Log Opened: Wed Mar 03 18:00:19 2004
<Misirlou> volsung: Pick me! Pick me!
*** Join: acolwell (~acolwell@209.210.138.240)
*** Join: xiphmont (~xiphmont@h004005a8a3c2.ne.client2.attbi.com)
*** Join: MikeS (~msmith@CPE-144-137-63-214.vic.bigpond.net.au)
<xiphmont> hihi
<volsung> HowdyHowdy.
<Misirlou> hallo
<vanguardist> m00
<goosedaemon> hum. the power is flickering a little here. i hope i don't get bumped off.
*** Join: Lee__ (~chatzilla@66.246.92.228)
*** Join: oddsock (~oddsock@h-67-101-202-59.CHCGILGM.covad.net)
<Arc> we're waiting for ralph still
<oddsock> howdy all 
<Arc> hey Ed
<Lee__> just sitting in
<Arc> rillian was just talking on #theora.. i wonder what happened to him?
<xiphmont> He's either off making tea or eBaying a small, third-world nation.
<xiphmont> let's give him a few more, if he doesn't arrive, I'll start it./
<xiphmont> [I too am making tea and getting situated]
*** Join: rillian (~giles@mist.thaumas.net)
<MikeS> (Monty already owns a small, third-world nation, so doesn't have to go and buy one on ebay)
<rillian> sorry, time zone problems
<rillian> shall we start?
<xiphmont> Yep.
<rillian> someone logging?
<goosedaemon> i am...
<rillian> First item is Release team report
<rillian> I can't say there's been any progress there
<rillian> ices2 release is still pending
<xiphmont> and 1.0.2?
<xiphmont> [I imagine not]
<rillian> I've not heard of any work in that direction
<rillian> btw, xiphmont can you fix the outstanding bugs in libogg?
<xiphmont> Ah, I need to coordinate with you, then.  I was unclear.
<xiphmont> 'CVS head is waiting to be released as 1.0.2, and has been since New Year'
<xiphmont> rillian: yes, I need to do that too.
*** Join: joeforker (dholth@3ffe:bc0:8000:0:0:0:0:262b)
<rillian> ok. sorry I misunderstood
<xiphmont> I'm more worried about build system, but there are outstanding bugs to worry about, yes.
<rillian> We Will Do vorbis 1.0.2
<MikeS> At least -tools also needs quite a bit of auto* work, 1.0.1 was pretty broken in that respect.
<joeforker> hi
<xiphmont> Yeah, well, I saw no traffic and didn;t tell you.  My bad.
<rillian> hi joeforker
<rillian> looks like the icecast folk aren't here
<oddsock> sure we are!
<xiphmont> BTW, it is a subminor release... I'll be happy with bugs fixed and assisting distros with packages.  
<rillian> oddsock: oops! sorry :)
<rillian> next item is codec and project reports
<rillian> I'll start with theora
<rillian> fluendo.com has given us funding to get to beta 1
<rillian> for which we are very grateful
<xiphmont> [jazzed!]
<derf_> What exactly does that mean?
<rillian> they've agreed to make a donation to xiph
<xiphmont> And we've agreed to use that to fund Theora.
<xiphmont> [that means Ralph]
<rillian> and I have a contract with them for doing part of the work
<MikeS> rillian: yes we is
<derf_> That doesn't sound like a donation.
<xiphmont> Some of it is not a donation; the lion's share is.
<xiphmont> "There is no contractual guarantee to deliver anything" however, we plan to use it for Theora.
<derf_> But okay. One-time fixed lump sum.
<rillian> the main sticking point has always been the spec, so I'm happy to subcontract with anyone who would be motivated to write for pay
<derf_> I'll write for money.
<xiphmont> That looks like a volunteer.
<rillian> derf_: feeling more confortable with the format?
<derf_> I like the term "contractor".
<derf_> rillian: Probably more so than anyone else on the planet.
<rillian> it's just last time I asked you wanted to finish your encoder first :)_
<derf_> Yeah. Well, it makes valid bitstreams now.
