See also: IRC log
<trackbot> Date: 21 August 2012
<adrianba> ScribeNick: johnsim
No comments on minutes - completed
Three overdue action items which are further down in the agenda -
Baseline documents and Bugzilla information
David: as far as i know it is complete (EME), but since no review, marked as work as progress. and clear key
does not work at all until we complete action item
and to provide link to old draft to prevent confusion
<trackbot> ACTION-1 -- David Dorwin to and editors to implement proposal 2b by revising EME specification -- due 2012-08-17 -- CLOSED
Adrian: this is complete and awaiting review
<trackbot> ACTION-2 -- Mark Watson to propose a resolution to bug 17673 -- due 2012-08-17 -- OPEN
Adrian: this is about how to update the spec to support --- to describe how init data will work for ISOBFF
Mark is not on the call today because he is traveling. We have had several discussions about this bug, and we wanted to sort this out and resolve before resolving clear key, which is the next action
I updated this bug and assigned to Johnsim who has been putting together the proposal text for this bug and the clear key which is the next action
<adrianba> johnsim: there are two different sources of initdata in general, maybe 3 if you slice it finer
<adrianba> ... the media and the app
<adrianba> ... for example MPD might contain initdata for one key system
<adrianba> ... also in ISO BMFF there is a PSSH box which contains initdata for different key systems
<adrianba> ... especially if common encryption is used
<adrianba> ... so you can have multiple from both places
<adrianba> ... the initdata from the needkey may need to be combined with the initdata from the app
<adrianba> ... we have several ideas about how to do that
<adrianba> ... the other factor to be taken into consideration is if common encryption is being used which may have several initdata blocks that you may want to replace
<adrianba> ... for example you may want to replace the location of the key (discussed in MPEG DASH)
<adrianba> ... these are some of the scenarios that need to be accommodated
<adrianba> joesteele: you're saying the app could modify the initdata?
<adrianba> johnsim: let's say the needkey includes some initdata but the app knows it doesn't want to use that data for one system
<adrianba> ... but wants to replace it
<adrianba> ... our thinking is that you should be able to provide the app specific data prepended to indicate that this is the preferred data for this acquisition
<adrianba> ... we know the scenario occurs, not sure if this is the right solution
<adrianba> ddorwin: why prepend rather than just including that?
<adrianba> johnsim: that's one idea - this is how to say one PSSH box should be used and not another without having to crack open the initdata
<adrianba> ... it's possible to parse through the data but we'd prefer not to do that
adrian: you get a needkey event, because the media contains multiple pssh, passes the concatenated pssh boxes as the init data
meanwhile the app sees some init data used by one of the key systems
without parsing through each system id it is, you can use the app provided one prepended, so that you could prioritize the app one over the media data or the other way around.
<adrianba> action-2 due in 1 week
<trackbot> ACTION-2 Propose a resolution to bug 17673 due date now in 1 week
<trackbot> ACTION-3 -- Adrian Bateman to propose resolution to bug 17682 -- due 2012-08-17 -- OPEN
Clarify scenarios where you need to replace entire init data and when you need to update a subset of the init data
Adrian: clarifying the Clear Key, how Keys and Key IDs are associated
David: original 0.1 had an add key parameter, and a key id parameter
init data parameter was redundent with key request
removed init data paramters but now there is no pair of parameters for key id and key in clear key
we need mappings from key id to key
Joe Steele: what is the parameter
David: just key, which could be a license
Adrian: why wold this be, create
session wit the init data, with the resulting session object,
and pass in the key, two consequtive calls rather than
... you call "createsession" on the media keys object, and that gives you a media key session object, when you call you pass in the init data
when you get the media key session, you call add key and you pass in the key in the same way, but that restricts clear key to one key per session, which is different form other systems
<adrianba> johnsim: in key rotation, would that be one session or multiple sessions?
David: one key for ten minutes, i
think that should be separate sessions, playing something for a
day, don't want 500 keys sitting around
... each time you wanted a key, that would be a new session, and then you would destroy the old one at some point.
