[04:47:00] alex-meetecho joins the room
[04:50:02] Barry Leiba joins the room
[04:53:50] <Barry Leiba> I'm in the "avtcore" room with my jabber client, but it doesn't seem to connect to the meetecho chat room.
[04:54:39] <Meetecho> We notived, investigating
[04:54:40] Jonathan Lennox joins the room
[04:54:42] <Meetecho> *noticed
[04:56:07] Matthew Jarosz joins the room
[04:56:25] Magnus Westerlund joins the room
[04:56:41] Erik Wilde joins the room
[04:56:46] Mike English joins the room
[04:57:01] Martin Thomson joins the room
[04:57:33] csp joins the room
[04:57:54] Brian Rosen joins the room
[04:58:05] Alessandro Amirante joins the room
[04:58:09] mt joins the room
[04:58:18] Erik Wilde leaves the room
[04:58:18] <Alessandro Amirante> test
[04:58:18] Youenn Fablet joins the room
[04:58:25] <mt> It's working via Matrix
[04:58:27] Erik Wilde joins the room
[04:58:50] Magnus Westerlund too joins the room
[04:58:56] <Meetecho> Fixed, reloading the page will fix the chat for Meetecho users
[04:59:20] Mike English leaves the room
[04:59:25] Mike English joins the room
[04:59:37] Shuai Zhao joins the room
[04:59:38] Cullen Jennings joins the room
[04:59:39] Barry Leiba_179 joins the room
[04:59:40] Jonathan Lennox_863 joins the room
[04:59:48] <Cullen Jennings> test
[04:59:50] <Barry Leiba_179> chat...
[04:59:54] <Shuai Zhao> testing...
[04:59:57] <Jonathan Lennox_863> I see Cullen's test
[04:59:58] <Brian Rosen> works
[05:00:02] <Cullen Jennings> Looks like it is working
[05:00:11] Ted Hardie joins the room
[05:00:18] Bernard Aboba joins the room
[05:00:19] Jonathan Lennox leaves the room
[05:00:22] Stephan Wenger joins the room
[05:00:22] <Barry Leiba> Thanks, Meetecho folks.
[05:00:26] Martin Pettersson joins the room
[05:00:36] <Stephan Wenger> chat chat chat
[05:00:38] <Meetecho> Barry Leiba: sure, apologies for the issue!
[05:00:50] Gonzalo Salgueiro joins the room
[05:00:58] Gunnar Hellstrom joins the room
[05:01:31] <Brian Rosen> I will jabber scribe, if you want me to relay to the room, preface with "mic:"
[05:01:40] Harald Alvestrand joins the room
[05:01:48] englishm joins the room
[05:02:07] <Cullen Jennings> https://codimd.ietf.org/notes-ietf-109-avtcore?edit
[05:02:28] <Cullen Jennings> That is link to tool if you want to help
[05:02:30] Mike Coleman joins the room
[05:02:43] Roni Even joins the room
[05:03:12] Justin Uberti joins the room
[05:03:23] Sergio Garcia Murillo joins the room
[05:03:50] Timothy Panton joins the room
[05:03:59] <Martin Thomson> I am muted.
[05:04:11] <Martin Thomson> Trying to debug meetecho ICE issues.
[05:04:38] Erik Wilde leaves the room
[05:04:41] Erik Wilde joins the room
[05:05:20] james welch joins the room
[05:05:51] Luke Curley joins the room
[05:06:43] Roni Even leaves the room
[05:07:13] Roni Even joins the room
[05:08:05] <Shuai Zhao> i dont see an option at Meetcho to upload a profile picture :)
[05:08:30] <Meetecho> Shuai Zhao: if you set one in your datatracker account, we use that
[05:08:44] <Meetecho> Otherwise we fallback to gravatar, if you associated a picture to your mail address
[05:09:13] Daniel Havey joins the room
[05:09:27] <Martin Thomson> Chairs probably need to bump people out of queue when they give them the floor.
[05:09:32] James Gruessing joins the room
[05:09:46] Yago Sanchez joins the room
[05:10:18] Magnus Westerlund leaves the room
[05:10:32] Magnus Westerlund joins the room
[05:11:23] roni_even joins the room
[05:11:24] Magnus Westerlund leaves the room
[05:11:33] Magnus Westerlund joins the room
[05:11:43] <Shuai Zhao> will do that for EVC
[05:11:55] Magnus Westerlund leaves the room
[05:12:14] Erik Wilde leaves the room
[05:12:23] Erik Wilde joins the room
[05:13:12] Magnus Westerlund joins the room
[05:13:13] Mihail Yanev joins the room
[05:13:13] Mihail Yanev leaves the room
[05:13:34] Paul Kyzivat joins the room
[05:13:57] Erik Wilde leaves the room
[05:14:01] <Cullen Jennings> not sure but I think if you upload a picture to your ietf datatracker profile, you might get it
[05:14:10] Mihail Yanev joins the room
[05:14:24] Mihail Yanev leaves the room
[05:14:37] <Shuai Zhao> @Cullen, but there is no option under my IETF account to even upload one..maybe I am looking at the wrong place..