<rillian> there's also been lots of other activity.
<derf_> That's not "finished", but it's what I meant by that.
<rillian> karlh patched icecast to stream theora
<rillian> and there's been some work on front ends to feed it
<oddsock> there are some issues with karlh's code
<rillian> anyway, I'd like to shoot for the end of april for theora beta 1
<oddsock> mainly, he has a whole mess of changes that his theora code relies on, and none of it is in CVS at the moment
<MikeS> There are many issues with karlh's code. Primarily that he keeps sending me 100 kB+ patches for review that change just about everything, so they're completely incomprehensible. 
<Arc> rillian yea beginning of april == bad time for release.
<rillian> I've noticed that with karlh's tree
<oddsock> this is the problem, we are kind of at an impass..
<rillian> so what's up with icecast et al.?
<MikeS> well, we need to do an ices2 release. icecast itself is fine - we've got a release out, we've more-or-less got a plan for the next release.
<rillian> agreed
<rillian> who's in charge of that besides karlh?
<Arc> last i heard from karlh there were three issues with beta4, and two of them were fixed. he said he was expecting the third to be fixed soon.
* rillian heard likewise
<oddsock> technically I am in charge, but I'm not doing much to facilitate the code changes karl has to commit
<rillian> so there are two things
<rillian> I think we should concentrate on releasing ices2
<rillian> and then merging some of the new features into the trees
<MikeS> I'm reviewing karl's changes, at least when he sends me non-ridiculous patches.
<rillian> MikeS: that includes the recent small changes that have been going in?
<MikeS> No, most of the small patches are being committed without review. They're generally sensible, this is ok.
*** Join: adiabatic (~boingo@dsl-206-55-130-248.tstonramp.com)
<rillian> ok. I'm not sure we can decide anything without karlh here
<rillian> unless someone else wants to volunteer to package his changes
<rillian> in reviewable chunks
<MikeS> He _is_ generally doing that, it just takes me several tries for each new feature to get a reviewable patch. So it's very, very slow going.
<rillian> so you feel progress is being made, unlike oddsock?
<oddsock> I feel progress is being made...at a snail's pace...which may be okay, but it's been a few months now, and we haven't released anything
<rillian> is there a release goals thing for the ice* projects?
<oddsock> which, again is okay, but I thought we were on a bit quicker schedule
<rillian> like TheoraTodo in the wiki?
<oddsock> yeah, we have a release plan
<oddsock> not for ices, but icecast...let me find the link
<oddsock> http://wiki.xiph.org/Icecast2ReleasePlan
<oddsock> unfortunatley no timeframes though
<rillian> that's fine, I was just curious
<rillian> shall we move on?
<MikeS> Yeah, the problem is that the only really active developer right now is karlH, and he's far too willing to take the "throw-the-kitchen-sink-in-too" approach to software development, so we have to guide him.
<MikeS> Yes, let's move on
<oddsock> yup
<rillian> ok. Monty? tell us about ogg and vorbis developments?
<xiphmont> Sure.
<xiphmont> (Also in a side conversation about the IEEE-ISTO)
<rillian> (maybe we can do that with the affero item on agenda)
<xiphmont> I want to have a full proposal of OggMux, containing feedback of all the recent discussions in fully-fleshed-out but not editied form tonight.
<xiphmont> No, I'm typing :-)  I'm here for the meeting first off :-)
<goosedaemon> oggmux?
<MikeS> goosedaemon: full multiplexed ogg.
<goosedaemon> kay.
<xiphmont> The past month saw lots of discussion, and acolwell has a knack for distilling points into clear, consice examples.  We're mostly in agreement, some details are left.
<MikeS> How multiplexing is required to happen, additional requirements on codecs, any spec changes, etc.
<xiphmont> We've had basic spec for that for years, but it glossed over all the details.
<xiphmont> This doc is intended to nail all those details down.
<xiphmont> Once that is drafted and out for further discussion, OggFile API first draft happens.
