--- Log opened Wed Aug 04 16:42:02 2004
16:42 <@rillian> hello world
16:49 -!- thomas_ [~thomas@226.Red-81-38-170.pooles.rima-tde.net] has joined #xiphmeet
16:53 -!- thomasvs [~thomas@176.Red-81-39-162.pooles.rima-tde.net] has quit [Client Quit]
16:54 -!- jack [~jack@host-64-179-26-66.pro.choiceone.net] has joined #xiphmeet
16:58 -!- MikeS [~msmith@CPE-144-137-40-253.vic.bigpond.net.au] has joined #xiphmeet
17:00 -!- rillian changed the topic of #xiphmeet to: Welcome to the Xiph.org meeting and discussion channel. Our wiki, where we normally schedule things, is still down. Monthly Meeting in progress; live log at http://westfish.xiph.org/~giles/200408_meeting.txt
17:00 -!- xiphmont [~xiphmont@host-64-179-26-66.pro.choiceone.net] has joined #xiphmeet
17:00 <@rillian> hi monty
17:00 <@rillian> shall we begin?
17:01 <@rillian> We're without an official agenda
17:01 <@rillian> so let's just go through the news
17:01 <@rillian> first, the hosting situation
17:01 <@rillian> we now, finally, have motherfish-iii installed and working on the Oregon State University Open Source Lab
17:01 <@rillian> who generously stepped forward to host us when our previous arrangements imploded
17:02 < xiphmont> I sent them stickers.
17:02 <@rillian> currently I think it's only hosting svn
17:02 <@rillian> with everything else still on westfish
17:02 <@rillian> but at least that's under control
17:02 -!- myidiym [~myidiym@24-50-57-104.sbtnvt.adelphia.net] has joined #xiphmeet
17:02 <@rillian> Also, we have a new volunteer webmaster team
17:02 <@rillian> who have been helping with the site maintenance
17:02 <@rillian> webmaster at xiph.org now goes to all of them
17:03 <@rillian> Atamido: do you have anything you'd like to share about that?
17:03 < xiphmont> there are multiple webmaster lists; are they handling only xiph.org now?
17:03 -!- Arc [~arc@208-59-116-202.c3-0.161-ubr1.lnh-161.md.cable.rcn.com] has joined #Xiphmeet
17:03 <@rillian> xiphmont: the idea was for all those to go away
17:03 <@rillian> I'm not sure if they're getting feedback at vorbis.com or now
17:03 < xiphmont> OK.
17:03 <@rillian> or not now
17:04 <@rillian> but they should :)
17:04 <@rillian> anyway, Atamido is one of the more active members of the new team
17:04 < xiphmont> Hello Atamido!
17:05 <@rillian> we've also moved most of the websites in the svn, to simplify maintenance
17:05 <@rillian> and allowing us to have a number of mirrors, at least for the static sites
17:05 <@rillian> and we have a new account on ftp.osuosl.org, aka downloads.xiph.org, where we should put high-bandwidth items
17:05 < xiphmont> how will we reintegrate personal sites?
17:05 < xiphmont> [I know the answer, others don't ;-]
17:06 <@rillian> that will probably be rsynced from a staging area on mf3
17:06 <@rillian> since we don't really want video files and so on in svn
17:06 <@rillian> jack and I have been working on an ldap scheme for account handling
17:06 < Arc> is it possible to get that working with bittorrent?
17:06 <@rillian> so it will be easier to share access between the various machines
17:07 <@rillian> and so we can partition shell, svn, and mail access
17:07 <@rillian> that will take another month or two, probably
17:07 <@rillian> unfortunately, we've still not recovered the wiki, or user data from mf-ii
17:08 <@rillian> personal web space will have to go on a separate domain, because of the mirror system
17:08 <@rillian> probably people.xiph.org, in analogy to debian
17:08 <@rillian> that will get done after the ldap stuff
17:08 <@rillian> in the meantime, use westfish.xiph.org
17:09 <@rillian> Arc: we can't run bt on downloads.xiph.org directly; that will have to run on mf3 or westfish
17:09 < xiphmont> I'm more concerned with currently broken links from everywhere.
