(MINUTES) HTML WG media telcon 2012-09-25 - MSE

HTML Media Task Force Teleconference - Media Source Extensions (MSE)

25 Sept 2012



The minutes are available in HTML here: http://www.w3.org/2012/09/25-html-media-minutes.html



or in plain text, below.



------------------------------------------------------------------------------
- DRAFT -

HTML Media Task Force Teleconference

25 Sep 2012

Agenda

See also: IRC log

Attendees

Present markw, +1.310.210.aaaa, pal, Matt, Clarke, +1.858.735.aabb, kstreeter, +1.206.218.aacc, ddorwin, paulc, johnsim, adrianba, acolwell, +1.650.214.aadd, BobLund, MartinSoukup, +1.617.253.aaee, strobe
Regrets
Chair SV_MEETING_CHAIR
Scribe johnsim
Contents
*Topics1.rollcall
2.minutes
3.action items
4.base line document
5.outstanding bugs
6.actions from the previous meeting

*Summary of Action Items

--------------------------------------------------------------------------------




<trackbot> Date: 25 September 2012

<paulc> Having trouble getting thru to Zakim

<paulc> Looking for a scribe for today

<scribe> scribe: johnsim

<paulc_> reconnecting

<paulc_> dropped first from zakim and then from the call

<paulc_> rejoiing

rollcall DONE

rollcall

minutes

<paulc_> http://www.w3.org/2012/08/28-html-media-minutes.html

noted

action items

none for media source extension

base line document

<paulc_> http://dvcs.w3.org/hg/html-media/raw-file/tip/media-source/media-source.html

<paulc_> Outstanding bugs: http://tinyurl.com/6pdnzej

outstanding bugs

actions from the previous meeting

bug 18400 -

<paulc_> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18400

Adrian working on this one and still pending

<paulc_> Bug 17000: Define a capability detection mechanism

<paulc_> https://www.w3.org/Bugs/Public/show_bug.cgi?id=17000

John and Adrian to discuss this item

17002 - two action items

<paulc_> http://lists.w3.org/Archives/Public/public-html-media/2012Sep/0025.html

series of questions and six different questions - long thread with Sylvia

Paul: need to get some input from Sylvia

aaron: filed bug against the HTML spec, and not sure what to do with the media fragment issue

<paulc_> HTML5 bug: https://www.w3.org/Bugs/Public/show_bug.cgi?id=18971

Aaron: i believe that is what unblocks the media source issue

<paulc_> What is status now of 17002: ) Bug 17002: Specify SourceID to AudioTrack.id & VideoTrack.id mappings

<paulc_> 6. Should I change the proposed fix to Bug 17002 to the following so the

<paulc_> solution is independent of the id field?

Adrian: this is an area where we are just starting to do the research - my preference without more info is to keep the string - more intuitive than the way the other methods work

Question is still outstanding and we will have to track this

bug 17006

<paulc_> Bug 17006: <track> Setting track language & kind when the information is in a manifest

email thread started by Aaron - some response - summarize where we are?

<paulc_> http://lists.w3.org/Archives/Public/public-html-media/2012Sep/0024.html

<paulc_> see Bob's response in http://lists.w3.org/Archives/Public/public-html-media/2012Sep/0027.html

Aaron: this one, i think i have plenty of information, still need to sort out what happens when language info is not in the track metadata
... media source not constrained to DASH, so have to handle case where language information is not present

Aaron will continue to do research on 17006

<paulc_> Bug 17094: Define segment formats for MPEG2-TS

Bob Lund sent email on this item

<paulc_> Bob's message: http://lists.w3.org/Archives/Public/public-html-media/2012Sep/0013.html

<paulc_> Aaron asked 8 questions in http://lists.w3.org/Archives/Public/public-html-media/2012Sep/0021.html

Aaron: i think this one is getting to the point where there needs to be a pretty technical discussion, transport stream spec has resulted in lots of questions, TS is very different than the other two formats we have

<adrianba> q

<adrianba> q

<adrianba> ?

bob lund: responded to your questions

<paulc_> See http://lists.w3.org/Archives/Public/public-html-media/2012Sep/0041.html for more answers

adrian: question on goal of MPEG2 TS - Aaron made the comment, adding MPEG2 TS support it should enable consuming HLS content. is that the goal?

Bob: HLS but also DASH profiles that use MPEG2 TS

yes

Mark: HLS support, assure that HLS compliant content can be playable in MSE, or second goal is that you can encode for MSE
... believe the second goal is what we are targeting

<markw> second goal is that you can encode HLS content that also works with MSE

Kevin: segment encryption scheme alignment with DASH content - IVs and Keys set in the DASH manifest - with signaling

<paulc_> Bug 16998: Change sourceAppend() to take a URL and optional range parameters

Bug 16998

Aaron: created 18962 & 18963

<paulc_> ) [Bug 18962] New: Allow appending with XHR

<paulc_> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18962

Aaron: created to satisfy the action item - split it into these two bugs - to gauge a sense of priorities of these two new features and to kick off work on them if considered high enough priority