<xiphmont> Also, for SXSW, I want to have a pass through the website done as well as an initial release of Postfish.
<rillian> how's that coming?
<xiphmont> How is which?
<rillian> I was particularly curious about the website
<xiphmont> Ah.
<Arc> yea and are we getting new logos or what? :-)
<xiphmont> I hope you're more curious about function than look. :-)
<xiphmont> Oh, I see.  Well, the Xiph fish morphs ever so slightly.
<xiphmont> (Matching the recent print doc logo)
<xiphmont> beyond that, no.  
<xiphmont> I was concentrating only on putting information where people expect to find it.
<adiabatic> (still waiting on word from monty on what the enw artist wants)
<xiphmont> The site maps are byzantine right now.
<adiabatic> <-on phone
<Misirlou> Do we have a vector version of The Fish?
<xiphmont> Of course.
<Arc> can you drop a url for it?
<xiphmont> The fish morph I'm doing.
<Misirlou> Yeah, I want something to drool over, too
<rillian> Arc: it's in vorbis/doc/xifish.pdf
<xiphmont> xiph.org/xiphmont/twirlfish
<xiphmont> I've not asked Melissa for wordmark.  Hm.  I'll take her out drinking tomorrow night.
<rillian> Ok. thanks for the update
<rillian> I've not heard of anything new with speex or flac
<rillian> what about the tools
<rillian> volsung?
<xiphmont> I have been in touch with Melissa, but also known her to be very busy recently.
<xiphmont> (Thus we don't have website visual work as she's been legitimately unavailable)
<rillian> (btw, one thing that should happen for the next ao release is to fix whatever broken the macosx plugin in panther)
<xiphmont> *YES*
<adiabatic> plugin?
<MikeS> I think tools haven't really been touched in ages. Well, except that I removed vcut from the build. 
<xiphmont> MacOSX support is one thing for the 'SXSW' portion of the agenda.
<rillian> I've looked at that a bit, but haven't traced the issue yet
<volsung> No news about speex or flac tools.
<MikeS> adiabatic: libao uses 'plugins' for output support (dynamically loaded)
<adiabatic> What of ours uses libao save for ogg123?
<volsung> That's it.
<volsung> Only other people who use it are non-Xiph.  (Like mpg321 and gaim.)
<rillian> ah, they're finally using our version? cool
<adiabatic> ...and who's clamoring for libao-on-OSX?
<rillian> adiabatic: me
<rillian> Ok, any other project updates?
<Arc> back to tools for a moment, if we can
<rillian> Arc: go ahead
<Arc> I'd like to discuss the ogg-tools/vorbis-tools + libtheora w/ tools included situation
<rillian> ah
<rillian> I'd say, either ogg* need to be renamed vorbis*
<Arc> I think we need to rip the examples/etc out of libtheora like pre-alpha3.  trying to build it on different machines w/o X is a pain because you have to mod it's build system to not build examples/*
<rillian> or then need to be rewritten from the ground up to handle video as well
<Arc> (since the player has SDL as a dependency, etc)
<rillian> Arc: that's called an autotools bug
<rillian> and is not a justification for removing the examples
<Arc> rillian or even with audio, there's been suggestions that flac/speex integration with ogg123/oggenc be justification to move it to ogg-tools
<rillian> *nod*
<MikeS> I don't think we need to go moving the actual code (it's basically just a pain to do so).
<rillian> and there's definitely a need for a more functional theoraenc
<adiabatic> Arc: for oggenc, oggenc just has fancier-than-wav-and-aiff loaders.
<xiphmont> yes, sorry about that ;-)
<Arc> some of the ogg-tools are also going to need updating soon anyways, to reflect the muxing drafts comming out when a decidion is made
<MikeS> I'm happy to have oggenc extended to support multiple ogg formats, as long as they're kept optional (i.e. I want to be able to just build a vorbis-only oggenc if I don't have everything else installed) if people want that. 