17:09 <@rillian> xiphmont: let the webmaster team know where you'd like redirects to go
17:10 <@rillian> That's all I can think of on the hosting
17:10 <@rillian> any other questions?
17:10 < xiphmont> ok.  Atamido has xiph email?
17:10 < xiphmont> oh, webmaster.
17:10 <@rillian> no. just write to the webmaster address
17:10 < xiphmont> Jack ha mocked me in real life.
17:10 <@rillian> or better, file a bug again the website at bugs.xiph.org
17:10 < xiphmont> ok
17:11 <@rillian> The other major piece from my corner is that derf has finished the draft of the theora spec
17:11 <@rillian> three cheers for derf!
17:12 <@rillian> so that's groovy
17:12 < Arc> yaa!
17:12 <@rillian> much cleanup still to be done
17:12 <@rillian> and since I've not made any progress on the code updates, beta 1 is still a couple of weeks off
17:12 <@rillian> but that's a load off
17:12 <@rillian> also, too bad the csiro people aren't here
17:13 <@rillian> we decided about 3 weeks ago to try adopting liboggz as the 'official' recommended mux library
17:13 <@rillian> kfish was going to port it to ogg2
17:13 <@rillian> and I was working on something that might turn into libfishsight
17:14 <@rillian> or whatever, an equivalent to libfishsound that handled video
17:14 <@rillian> derf_: you've also been working on your experimental theora library. what's new there?
17:15 < derf_> Well, the decoder now produces bit-exact output compared against the current reference.
17:15 < derf_> I still need to add out-of-loop postprocessing and do some minor API cleanups.
17:15 < derf_> On the encoder side, I've started to take a look at why the HVS system produces wacked thresholds.
17:16 < derf_> I've at least confirmed it's not a problem with the quantization matrix generation.
17:16 <@rillian> so is your encoder still not doing a better job than the reference?
17:17 < derf_> Correct.
17:17 <@rillian> ok
17:17 < derf_> I will at least make the remaining changes to the decoder tomorrow. I'm not going in to work.
17:17 < derf_> And I'll see what I can find on the encoder side.
17:17 <@rillian> sounds good
17:18 -!- j^ [~j@135.bootlab.org] has quit [Read error: 113 (No route to host)]
17:18 <@rillian> xiphmont: do you want to tell us what you've been up to?
17:18 < xiphmont> Sure
17:18 < xiphmont> AoTuV tuning integration happened.
17:18 < xiphmont> That also involved some bugfixes, and additional testing of the new bitreservoir rate management.
17:19 < xiphmont> Right now, I'm updating libvorbisenc, libvorbisfile and libvorbis docs
17:19 < xiphmont> as well as integrating outside doc submissions relating to libvorbis from the BBC.
17:19 <@rillian> yay.
17:19 < xiphmont> After that, time to start dealing with vorbis and tremor patches.
17:19 <@rillian> so what does the release plan look like for 1.1?
17:20 <@rillian> after the patch round?
17:20 <@rillian> or should we do 1.1.0 immediately
17:20 < xiphmont> 1.1 is signed off for release.  Just needs release engineering.
17:20 <@rillian> oh, ok.
17:20 < xiphmont> 1.1.0 should be Done.
17:20 < thomas_> does that mean I can't do build fixes anymore ?
17:20 < xiphmont> we keep working of course.
17:20 <@rillian> so 1.1 as tagged can go
17:20 < thomas_> I'd at least like it do pass make distcheck
17:21 < xiphmont> thomas_: Persdonally, I think of buildfixes as part of release engineering,.
17:21 <@rillian> thomas_: that means work on a branch from the 1.1rc1 tag
17:21 <@rillian> make an 1.1rc2
17:21 <@rillian> and then we'll release that
17:21 < xiphmont> right
17:21 < thomas_> rillian: mail me with more specifics, I don't know how you traditionally have done it
17:21 < xiphmont> wait
17:21 <@rillian> thomas has been doing a lot of fixups on the autotools stuff. thank you!
17:22 < thomas_> is there any hope I can at some point convince you guys to not put letters in release versions at all ?
