you join the meeting and if you have any topics in your mind
15:01:30 <kaz> ... please let me know
15:01:46 <kaz> ... some discussion around next-generation audio, etc.
15:02:00 <kaz> ... it's just an open invitation to you all
15:02:08 <kaz> ... happy to hear that
15:02:27 <kaz> ... that's all for today
15:02:31 <kaz> topic: AOB
15:02:41 <kaz> cpn: anything else for today?
15:02:42 <kaz> (none)
15:02:53 <kaz> cpn: tx for all your discussions!
15:02:57 <kaz> ... next meeting will be held in August
15:03:01 <kaz> [adjourned]
15:03:05 <kaz> rrsagent, make log public
15:03:09 <kaz> rrsagent, draft minutes
15:03:10 <RRSAgent> I have made the request to generate https://www.w3.org/2024/07/02-me-minutes.html kaz
15:03:57 <alicia> Chris, it was you that said that were involved in C2A in the context of the BBC, right?
15:05:03 <alicia> I see
15:06:01 <alicia> another thing I'm wondering a lot is how they will actually handle editing. I know it is part of Adobe C2A proposal by design, but I really wonder about the specifics, since that's not a trivial thing at all.
15:07:43 <alicia> afaik they want information like what edits have been done to be embedded in the manifests, so if e.g. the BBC shows some footage from I don't know... Reuters, in a picture in picture, the frame from Reuters can be validated as such
15:08:24 <cpn> *me So I think technically it is possible to do that, but there's still an editorial decision as to whether to include that information
15:08:50 <alicia> but in the real world people doing video editing have a lot more freedom than doing some PiP, there is color grading, there is croping, there is noise reduction, there are effects...
15:09:24 <alicia> so making a video editor that actually produces those manifests could get quite tricky
15:09:54 <alicia> especially as a 3rd party vendor who hasn't been involved at close length with the spec
15:10:48 <alicia> and that is before you consider things like patents, that a few companies (let's say Adobe) could use to make themselves a de facto monopoly on c2a content video editing
15:11:38 <alicia> that is good to know, and then thankfully one concern I can cross out of the list
15:12:38 <alicia> technically I can imagine Adobe being technically capable of encoding this kind of metadata from e.g. Premiere, since I've seen what they do with XMP metadata in Lightroom and it's almost scary.
15:13:57 <alicia> (all your edits are stored inside the JPEGs you export for everyone to see unless you strip them! along with what looks like serial numbers of your copy of lightroom or similar)
15:15:53 <alicia> Privacy is another thing that concerns me a lot about C2A... I think a system like C2A makes a lot of sense for the BBC, but I would dread a world in which C2A is expected or feared to become mandatory in any way for consumer video
15:18:39 <alicia> In my head, C2A makes sense in a similar niche as software digital signature. Something most people would never need to interact with further than clicking a few prompts in their OS occasionally but that helps diagnose and prevent authenticity problems.
15:20:19 <alicia> That "is this video real or is it AI generated" use case is truly a dystopian one, and if Adobe wants to push for it I hope it fails.
15:22:37 <alicia> not if the price to pay is mass surveilance
15:23:10 <alicia> which is why I say C2A makes sense for certain niche applications
15:24:57 <alicia> for instance: if a certain press agency starts C2A signing all their content and makes that well known, it is very easy to prove any manipulation later done to content sourced from them, because you know the signed content exists and you can compare it to the unsigned content.
15:29:37 <alicia> and there would definitely be value on having an automated tool for that kind of verification!
15:30:57 <alicia> and there isn't any privacy concerns with that particular use case because in that use case, the agency publishing the content has no need for anonimity or pseudonimity, only their sources do.
15:32:22 <alicia> on the other hand, imagine a world where C2A is somehow made the norm and every video ever recorded can be traced back to a specific camera bought by a specific person. Ironically, even those same news agencies now would suffer rather than benefitting from C2A, since they couldn't e.g. publish an interview with an anonymous source without revealing
15:32:22 <alicia> it.
15:32:30 <alicia> on the other hand, imagine a world where C2A is somehow made the norm and every video ever recorded can be traced back to a specific camera bought by a specific person. Ironically, even those same news agencies now would suffer rather than benefitting from C2A, since they couldn't e.g. publish an interview with an anonymous source without revealing
15:32:30 <alicia> their sources.
15:34:32 <alicia> the latter concern is mostly political in nature
15:35:46 <alicia> as in, as much as C2A can be useful in some cases, I would fight any push to make it expected for general content and any politician claiming enforcing it be worth it for fighting AI disinformation or whatever
15:37:34 <alicia> in their proposal Adobe included the cameras signing the content they record, so I really worry they want anything like that
15:38:39 <alicia> almost as if we cannot even trust the person signing the content they produced, but instead the camera vendors have to get involved to prove they weren't manipulating it...
16:36:52 <Zakim> Zakim has left #me