<rillian> MikeS: sounds good
<MikeS> I understand the naming issue, but I think users have been using oggenc for years, renaming it would confuse.
<rillian> I'd vote for extending oggenc to support all our formats then
<volsung> Audio and video?
<volsung> Or just audio?
<rillian> we can start with just audio
<Arc> I think it should do theora too. especially since I'd like to see theora doing speex/flac :-)
<adiabatic> I vote against.
<MikeS> On the other hand, I'm not sure a) how suitable oggenc's internals are to video, and b) how much people actually want to use a simple command-line tool for video encoding
<Lee__> I do daily
<volsung> I'm just afraid of getting something that looks like mencoder or transcode.  :)
* adiabatic waits for the floor
<xiphmont> I think that it needs to be a new tool with the same name.
<rillian> MikeS: I think the issue is more of how to handle multiplexed output
<Lee__> as do most people at archive.org
<xiphmont> We can do a good job of an all-codec tool even if other groups have failed through laziness.
<rillian> that's what affects the code internals and the ui
<MikeS> i.e. We can't go and throw decoding-ten-thousand-video-formats into oggenc, and I don't think it'd be that useful without that.
<rillian> Lee__: yes, obviously there's a need for such
<MikeS> Lee__: note: "simple command-line tool", not just "command-line tool". i.e. as volsung says, I don't want to turn oggenc into some monstrosity like mencoder.
<Arc> MikeS, well hopefully OggFile will be out sometime before we have two or three different video codecs, so we wont need to integrate them into the encoder
<xiphmont> Oh, you mean *inputs*, sorry.
<rillian> and clearly, we want mencoder and ffmpeg to support theora output for just those situations
<joeforker> some have suggested making a decoder that outputs something like yuv4mpeg + pcm in ogg, and letting an encoder parse that from stdin.
<xiphmont> I only meant outputs.
<derf_> Also, what makes us less lazy than anyone else?
<MikeS> Arc: err... no, you missed the point: people don't want to have to find other tools to produce whatever obscure raw format the theora example encoder requires - they want to be able to feed it divx, or mpeg-2, or whatever, and have it all "just work".
<Arc> MikeS, we can't support those formats anyways. patent issues. so why discuss it?
<MikeS> xiphmont: yeah, lots of outputs we can do a good tool for. But I'm wondering how useful it'd be without hitting problems with input formats that we don't (or can't due to patents) support.
<MikeS> Arc: that's my point! 
<Misirlou> derf_: heh
<xiphmont> Well, I think that thread started by accident.  
<Arc> but I do think that oggenc should eventually support theora. and writ. and mng/ming, and whatever else... to encode stuff to.
<xiphmont> Let's offline it
<rillian> agreed
<adiabatic> it = ?
<MikeS> ok
<MikeS> adiabatic: it = discussions on the future of oggenc and other ogg-* tools.
<rillian> do we have consensus that oggenc should at least be able to output speex and oggflac?
<adiabatic> rillian: Not from me.
<rillian> rough consensus then :)
<Arc> adiabatic why dont you expand on your objection?
<adiabatic> Sure.
<adiabatic> flac --ogg takes a number of arguments that have everything to do with the particulars of FLAC encoding and would simply be useless, confusing clutter elsewhere.
<rillian> fair enough
<MikeS> rillian: I'm willing to put in some work on generalising the output stages of oggenc somewhat, if other people (who know the speex/flac APIs - I'm pretty short on time at the moment) want to write the actual encoding bits.
<adiabatic> So not only do you have the problem of large swaths of manpage dedicated to flags you won't use...
<Arc> can OggFlac encoding be done without those extra arguments?
<adiabatic> ...you're also going to have to resolve what happens when the user uses... (one sec)
<MikeS> adiabatic: I'd suggest subsuming those within the --advanced-encode-option option.
<MikeS> and/or only supporting fairly basic usage in oggenc.
<rillian> let's not start the discussion again
<xiphmont> this is a more relevant discussion, but it belongs offline as well.