[05:16:30] <alex-meetecho> Shuai Zhao: you shuold email the secretariat to have you picture associated to your datatracker account
[05:16:58] <Shuai Zhao> thats probably the only way to do it. Thanks @alex
[05:17:39] Roni Even leaves the room
[05:17:42] Roni Even joins the room
[05:17:45] Roni Even leaves the room
[05:17:54] <alex-meetecho> yes that's the only way AFAIK
[05:17:54] Roni Even joins the room
[05:20:09] <Magnus Westerlund too> For people information C238 is down to 8 documents that needs approval from the authors. One of them I am the last one. There are some more people in this meeting that has work to do, but I will not name names.
[05:20:18] <Daniel Havey> We have no jabber room?
[05:20:30] <James Gruessing> This chat is the jabber room.
[05:20:39] <Ted Hardie> Name names!
[05:20:51] <Brian Rosen> it should be the same as avtcore jabber room
[05:21:02] <englishm> Yes, it is linked on this side
[05:21:54] roni_even leaves the room
[05:22:10] Lucas Pardue joins the room
[05:22:28] <Magnus Westerlund too> Ok Ted: Justin, MT, Cullen have work to do.  
[05:22:56] <mt> I don't have work to do, it's ekr
[05:23:39] <Magnus Westerlund too> Ok, I didn't check if people have approved their part, but you are a co-author on a non-approved document.
[05:24:47] Mo Zanaty joins the room
[05:24:48] Magnus Westerlund leaves the room
[05:24:58] <Ted Hardie> Justin was active in the github repos very recently, so that's in process.  Is Cullen also for JSEP?
[05:25:04] <Justin Uberti> We are still working through the github pilot process for JSEP
[05:25:06] <Daniel Havey> How will this lose 3 work with new packet recovery methods like RACK?
[05:25:16] Magnus Westerlund joins the room
[05:26:13] Magnus Westerlund leaves the room
[05:26:16] Magnus Westerlund joins the room
[05:26:18] Waqar Zia joins the room
[05:27:06] Magnus Westerlund leaves the room
[05:27:20] Magnus Westerlund joins the room
[05:27:46] <Jonathan Lennox_863> Apologies for missing you in the queue Daniel, I had the participant list scrolled down
[05:27:54] Timothy Panton leaves the room
[05:27:55] <Cullen Jennings> I'm sure Auth 48 is top of mind for Justin, EKR, and I
[05:28:04] Timothy Panton joins the room
[05:28:07] Magnus Westerlund leaves the room
[05:28:39] Magnus Westerlund joins the room
[05:28:48] Magnus Westerlund leaves the room
[05:28:51] Magnus Westerlund joins the room
[05:28:51] Timothy Panton leaves the room
[05:28:57] Timothy Panton joins the room
[05:29:02] Magnus Westerlund leaves the room
[05:29:05] Magnus Westerlund joins the room
[05:29:44] Timothy Panton leaves the room
[05:29:52] Magnus Westerlund leaves the room
[05:30:20] <Cullen Jennings> And thanks to Justin for making lots of progress
[05:30:25] Timothy Panton joins the room
[05:30:31] Timothy Panton leaves the room
[05:30:34] Timothy Panton joins the room
[05:30:40] Magnus Westerlund too leaves the room: Connection failed: connection closed
[05:32:14] <Daniel Havey> No worries Jonathan.
[05:33:13] Magnus Westerlund joins the room
[05:34:21] Magnus Westerlund too joins the room
[05:34:36] tbr joins the room
[05:35:44] tbr is now known as Tim Bruylants
[05:36:43] <Magnus Westerlund too> MT you have not approved UKS for publication.
[05:37:17] <mt> I did several times already.  I will do so again.
[05:37:21] <Daniel Havey> I'm 100% in support of encrypting transport headers.
[05:37:33] <mt> It would be fine if it weren't for the fact that my co-author was so recalcitrant.
[05:37:40] <Daniel Havey> Guards against ossification.
