[13:48:18] mark.ellison@jabber.org joins the room [13:48:37] mark.ellison@jabber.org leaves the room [18:33:26] tomkri joins the room [21:04:46] csp joins the room [21:26:56] jiang_xing_feng joins the room [21:27:06] jiang_xing_feng leaves the room [21:27:29] JonathanLennox joins the room [21:27:30] Simon Perreault joins the room [21:28:07] Hi, thanks! [21:28:11] Special thanks to Colin [21:28:13] alfredh joins the room [21:28:27] Agenda for monday... [21:28:35] Audio is working well [21:28:42] Good [21:29:42] Agenda for tuesday... [21:30:08] (These are the group status slides, obviously) [21:30:10] Note well... [21:30:16] jiang_xing_feng joins the room [21:30:40] Document status... [21:31:03] Jean-Francois joins the room [21:31:07] Dan York joins the room [21:31:24] Publication queue... [21:33:15] draft-ietf-avt-seed-srtp-06... [21:34:06] alter ego Shiraz joins the room [21:34:35] Jim joins the room [21:35:09] SVC Video payload: Thomas Schierl [21:35:43] What has happened since last IETF... [21:36:11] What is new - General... [21:36:17] dmw.winter joins the room [21:36:40] Changes from version 13 - New Packet Types... [21:37:33] ysuzuki joins the room [21:37:53] Dave Oran [21:38:42] Changes from version 13... [21:38:56] omerboyaci joins the room [21:39:18] Changes from version 13 - Signaling [21:39:46] scottmcglashan joins the room [21:40:47] Changes from version 13 - Signaling - New SDK parameters... [21:40:50] SDP [21:41:30] scottmcglashan leaves the room [21:43:34] Changes from version 13 - Signaling - sprop-operation-point-info example [21:43:39] scottmcglashan joins the room [21:46:53] Changes from version 13 - Signaling (cont.) [21:46:58] [And that was me at the mic just now] [21:47:20] scottmcglashan leaves the room [21:48:09] New slide, same title [21:48:11] New slide, same title [21:48:25] Signaling session dependencies for MST [21:49:11] Open issues [21:49:16] Simon Perreault leaves the room [21:49:51] (can't hear questions) [21:50:25] Thomas Wenger is asking for each point to be gone through in order to make notetaking easier [21:50:41] ok [21:51:06] And also asking for summary of comments on the list [21:51:18] Next steps [21:51:28] Volunteers for review? [21:52:41] Are people able to hear Roni from the the chair's mic? [21:52:50] I can hear Roni [21:53:10] Good, thanks. They had mentioned something about that mic not showing up on the recording [21:54:51] Payload format for H.264 -- Roni Even [21:55:21] Outline [21:56:03] Overview of activities since Dublin [21:56:29] Summary of technical changes [21:57:07] Summary of technical changes - bug fixes (1) [21:57:52] Summary of technical changes - bug fixes (2) [22:00:48] jiang_xing_feng leaves the room: Replaced by new connection [22:01:03] Simon Perreault joins the room [22:01:35] magnus joins the room [22:02:00] Summary of technical changes - enhancements (2) now [22:02:51] Backward Compatibility Issues (1) [22:03:57] Backward Compatibility Issues (2) [22:06:32] (3) [22:06:56] JonathanLennox leaves the room: Replaced by new connection [22:07:24] Moving to Open Issues [22:09:06] Jonathan Lennox@mike on open issue #2 [22:09:37] simov joins the room [22:09:42] the issue which I mentioned offline is that rather than having the ssrc-based param specific to H.264, we define something more generic in the MMUSIC decoding dependency draft [22:10:57] Jonathan also proposed to define the sprop param at the fmtp attribute level [22:11:08] --- new slide: Request and question to WG [22:11:41] (end of presentation by Roni) [22:12:34] --- Thomas Schierl on draft-schierl-avt-rtp-multi-session-transmission-00 [22:12:55] Slide: What is it about? [22:13:06] JonathanLennox joins the room [22:13:22] Slide: RTP Solution [22:14:09] Slide: problems Data Alignment [22:16:17] Joerg at the mike: what magnitude are we talking about re: clock skew? [22:16:31] Thomas: tens of milliseconds in imprecision [22:17:08] Magnus W: when it comes to clock skew in reference to clock, it is up to 15% [22:17:24] "15% too fast" [22:17:35] Dave Oran: confused [22:17:55] Dave: are you worried about equipement that is too old that you cannot get a good system clock tick [22:18:20] Jim leaves the room [22:18:21] That's the point I made in email: you don't need to use gettimeofday() to get the SR clock [22:20:22] (sorry, was at the mike to pass on Colin's comment - Jonathan L will voice it as he is next online) [22:21:10] Magnus and Dave Oran are making my point pretty well, actually :-) [22:21:56] Jonathan Lennox at the mike [22:22:55] Magnus: explaining how to compute clock skew and what reference point to take [22:23:12] Roni: suggest that authors write some example of the pbs and put it to the list [22:24:57] Next slide - slide 5 [22:25:52] Colin, good, that we are agreeing. It is unfortunate that this isn't better described in the RTP spec, nor you book if I remember correctly. [22:27:15] Next slide - 6, possible solutions and source correlation [22:27:40] Magnus: Steve Casner had some fairly detailed notes at some point, but I don't think they were ever published. It would be good to write something to explain how to implement this on common platforms. The spec works, but is often badly implemented. [22:28:15] Next slide - 8 [22:28:37] Yes, I agree that it is all in the implementation. But, it takes quite a lot of time to understand what the specification really is saying. [22:28:56] Next slide - 9 [22:29:50] Dave Oran: ask for definition of RTP timestamp alignment [22:30:14] This timestamp alignment doesn't work for synchronising non-layered streams. You'd need to implement two mechanisms. [22:30:40] DAve: don't know how to do this without signaling it [22:30:43] (one for layered codecs, one for lip-sync) [22:32:41] Xavier Marjou: timestamp sync is complicated. Stronly support timestamp alignment [22:32:43] If you implement the timestamp solution, you also need to implement the SR mechanism for lip-sync. it doesn't help. [22:33:03] Jean-Francois leaves the room [22:33:37] Colin: yes, but the requirements for lip-sync and for layered codecs are very different [22:33:52] Sure, but the SR mechanism addresses both [22:33:59] [I'm taking over slide presentation again now[ [22:34:16] Cullen Jennings joins the room [22:34:30] Joerg is presenting Colin's slides: current slide: Synchronization of RTP flows [22:35:09] Slide: How fast does RTP synchronize flows [22:36:45] No - RFC 3550 explicitly states unicast sessions [22:37:13] not two participants [22:37:16] "Beginning of Session" is not the same as "when a source starts" [22:37:30] Slide: SSM flows [22:38:18] According to 3550, this is when a source starts in a unicast session. [22:38:39] Consider a MCU relaying streams [22:38:46] Jean-Francois joins the room [22:39:07] Slide: faster synchronization: packet loss [22:39:08] we should take it to the list - I don't see the issue [22:40:10] It's the bullet on the next slide, actually. :-) [22:40:15] Video switching [22:40:28] ah, ok [22:41:20] Slide: Summary [22:43:47] Next presentation: Rapid Sync RTP Multicast [22:44:02] dmw.winter leaves the room [22:46:21] There are two drafts on this; great commonality; they want to draft a single submission [22:46:43] Bill Ver Steeg: Rapid Synch for RTP Multicast [22:46:52] jiang_xing_feng joins the room [22:46:58] Orit Levin: Microsoft declares IPR on both of these drafts [22:47:19] Bill Ver Steeg: Cisco declared IPR on draft -00 [22:47:29] Slide: A Typical Multicast Join Scenario [22:50:29] [Sorry, falling behind] [22:50:36] Slide: Rapid Synchronziation [22:51:22] Slide: RMS Request [22:51:32] Slide: RMS Information [22:52:08] simov leaves the room [22:52:36] Slide: RMS Termination [22:53:24] Slide: Multicast Join XR Report Block [22:53:48] Slide: Issues in Control Plane [22:54:35] Slide: So we need redundancy in CP [22:55:30] Slide: SDP Example [22:55:35] Slide: Open Source Implementation [22:56:19] Slide: Next Steps [22:56:36] Joerg Ott: draft makes assumption you can always prime faster than realtime [22:57:25] Simon Perreault leaves the room [22:57:43] Dave Oran: yes, it doesn't make sense to do if you can't [22:58:08] Dave: We'll send to the list our IEEE paper that discusses what you can do [22:59:26] (can't hear Orit) [22:59:40] Next presentation: Orit Levin, Extensions to RTCP Feedback Mechanism for Burst Streaming [23:00:10] better, thanks [23:00:20] Slide: Introduction [23:01:41] Slie: Transport Layer Mechanism [23:04:24] Slide: Variations in Network Topologies and Constraints [23:08:13] Joerg Ott: How often do you update information about bandwidth constraints? [23:08:31] Slide: Example Topology [23:09:54] Slide: computers have no notion of what their access link rate is at any given time [23:10:01] Er, Joerg, not slide [23:13:18] Joerg: your wi-fi device has no idea if someone turns on the microwave and jams your 802.11g [23:13:50] Magnus (with Transport Area Director hat): you need to have a solid story about how this is handled in generic cases [23:15:19] Steve Casner: scaling and delay assumptions? [23:15:27] Next slide: Scalability and Efficiency [23:17:30] omerboyaci leaves the room [23:17:34] Dan York leaves the room: Computer went to sleep [23:18:39] Next slide: Receiver Driven Burst Mechanism [23:20:48] jiang_xing_feng leaves the room [23:24:23] [Didn't catch name]: What happens if SCI message gets lost? [23:25:03] Dan York joins the room [23:25:03] Roni Even: further discussion of the problem [23:25:12] Ali Began? [23:25:32] I think so [23:25:42] Dan York leaves the room [23:28:04] Dave Oran: can control-theory people give input on stability of the mechanism? [23:28:41] The other scheme is stable, there's a proof in the IEEE paper [23:29:00] Slide: Why using RTCP BYE is not a good idea [23:30:59] Jean-Francois leaves the room: Computer went to sleep [23:31:04] Joerg: if burst doesn't stop, client retransmits stop request. What happens if set-top box crashes? [23:31:48] Cullen Jennings leaves the room [23:32:26] Next slide: comparison of primitives of two drafts [23:32:34] alter ego Shiraz leaves the room [23:32:42] Next slide: this mechanism is faster than satellite [23:32:49] Next slide: next steps [23:34:15] Slide: agenda for tomorrow [23:34:20] Session ends [23:34:28] JonathanLennox leaves the room [23:34:37] csp leaves the room [23:35:28] ysuzuki leaves the room [23:38:10] alfredh leaves the room [23:46:38] Jean-Francois joins the room [23:47:07] magnus leaves the room [23:47:24] Jean-Francois leaves the room