17:22 < thomas_> or should I give up before I start on that ? :)
17:22 <@rillian> thomas_: the stable releases don't
17:22 < thomas_> rillian: you're very welcome
17:22 <@rillian> this will be libvorbis-1.1.0
17:22 < xiphmont> I'd like head to reflect upcoming 1.1 final in progress.
17:22 -!- dev0 [Tobias@dev0.registered] has joined #xiphmeet
17:22 < xiphmont> and non-1.1 work to branch for now.
17:22 < thomas_> rillian: doesn't matter much, since non-stable releases of all of your stuff get packaged regardless
17:22 < xiphmont> that's been our traditional flow.
17:22 <@rillian> xiphmont: ok, that works too
17:23 < thomas_> ok, so I fix on head ?
17:23 <@rillian> you'll work on a branch
17:23 <@rillian> and we just do release engineering on head
17:23 < thomas_> ok, understood
17:23 <@rillian> xiphmont: current head is equally safe to release?
17:23 < xiphmont> If his fixes are needed for release, they go on head.  If they're just further work, branch.
17:23 < xiphmont> head is currently safe to release AFAIK.  We need a sweep through tools.  Libs are certified.
17:24 <@rillian> ok
17:24 <@rillian> thomas_: lets run with that then
17:24 < dev0> (hint: get Case to do the Win32 release)
17:24 < thomas_> ok
17:24 <@rillian> MikeS: what's new with icecast?
17:24 < thomas_> did you guys bump the .so versioning correctly this time as well ?
17:24 < thomas_> or does that still need checking ?
17:24 < xiphmont> 'part of the release engineering'
17:24 <@rillian> thomas_: that still needs checking
17:24 < thomas_> ok, will do
17:25 < xiphmont> I don;t bump anything that's part of the build system, just versioning strings in the source.
17:25 < thomas_> 1.0.1 went out with the same lib versioning even though api was added, want to avoid that this time
17:25 < xiphmont> thomas_: agreed.
17:25 < MikeS> ok, so icecast:
17:26 < xiphmont> [Oh, also spent a week on Audacity bugfixes and scalability.  That's not Xiph work, but it counts as Friends of Xiph in the same basic domain]
17:26 < bond> xiphmont: will vorbis 1.1 support channel coupling for multichannel?
17:26 < MikeS> nothing much has been happening in a user-visible way. I've been talking with karlH to get some of the more reliable bits of his branch merged back; that's currently in a seperate branch for testing/merging, and will go into mainline in the next week or so.
17:27 < MikeS> I've been away on holidays, so I haven't been doing anything more than that.
17:27 < xiphmont> bond: for > stereo, no.  That's stil coming.
17:27 < bond> k thx
17:27 < dev0> xiphmeet: Is that still in que for Vorbis I?
17:27 < xiphmont> Writing the dev proposal for 5.1 is on tonight's docket.
17:27 < dev0> s/que/queue
17:27 < xiphmont> yes, Vorbis I will get that.
17:27 -!- jmspeex [~jmspeex@pxy-2a.kyoto-u.ac.jp] has joined #xiphmeet
17:28 < bond> :)
17:28 <@rillian> hi jean-marc
17:28 < jmspeex> hi
17:28 <@rillian> how's japan? :)
17:28 < jmspeex> After 15 minutes, the proxy here finally decided it would allow me to log in...
17:28 < jmspeex> It's nice, but I've only been there for 4 days, so I'm still lost a bit.
17:29 <@rillian> MikeS: thanks for working on that. I'm glad stuff is starting to trickle back in
17:29 <@rillian> I'll also mention that the kh branch of icecast does support theora quite well
17:29 <@rillian> we've done quite a few tests now
17:29 < jmspeex> Is there a log?
17:30 < derf_> http://westfish.xiph.org/~giles/200408_meeting.txt
17:30 <@rillian> http://westfish.xiph.org/~giles/200408_meeting.txt
17:31 <@rillian> Jan's written some nice commandline utilities that make live stream creation easy as well
17:31 <@rillian> the main issue is cpu, since theora encoding is still quite slow
17:31 <@rillian> jmspeex: do you want to give an update on what's happened with your work?