[05:37:44] Praneet joins the room
[05:40:51] Hiroyuki Goto joins the room
[05:40:55] <Harald Alvestrand> Can we lose 1-byte header extensions while we're at it?
[05:42:26] <Mo Zanaty> Will encrypted CSRCs cause false frequent SSRC space collisions by polluting the space very quickly?
[05:42:50] <Martin Thomson> Mo, are you assuming that someone will not remove encryption?
[05:43:31] <James Gruessing> Or can't because they are a middlebox?
[05:43:33] <Magnus Westerlund too> @MT: Regarding AUTH48 approval, I have also noted on several occurrances that you need to be very explicit that you approve publication or RFC-editor will not interpret it correctly. We will soon be down to only your co-author not having approved their documents.
[05:44:13] <Mo Zanaty> So all middleboxes must be in on this, right?
[05:49:24] Colin Perkins joins the room
[05:52:52] <Martin Thomson> Left to Right?
[05:53:34] <Shuai Zhao> normally, the right side has the latest change from the diff tool
[05:53:48] <Martin Thomson> L -&gt; R was what I was referring to.
[05:54:10] <Brian Rosen> "Lost" to "Received" I think
[05:54:33] <Brian Rosen> but Left to Right is hilarious
[05:55:40] <Martin Thomson> Imagine if we used different words for Boolean values: sinister and dexter would be my choice.
[05:55:51] <Brian Rosen> love it
[05:58:53] Hiroyuki Goto leaves the room
[06:00:01] <Harald Alvestrand> (clap clap)
[06:01:05] csp leaves the room
[06:02:22] <Martin Thomson> It's coaxial.
[06:02:59] <Brian Rosen> That;s because it's shielded
[06:07:22] <Cullen Jennings> agree with what magnus said on this
[06:08:33] <Martin Thomson> magnus: your microphone is saturating a little and distorting
[06:08:59] <Magnus Westerlund too> I will adjust down the level even further.
[06:09:13] <Martin Thomson> So many levels to adjust, I'm sure.
[06:10:02] <Magnus Westerlund too> Can Justin mute?
[06:10:20] <James Gruessing> Someone's making me hungry..
[06:10:22] <Daniel Havey> Somebody is munching a little too close to the mike ;).
[06:11:20] Shuai Zhao leaves the room
[06:12:02] Shuai Zhao joins the room
[06:13:55] Qi Pan joins the room
[06:15:40] Hiroyuki Goto joins the room
[06:18:34] <Justin Uberti> lol sorry folks had a bowl of :pretzel:s
[06:22:43] Jiao Kang joins the room
[06:22:48] Jiao Kang leaves the room
[06:24:57] <Martin Thomson> Time check: we have 35 min left and 45 min of scheduled time.
[06:25:19] <Jonathan Lennox_863> Yes.  One presentation and presenter isn't here, so that should be fine, but let's try to move quickly.
[06:27:32] <Martin Thomson> I've only just started looking into sframe (I only recently came on as chair), and I still don't understand the conceptual model.
[06:28:24] <Jonathan Lennox_863> I've closed the queue
[06:28:42] Hiroyuki Goto leaves the room
[06:29:03] <Roni Even> this reminds me a usage when you want to have a propriety encryption between the encode and decoder as part of the encoding. was done in circuit switch for miltary use
[06:29:05] <Justin Uberti> mt: imagine the encoder is also an encryptor, and then the output of the encryptor is packetized.
[06:29:09] <Martin Thomson> We might need the help of this group.
[06:29:37] <Martin Thomson> Justin: I understand those words, but I'm looking for the next level of specificity.
[06:30:26] <Justin Uberti> the cleartext metadata is hoisted into something the SFU can access, and then the rest is simply split up into units smaller than a MTU.
[06:30:53] Yago Sanchez leaves the room
[06:30:55] Colin Perkins leaves the room
[06:31:09] <Martin Thomson> Is that the usual approach for packetization?  Split up the bytes so that they can fit into an MTU.
[06:31:36] Yago Sanchez joins the room
[06:31:44] <Justin Uberti> for VP8 and VP9 and AV1, yes. (and defining the metdata)
[06:31:46] <Magnus Westerlund too> Justin, I think that is over simplifying it. Becasue the real choice for the structure and what you can do on RTP will depend on what you declare as IDU and put in individual Sframes.
[06:31:57] <Roni Even> or everything is in the payload itself and not transparent to RTP
[06:32:23] <Justin Uberti> sure, this is still at the 1000 foot level. But the goal here is to avoid having the SFU having to mine through the packet.