... need clarity on key rotation scenarios and make sure we support them .... it would be nice if the more common key rotation scenarios are supported by clear key
adrian: single key per session, second, we need to examine scenario where multiple keys are need (e.g. key rotation), if common need to work for single session
joesteele: multiple keys - one for audio and video - a non-rotation issue we have with clear key
markw: adaptive bitrates using different keys is another scenario.
adrian: the question are those a
single session or multiple session when we look at the media
key object, session id, single key negotiation withthe license
... diff keys audio/video, diff keys based on bitrates, multiple needkey events, and each would be a separate session,
Mark: anything is possible, CDM can match keys with lienses, keys needed for media, whether a new session is needed, we haven't you need a separate session for each key
Adrian: the current model assumes a single init data corressponds to a single session
there is no way to add more init data to a session
mark: it think that is okay. audio and video file with init data, application doesn't know if those init data contain information - might know how the content provider created the content
Adrian: propose text change for Clear Key case for single key per session, and maybe we can have more discussion whether that is sufficient once we have synced everyting back into the spec and we can review it
<adrianba> ACTION-3 due in 1 week
<trackbot> ACTION-3 Propose resolution to bug 17682 due date now in 1 week
Adrian: extending this action for a week as well
<trackbot> ACTION-4 -- David Dorwin to propose a solution for bug 17672 -- due 2012-08-17 -- OPEN
David: webm version of the ISO action #2. I have not made any progress on the text, though it is coming together (WebM for encryption). Request two weeks
<adrianba> ACTION-4 due in 2 weeks
<trackbot> ACTION-4 Propose a solution for bug 17672 due date now in 2 weeks
Adrian: last meeting, mentioned
need for more discussion, David sent this mail and Mark
responded, so question is
... what state is the spec compared to these two emails? do we currently describe what is mentioned in the two emails or work needed on the spec
David: the spec has not been
updated per the emails and object oriented design has big
impact. Removed all but the introductory text because
incompatible, so we need to go back and fill that in
... Still some more decisions to be made first
Adrian: is this something we should push for the agenda on the next meeting and have some discussion ahead of that?
Mark: i think so, now that we have the object oriented updates, some ideas in the email thread, but we need to work through that
Adrian: could you follow-up on your previous mail and suggest changes ?
<adrianba> ACTION: mark to follow-up on Key Release mail thread now that the OO changes have been made - discuss on next call [recorded in http://www.w3.org/2012/08/21-html-media-minutes.html#action01]
<trackbot> Created ACTION-5 - Follow-up on Key Release mail thread now that the OO changes have been made - discuss on next call [on Mark Watson - due 2012-08-28].
Adrian: given the discussion on
this call we still have some work to do for things on the top
of the list, make sense for editors to get together in advance
of the next call and triage the bugs
... make sure they are still valid, maybe some can be closed, maybe some are simple and we can just action them.
... we will try to follow-up on that before the next meeting
Joe Steele: understanding, every other week?
Adrian: alternating between EME
and MSE - trying to keep as predictable as possible.
... Chair and scribe for next meeting - Paul back - who will volunteer for scribe?
Joe Steele: will scribe next meeting
This is scribe.perl Revision: 1.136 of Date: 2011/05/12 12:01:43 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Found ScribeNick: johnsim Inferring Scribes: johnsim Default Present: Matt, adrianba, +1.425.202.aaaa, ddorwin, acolwell, +1.818.560.aabb, +1.206.861.aacc, +1.408.536.aadd, bwatts, +1.613.491.aaee, Suzie, Johnsim, MartinSoukup, Clarke, +44.303.040.aaff, joesteele, duncanr, +1.215.286.aagg, markw, +1.760.533.aahh, +1.858.485.aaii Present: Matt adrianba +1.425.202.aaaa ddorwin acolwell +1.818.560.aabb +1.206.861.aacc +1.408.536.aadd bwatts +1.613.491.aaee Suzie Johnsim MartinSoukup Clarke +44.303.040.aaff joesteele duncanr +1.215.286.aagg markw +1.760.533.aahh +1.858.485.aaii Agenda: http://lists.w3.org/Archives/Public/public-html-media/2012Aug/0034.html Found Date: 21 Aug 2012 Guessing minutes URL: http://www.w3.org/2012/08/21-html-media-minutes.html People with action items: mark WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]