17:32 < jmspeex> rillian: sure
17:32 < jmspeex> Basically, I've released 1.1.6 and 1.0.4 just before flying to Japan.
17:32 < jmspeex> I've been working on a small low-latency VoIP application.
17:33 < jmspeex> The goal is mainly to show people how to write a VoIP app that has decent quality.
17:33 < jmspeex> Otherwise, all the details are in the 1.1.6 and 1.0.4 release notes...
17:34 < jack> are you going to commit that to the repo?
17:34 < jmspeex> The VoIP app?
17:34 <@rillian> yes
17:34 <@rillian> we're all quite excited about it :)
17:34 < jmspeex> Sure, I'll make a small xiph project with it.
17:35 <@rillian> nifty
17:35 <@rillian> and congrats on your releases
17:35 < jmspeex> Note that it doesn't use any connection protocol. All the code is for the communication itself.
17:35 < jmspeex> (but it uses wideband and the quality is much better that what's currently available)
17:35 <@rillian> that's really the hard part, isn't it?
17:36 < jmspeex> Depends for whom ;)
17:36 < dev0> jmspeex: I discovered another nice project using speex some days ago... shtoom, which is an SIP/RTP based VoIP app.
17:36 <@rillian> jmspeex: anything else?
17:37 < jmspeex> Well, for that app, I had to actually make the Speex jitter buffer work (I think it's pretty decent now).
17:37 <@rillian> always nice to have application feedback :)
17:37 < jmspeex> And I've been working a bit on acoustic echo cancellation, but the results still sucks.
17:38 < MikeS> by the way: someone asked earlier about feedback@vorbis.com and the webmaster list: it goes to the webmaster list, but gets caught in the moderation queue. I went through ~1500 messages in that queue a week and a half ago, but I can't do that regularly.
17:38 <@rillian> sorry you couldn't make it to Bordeaux last month. would be have nice to meet in person
17:38 <@rillian> MikeS: ok, we should add it to the exception list
17:38 < jmspeex> rillian: see you next year ;)
17:38 <@rillian> thanks for doing that
17:38 <@rillian> Ok. also vorbis-related
17:39 <@rillian> jack wrote a python wrapper for vorbisfile last week
17:39 <@rillian> it's quite nice; you just hand it a python file-like object and it will do it's thing
17:39 <@rillian> it's in the repository under http://svn.xiph.org/trunk/vorbisfile-python/
17:39 < MikeS> rillian: yes. I don't know how to do that. (I think there are other addresses that are suffering the same fate)
17:39 <@rillian> if anyone wants to play with it
17:39 < Arc> jack: did you commit that back to the py-vorbis author?
17:39 < jack> and unlike the current pyvorbis, it supports the complete vorbisfile api.
17:39 < jack> it's in trunk/vorbisfile-python
17:40 <@rillian> it's a separate implementation; doesn't really make sense to try and merge it with py-vorbis
17:40 <@rillian> Arc: what have you been working on?
17:41 < Arc> well I fixed all the ogg2 bugs/enhancements I needed for py-ogg2, currently in branches/ogg2-arc and trunk/py-ogg2
17:42 < Arc> awaiting monty to look over the api extensions I made to libogg2
17:43 < Arc> there's now an Ogg reserializer in py-ogg2's examples directory (nice little tool) and I've started work on Xinloe (* is a non-linear Ogg editor), which can currently open a muxed/chained Ogg file of basically any codec and, from speex theora or vorbis, display the length/bitrate from the stream
17:43 < Arc> has nice little 16x16 icons for each based on sheryl's designs that it represents them with
17:44 <@rillian> Arc: you said you'd had some correspondence with Josh Coalson, the FLAC author?
17:44 < Arc> the plan is that since we currently dont have any way to either decode theora from python or display video, that it'll output a http stream to a (usually local) media player for visual feedback of what you're editing
17:46 < Arc> yea while I was going over this I found that the OggFLAC implementation is... lacking, and highly redundant.  basically, besides the header issues (like implementing seektables in Ogg), each packet contains dual CRC's (on top of Ogg's) and sync-patterns, etc
17:47 <@rillian> ouch. so the whole embedding might change?
17:47 < Arc> Josh reinterated that the current OggFLAC implementation is not finished, and invited help on getting it polished up.