[06:32:44] <Justin Uberti> so we extract the meta for each IDU, and then pack the IDUs into RTP packets to fit within MTU.
[06:32:54] <Martin Thomson> I heard that the SFU routinely does sift through things like NALUs.
[06:33:07] <Justin Uberti> well, a NALU is technically an IDU.
[06:33:27] <Stephan Wenger> wish that were true.  But most SDUs need to look into slice headers and such
[06:33:50] <Martin Thomson> I heard that small NALUs could fit multiple per IDU.  And Stephan did make a point about slice headers.
[06:33:52] <Magnus Westerlund too> Justin, that is still the 1000 foot view. When you drill down into this and look at what has been done before there are actually quite some challenges here.
[06:33:56] <Jonathan Lennox_863> There would need to be a document somewhere that says that the H.264 IDU is a NALU, whereas a VP8 IDU is a frame.  Presumably that's in SFrame
[06:33:56] <Stephan Wenger> NAL unit headers were designed to take care of all SFU needs, but we missed too much when designing them.  True for all H.26x standards
[06:33:57] <Justin Uberti> right; even with VP8/9 SFUs need to look at the first payload bits.
[06:35:12] <Justin Uberti> If we can extract the needed metadata into a common cleartext metadata block, then we can just encrypt the IDUs entirely and simply chop them up according to MTU.
[06:35:18] <Magnus Westerlund too> You will also have the aspect that if you can move part of the IDU to the meta data or need to interpret and set a META data header for each IDU to be useful.
[06:35:31] <Martin Thomson> ISTM that you have a function that produces chunks of bytes and then you need a function that tells you what parts of those chunks needs encryption and then those inputs can be fed into a function that encrypts just those bits that need encryption, authenticates the rest and wraps them in sframe goop.  At that point, your SFU can still get what it needs, but it has to sift through the sframe goop to get to the codec goop.
[06:35:57] <Magnus Westerlund too> Then you have those aspects of decoding order, vs what video frame they will result in after decoding and where on the timeline that happens.
[06:36:21] <Justin Uberti> sure, DON could be part of the meta.
[06:36:29] <Martin Thomson> Random access should be possible on a per-IDU basis, so I don't see that being a problem
[06:36:38] <Martin Thomson> Unless you bundle too much in an IDU.
[06:37:00] <Roni Even> to me it looks like you do not want the SFU to look and do anything with the media. Just use RTP for transport level. Like I wrote before, in H.320 we had a V.35 IF that provided us with the exit from an encryption and the media was packtetize , military use
[06:37:14] <Justin Uberti> mt: yeah, one could do that, in which we say that the packetization is as defined by the codec and we simply mask the encryption to only touch the bits that aren't considered metadata.
[06:37:36] <Martin Thomson> That seems most consistent with Colin's views on codec specificity.
[06:37:40] <Jonathan Lennox_863> There would need to be agreement on which bits those are, though.
[06:37:46] <Martin Thomson> It is also the worst in terms of maintenance.
[06:37:53] <Martin Thomson> Jonathan, that's the easy part :p
[06:37:54] <Bernard Aboba> Martin: In Sframe, SFM by definition does not parse packet payloads.
[06:38:14] <Martin Thomson> Bernard: what I've heard here is that this is a definitional error.
[06:38:16] <Bernard Aboba> That is what distinguishes it from a MANE, which does.
[06:38:23] Waqar Zia leaves the room
[06:38:27] Waqar Zia joins the room
[06:38:41] <Jonathan Lennox_863> Martin: I think it's a lot more complicated than you necessarily think.
[06:38:54] <mt> No doubt.
[06:39:07] <Justin Uberti> it comes down to whether we want SFU codec independence to be a goal, I suppose.
[06:39:24] <Justin Uberti> If we have a common meta and packetization, then the SFU can be codec-agnostic.
[06:39:24] <Sergio Garcia Murillo> as an SFU implementator, yes, please.. :)
[06:39:32] <Stephan Wenger> that's too hard a problem.
[06:39:33] <Magnus Westerlund too> I think this will be quite similar to the effort we had to do when actually arriving at RFC 7656: A Taxonomy of Semantics and Mechanisms for Real-Time Transport Protocol (RTP) Sources but for Encoded Media.
[06:39:45] <Justin Uberti> which I think was a huge success!
[06:40:24] Waqar Zia leaves the room
[06:40:42] <Roni Even> what is the SFU topology, to me it looks like it only switch the media
[06:41:46] Waqar Zia joins the room
[06:42:35] <Jonathan Lennox_863> Whoops, sorry Mo, I didn't mean to remove you from the queue
[06:44:42] <Shuai Zhao> &nbsp;@jonathn, can you move one slide up?