<xiphmont> adiabatic has a point, but let's not discuss it now.
<adiabatic> -a (--analyze in FLAC; name of artist in oggenc)
<rillian> Ok. moving on
<rillian> SXSW
<adiabatic> This is messier than "just s/oggenc/vorbisenc/".
<rillian> monty or vanguardist, any issues?
<rillian> what do we have planned?\
<xiphmont> Who all is actually coming?
<volsung> vanguardist is afk.
<volsung> But he sent me the letter confirming our booth at SXSW.
<xiphmont> For the most part, we're just being visible and assisting Dave Rose's crew.  I'll be doing an interview.
<rillian> Dave Rose being?
<volsung> doppler
<xiphmont> doppler; he's implemented all the live show Ogg streaming, and the fact that all the SXSW canned music streams are Ogg.
<volsung> Guy doing the webcasting of some of the shows.
<Arc> xiphmont does the xiph crew have shirts or something that identifies them as such? ie, fish logo polo shirts, printed teeshirts, something?
<rillian> so we'll be helping with the webcasting. that should be inpirational in the tools dept. :)
<volsung> Yeah, as far as I know, Manuel, Monty and myself will be working the booth.
<xiphmont> Arc: not right now.
<volsung> Manuel gets in Thursday night or Friday, i think.
<xiphmont> "Xiph Heavy Opera Company"
<Arc> xiphmont is there a plan to do such a thing? if nothing else teeshirts are quick and easy
<Misirlou> Acme Media Co.
<rillian> any other coordination questions?
<xiphmont> I'll look into it.
<volsung> Uh, what are we going to have at the booth?
<xiphmont> I actually need folks' itineraries.
* vanguardist pokes his head in
<vanguardist> I arrive Friday at 9am.
<xiphmont> email it!
<vanguardist> pfm said today that he'll be there for one day (possibly overnight) 
<xiphmont> arrive/leave, flights, etc.
<vanguardist> since he wants to see someone play saturday night
<Arc> SXSW is next weekend right, not this one?
<volsung> xiphmont: I live here, so you get to tell me where to be and when.  (I'll email contact info.)
<xiphmont> OK
<volsung> Arc: Two weeks.
<vanguardist> Arc: no, 17th
<Arc> for the booth/etc, is there interest in doing a live theora stream for demo'ing?
<vanguardist> heh
<vanguardist> volsung has some crazy stuff going on this laptop :)
<vanguardist> adding something else like theora would be cool 
<volsung> Heh, that's mostly just me showing off.  Not a real demo.  :)
<volsung> With the magic of virtual PC, I hope to be able to show Debian sourcing a stream to Win32 icecast being played with an OS X client.
<volsung> We'll find out how slow my box gets.
<rillian> Arc: I think our booth is in the music section
<Arc> I can't guarentee anything but support for this one camera I have (which outputs JPEG frames, fairly unique) but I'm confident I'll have it streaming to karlh's mod'ed Icecast by this weekend.
<Arc> you could, ie, have a camera overlooking the crowd from the booth
<Misirlou> DV -> Icecast?
<Misirlou> (In the future.)
<Arc> we don't have DV yet. eventually, not now.
<rillian> yes we do
<rillian> ffmpeg2theora does it
<Arc> we do?
<rillian> (it's just a little slow on a 3 GHz machine)
<Arc> rillian does ffmpeg2theora stream to libshout?
<rillian> oh, not that part. sorry. but that's the part you're writing.
<Arc> sorry off topic.
<rillian> quite
<rillian> that it for SxSW?
<rillian> Next item, Xiph.org as an Affero client
<rillian> who put that on? Want to tell us about it?
<Arc> I did.
<MikeS> how does anything manage to be slow on a 3 GHz machine?
<Arc> so I found this group called Affero (they're somehow affiliated with freenode)..