17:47 < Arc> yea but it's not the first time the embedding has changed, they just recently got Ogg granulepos getting written properly, and their OggFLAC seek support is currently in CVS
17:48 <@rillian> well, I certainly want that to work with the oggmng stuff
17:48 <@rillian> so I can take a look
17:48 < Arc> but it really does have to change. besides not being highly redundant, it has weird stuff like "uncompressed MD5" in the header, thus making it completely unsuitable for live recording/streaming
17:50 < Arc> FLAC is basically both a codec and a container format, mixed into one, so this is a process of seperating the two so FLAC (the codec) can be used in either FLAC (the container) or Ogg
17:50 <@rillian> *nod*
17:50 <@rillian> sad to hear that
17:50 <@rillian> good thing no one is using ogg flac :)
17:50 < Arc> yea as Josh said in a recent email
17:51 < Arc> I'd like to work towards getting the FLAC team more integrated with Xiph.. what is the status of migrating them to Xiph servers/SVN/etc? or should we make that a seperate agenda item?
17:51 <@rillian> that's all up to them
17:51 <@rillian> we have mailing lists
17:52 <@rillian> we're happy to let them put stuff in svn and host their website
17:52 < Arc> how can we transfer their mailing list archives and cvs history?
17:52 <@rillian> the ball's always been in josh's court
17:53 < Arc> I don't think they "know how" to do the migration.. their website says to this day that they are working on migrating their lists over
17:53 <@rillian> Arc: if they give me an mbox of the list archives and a cvsroot tarball, I can do the import
17:54 < Arc> rillian: it's on sourceforge, so I don't think they have access to that stuff any more than any user does
17:54 <@rillian> they can ask nicely
17:54 < Arc> sourceforge uses mailman and, i think, pipermail for archives.. 
17:54 <@rillian> they used to use pipermail
17:54 -!- |bond| [~bond@u-122-163.adsl.univie.ac.at] has joined #xiphmeet
17:54 <@rillian> they may have turned it off; it stopped being accessible some time ago
17:54 < Arc> ok so this is a special request they have to make of sf?
17:55 < thomas_> you can ask for a tarball of mail archives too
17:55 <@rillian> anyway, last I checked project admins could download a tarball of their cvs repository
17:55 < thomas_> they just have to log a support request for that
17:55 <@rillian> but I'm sure they'll have to ask for list archives
17:55 <@rillian> just file a support request
17:55 < thomas_> that's how we imported all of our gst stuff into gmane
17:55 < Arc> ok
17:56 < Arc> does josh have an account on mf3/westfish?
17:56 <@rillian> no
17:56 <@rillian> he needs to send me an ssh key
17:56 -!- thomas_ [~thomas@226.Red-81-38-170.pooles.rima-tde.net] has quit [Client Quit]
17:56 <@rillian> well, it would be great if they did finally come over for real :)
17:57 < Arc> ok, cool. I'll let him know that then
17:57 <@rillian> Ok, I think that wraps things up for the news
17:57 < xiphmont> we need to invite him to our parties, duh.
17:57 <@rillian> any questions, or other things going on?
17:57 < xiphmont> [and he needs to show up]
17:57 < Arc> xiphmont: I don't get invited to parties :-P
17:58 < xiphmont> You sound like my roommate...
17:58 < MikeS> I have one question: now that mf3 is up, are we back to using normal ssh-authenticated svn, rather than the temporary thing with http?
18:00 < Arc> (which is probobally a good thing since I usually can't attend anyways!)
18:00 < myidiym> rillian: what's our agenda today, and where are we on it?
18:01 < myidiym> oh, and this is Brett
18:01 <@rillian> oh, hi brett
18:01 < MikeS> myidiym: we didn't have a formal agenda, and we're at the end of it unless you want to bring something up
18:01 <@rillian> our agenda is: progress reports, random questions
18:01 < MikeS> (who are you, Brett?)