<strobe> matt: heh, actually i just misspelled my own name

<Zakim> adrianba, you wanted to talk about appending with XHR

Aaron: using XHR and interfacing it to media source, using default XHR beaviors

Adrian: mixed feelings about using XHR here. XHR interface does include all of the things - generic HTTP requests - but can't quite see how to map the data from the request and feeding it into the media engine without going through javascript

Mark: i think XHR approach is the right one, but i know if this is a high priority, just an optimization for constrained platforms, you can get a blob from XHR, without data being copied into javascript stack
... i think we can do is to propose to the web apps working group

Adrian: Mark has reminded me implicitly that we do have a proposal from MS for stream API - that might be a way of achieving this - for returning a stream object from XHR

<paulc__> Paul is back

<adrianba> http://lists.w3.org/Archives/Public/public-webapps/2011OctDec/1494.html

<adrianba> http://dvcs.w3.org/hg/streams-api/raw-file/tip/Overview.htm

<strobe> agreed, not too urgent from what we've seen

<paulc__> [Bug 18963] New: Provide a mechanism for rate limiting appending

<paulc__> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18963

Aaron: provide a way for UA to signal when it is getting overflowed with data
... when it reaches a place where it will have to remove media from the source buffer, runningout of memory, signal this so the application can back off appending
... then get a signal to say "okay to append more data" . need to get a sense of what people think of this and relative priority of the work

Strobe: one thing we have seen, have a fixed discovery mechanism, and notification that either eviction is about to happen or has happened.
... being able to figure this out in before hand would be helpful in defining a buffer strategy

<Zakim> adrianba, you wanted to talk about general limiting signals

Adrian: we haven't gotten too far into thinking about this specific question, which is potential or actual size of the append buffer - just want to highlight a more general question which has come up at MS
... may be signals from UA to application that indicate that a different rate of appending needs to be done or a different bitrate is desirable -
... ex) network working great but device is not capable of doing processing and rendering, you might give the user a worse experience than if you dropped to a lower bitrate
... there were some bugs in HTML5 postponed in this area
... similar feature area - ways UA needs to signal that a different bitrate might be desirable - buffers getting full or quality not good

<paulc__> [Bug 18920] New: addSourceBuffer parameter type should be optional

<paulc__> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18920

<MartinSoukup> for buffer management would threshold events be helpful to allow the appender to understand back-pressure?

Aaron: main reason to suggest this one, type parameter - a little back and forth whether UA should detect the byte stream format and reject if not the type specified

strobe: in favor of failing early - helpful to get that up front

Aaron: mhy preference is to say what type you are going to append and then get an error if it is not correct

adrian: believe content type should be specified

<paulc__> [Bug 18709] Add SourceBuffer.remove() method

<paulc__> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18709

Aaron: just wanted to make sure - changed "flush" to "remove" and changed parameters - wanted some buyoff from the group before adding it

<markw> +1

<adrianba> +1

<strobe> lgtm

<strobe> +1

<strobe> erm, conventions

<markw> ship it

<paulc__> [Bug 18601] Contradictory requirements for initialization segments

<paulc__> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18601

Aaron: this bug is the case where part of the spec says - init segments are optional - and other parts which absolutely rely on the first init segment -
... wanted to discuss why the original language was there to make it optional - since formats in specs use them - or should we remove the language around being optional
... should init segments still be optional?

Kevin: there are formats that can be self initialized.
... I think having that flexibility might be valuable. one thing to consider doing is rather than evaluate all cases where required, treat self-initialized segments as initialization segments WITH each segment that is appended

Bob: I think we may want to wait until we decide what we will do about MPEG2 TS - that format may be unusual - init must be replicated in media segments
... might want init segment to be optional in that case

Aaron: the one thing i wanted to point out, init segment in dash context, what we call MSE Init Segment followed by what we call media segment in MSE
... encourage people to add comments to this bug

other items on agenda - ran out of time

Aaron: another update to the specification before the next meeting

Scribe for next call not determined

Summary of Action Items
[End of minutes]

--------------------------------------------------------------------------------

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2012/09/25 16:03:43 $


--------------------------------------------------------------------------------

Scribe.perl diagnostic output
[Delete this section before finalizing the minutes.]
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)

Succeeded: s/q//
Found Scribe: johnsim
Inferring ScribeNick: johnsim
Default Present: markw, +1.310.210.aaaa, pal, Matt, Clarke, +1.858.735.aabb, kstreeter, +1.206.218.aacc, ddorwin, paulc, johnsim, adrianba, acolwell, +1.650.214.aadd, BobLund, MartinSoukup, +1.617.253.aaee, strobe
Present: markw +1.310.210.aaaa pal Matt Clarke +1.858.735.aabb kstreeter +1.206.218.aacc ddorwin paulc johnsim adrianba acolwell +1.650.214.aadd BobLund MartinSoukup +1.617.253.aaee strobe
Agenda: http://lists.w3.org/Archives/Public/public-html-media/2012Sep/0033.html

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Found Date: 25 Sep 2012
Guessing minutes URL: http://www.w3.org/2012/09/25-html-media-minutes.html
People with action items:

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.


[End of scribe.perl diagnostic output]

________________________________
John C. Simmons | Media Platform Architect | Microsoft Corporation | direct 425-707-2911  | mobile 425-269-5759