<rillian> *nod*
*** Quit: oddsock (Read error: 54 (Connection reset by peer))
<Arc> they are apparently like paypal, but for donating money in exchange for tech support/etc.  ie, someone helps you with Icecast, you drop them $5, the money goes to their preferred non-profits
<Arc> 501c3 isn't required (even non-profit status isnt required), but Xiph could quite easily be setup as a charitable org on there, and we could each drop ID numbers for affero. I don't know if we'll see much out of it, but it may be a small source of income
*** Quit: Lee__ ("ChatZilla 0.9.52B [Mozilla rv:1.6/20040227]")
<rillian> I've not heard of anyone making a helpful amount of cash through this site, but it is a nice idea
* Misirlou gives Monty all of his allowance
<Arc> i figure that we're doing tech support through freenode anyways, and it doesnt cost anything, we might as well play the game and hope it catches on?
<volsung> What's their cut?
<Arc> volsung you'd have to check their site for that, I'm not sure.
<rillian> Arc: can we sign up as individuals without having to be involved as an organization? (except possibly as a benificiary)
<rillian> maybe you should see if you can make anything answering icecast config questions :)
<Arc> yea, we would need to be signed up as individuals I think. but Xiph would need to be signed up as a benificiary for money to go to Xiph
<rillian> I think we should be in there anyway
<rillian> as a cause
<adiabatic> Arc: So, what would we need to do to sign Xiph up as a beneficiary, assuming it hasn't already been done?
<Arc> adiabatic I think anyone here could do it. I just wanted to discuss @ meeting first
<rillian> Arc: why don't you go ahead and register and see if you can get xiph in as a cause
<Arc> ok.
<rillian> the affero people are cooler than the paypal people, so I'd rather the cut go to them :)
<rillian> thanks
<rillian> Last item is utility standardization
<volsung> From the looks of one of their demo screenshots, they take 6%.
<rillian> this is related to the discussion we tabled earlier
<rillian> sounds like a reasonable idea to me
<xiphmont> utility standardization also has to do with how much specialization OggFile will pass through, and what it will try to standardize.
<rillian> any thoughts on that?
* rillian suspects "we'll see" is the correct answer
<adiabatic> "utility standardization"?
<adiabatic> (it's not in the minutes, so I'm not exactly sure what you're asking)
<rillian> adiabatic: agenda is at http://wiki.xiph.org/MonthlyMeeting200403
<adiabatic> o3. Got it.
<adiabatic> Standardization sounds nice; what sorts of things would such an effort entail?
<rillian> In terms of the VorbisFutureRelease that got moved to its own page (thanks arc?)
<rillian> we're still looking for uncompressed 5.1 soundtracks
<rillian> please keep asking
<rillian> I shopped around at the neuro conference last week
<rillian> but nothing helpful
<adiabatic> rillian: Sounds like a nice idea, but we'd probably have enough pandemonium with a switch to "vorbisenc" from "oggenc".
<rillian> "activists don't use dolby surround"
<rillian> Ok, standardization does indeed sound nice
<rillian> if there are no further comments, I move we adjourn
<adiabatic> Where do we move the offlined discussions to?
<rillian> their appropriate channels, or here after the official meeting
<adiabatic> Gotcha.
<rillian> (or the mailing lists)
<adiabatic> Adjourning sounds like a swell idea, then.
<xiphmont> 'kay
<MikeS> Thanks for chairing, rillian. 
<rillian> Ok. thanks *everyone* for coming
<xiphmont> thank you Ralph
<adiabatic> goosedaemon: got log?
<goosedaemon> yeah.
<Arc> oh btw
<goosedaemon> is this where i /log off?
<rillian> adiabatic: can you post the raw?
<adiabatic> rillian: sure
<Arc> everyone should thank goosedaemon, he'll be putting out the next OT.
<adiabatic> goosedaemon: gimme, please.
<rillian> goosedaemon: yep. we're all through here
*** USAGE: /LOG [START|STOP|PAUSE|RESUME]
*** Logging #xiphmeet to I:\#xiphmeet.log
Log Closed: Wed Mar 03 19:23:12 2004