18:01 <@rillian> we just started the second bit
18:01 <@rillian> MikeS: my thinking was that http auth is really nice
18:01 <@rillian> because the icecast people can do their recursive checkout thing
18:02 <@rillian> and we don't have to give shell accounts for every svn write account
18:02 <@rillian> otoh, plaintext auth makes me nervous
18:02 < myidiym> MikeS: I'm the "standards guy" from IEEE-ISTO
18:02 <@rillian> so I was thinking we should switch to https for write access
18:02 < MikeS> myidiym: ah. hi!
18:02 -!- thomas_ [~thomas@52.Red-83-35-138.pooles.rima-tde.net] has joined #xiphmeet
18:03 < MikeS> rillian: yeah, the plaintext bit was what made me nervous... is it possible to at least set it up as https only?
18:03 < MikeS> (is it even possible to use https?)
18:03 < thomas_> (crap dsl.  like I said after dropping out - if they log a support request for a mailing list archive, sf gives it to you.  gst used it to import all mailing archives in gmane)
18:04 <@rillian> MikeS: that is possible
18:04 < MikeS> sorry, it looks like you were ahead of me. But is it possible to configure it so that it won't _allow_ http (only https)?
18:04 <@rillian> it will still muck with the external only, because the url still changes
18:04 < myidiym> btw, I don't have anything to bring up.
18:04 < MikeS> thomas_: nah, you said that before you dropped out :)
18:04 < thomas_> MikeS: ah, good :)
18:04 < xiphmont> myidiyum: I do
18:04 -!- bond [~bond@u-122-163.adsl.univie.ac.at] has quit [Read error: 110 (Connection timed out)]
18:05 <@rillian> xiphmont: please go ahead
18:05 -!- |bond| is now known as bond
18:06 < xiphmont> We currently have no dedicated executive roles, that much is obvious.  For that reason, I'm going to transition out of coding, maybe altogether, to a) keep non-coding stuff rolling and b) hopefully more successfully funraise.
18:06 < xiphmont> At least, that's what I'm thinking now.  We don't have a sugardaddy and coasting is doing no good.
18:07 < Arc> xiphmont: what would happen to the projects you maintain?
18:08 < Garf> who's going to improve vorbis then?
18:08 < xiphmont> People take over coding maintainence or nothing happens.  But I do think there's enough work being submitted now to keep it moving.
18:08 < xiphmont> Garf: for the time-being, whoever steps up and not me.  I'll continue docs, but no new code for a while.
18:09 < xiphmont> That's if we truly decide we want to grow beyond 'three hackers and a copy of emacs in someone's basement'.
18:10  * Garf cries Emmett were areth thou
18:10 <@rillian> xiphmont: well, I suppose I'm happy someone's finally volunteering for that role
18:12 < xiphmont> Honestly, I'm not convinced my special brand of control-freakdom is appropriate to maintaining the codec at this point.
18:12 <@rillian> hehe
18:12 < myidiym> vorbis sits in a list of codec options that exist today and it enjoys a certain status in that list... how many months do you think we have in a "non-monty" Xiph.org before vorbis would begin to slip in its relation to other codecs available today?
18:12 < myidiym> "non-coding-monty" that is
18:12 <@rillian> it won't ever slip in relation to what's available today
18:13 < myidiym> ok
18:13 <@rillian> it will slip compared to whatever mpeg's latest audio flavour is each time they release a new one
18:13 < myidiym> any idea what kind of timeline mpeg is on?
18:13 < xiphmont> myidiym: oh, we're already slipping, but that's mostly not being on the bleeding edge.  
18:13 < Garf> you're still ok at the high bitrates...for now ;-)
18:14 < jack> I
18:16 <@rillian> ok, seems like there's not much further discussion
18:16 -!- thomas_ [~thomas@52.Red-83-35-138.pooles.rima-tde.net] has quit [Client Quit]
18:16 <@rillian> I move we close the meeting
18:17 < myidiym> g'night, g'day, g'bye
18:17 -!- Garf [Garf@Garf.active.supporter.pdpc] has left #xiphmeet ["(x) red pill ( ) blue pill"]
18:17 <@rillian> I'll take that for a second :)
18:17 < Arc> heh
18:17 <@rillian> thanks everyone for coming
18:17 <@rillian> this meeting is adjorned
--- Log closed Wed Aug 04 18:17:35 2004