[06:46:30] <Magnus Westerlund too> Roni: for an SFU to do what it can do today, you do need to be able to select the appropriate subset of SFRAMEs that are useful for a decoder as well as determine what it need to repair to serve its downstream consumers. That is not necessarily a simple due to decoding order and scalability layers.
[06:48:07] <Justin Uberti> right; for example you might choose to cache P pictures to allow a new joiner to catch up, rather than have to FIR.
[06:48:18] <Justin Uberti> (or GDR :)
[06:53:05] <Roni Even> but since the Sframe is constructed before the RTP layer it can be just a blob of data that is to be switched, propriety encryption. I am not sure what are the requirement from the payload that are sent to the RTP layer. Will there a payload specification for the encrypted frame, will it be negotiated?
[06:54:05] <Justin Uberti> yes, some negotiation will be needed.
[06:55:02] <Roni Even> currently for payload a public available specification is required for an RTP payload (I think that is what Colin commented
[06:56:40] <Roni Even> this may make it not codec agnostic
[06:57:20] Roni Even leaves the room
[06:57:44] Martin Thomson leaves the room
[06:57:52] Brendan McMillion joins the room
[06:59:00] Paul Kyzivat leaves the room
[06:59:57] Timothy Panton leaves the room
[07:00:10] Timothy Panton joins the room
[07:00:20] Timothy Panton leaves the room
[07:00:23] Timothy Panton joins the room
[07:00:26] <Magnus Westerlund too> So there is one important assumption here that do affect us in regards to signalling. It is assumed that SFRAME capable WebRTC endpoints will have some application specific method to determine what an SFRAME contain when it comes to codecs.
[07:00:33] Timothy Panton leaves the room
[07:00:36] Timothy Panton joins the room
[07:00:41] Timothy Panton leaves the room
[07:01:37] <Justin Uberti> well, that could be represented in PT.
[07:01:56] <Harald Alvestrand> That's the current implementation. It would be great if SFRAME signalling in sDP could contain that information.
[07:02:23] <Harald Alvestrand> but there's no way an intermediate node can verify that this is in fact what's interchanged
[07:02:31] Timothy Panton joins the room
[07:03:06] <Magnus Westerlund too> Ok, then we do get into the intersting question of a name space. But, clearly it possible.
[07:03:19] Barry Leiba leaves the room
[07:03:43] <Tim Bruylants> thanks all.. bye
[07:03:44] Mike Coleman leaves the room
[07:03:45] Ted Hardie leaves the room
[07:03:46] Cullen Jennings leaves the room
[07:03:47] Youenn Fablet leaves the room
[07:03:49] Magnus Westerlund leaves the room
[07:03:49] Barry Leiba_179 leaves the room
[07:03:50] Matthew Jarosz leaves the room
[07:03:51] Bernard Aboba leaves the room
[07:03:52] Daniel Havey leaves the room
[07:03:52] Shuai Zhao leaves the room
[07:03:53] Stephan Wenger leaves the room
[07:03:53] Justin Uberti leaves the room
[07:03:54] james welch leaves the room
[07:03:55] James Gruessing leaves the room
[07:03:56] Harald Alvestrand leaves the room
[07:03:57] Mike English leaves the room
[07:03:57] Brendan McMillion leaves the room
[07:03:57] Gonzalo Salgueiro leaves the room
[07:04:00] Brian Rosen leaves the room
[07:04:01] Timothy Panton leaves the room
[07:04:05] Gunnar Hellstrom leaves the room
[07:04:07] Sergio Garcia Murillo leaves the room
[07:04:08] Luke Curley leaves the room
[07:04:19] Tim Bruylants leaves the room
[07:04:20] Magnus Westerlund too leaves the room
[07:04:20] Lucas Pardue leaves the room
[07:04:22] Mo Zanaty leaves the room
[07:04:52] Waqar Zia leaves the room
[07:04:52] Qi Pan leaves the room
[07:04:52] Praneet leaves the room
[07:04:52] Jonathan Lennox_863 leaves the room
[07:04:52] Alessandro Amirante leaves the room
[07:04:52] Martin Pettersson leaves the room
[07:04:52] Yago Sanchez leaves the room
[07:05:53] Meetecho leaves the room
[07:29:08] mt leaves the room
[07:30:26] alex-meetecho leaves the room
[07:30:29] undefined joins the room
[09:04:32] undefined joins the room
[18:35:02] undefined joins the room