[gst-devel] Daily IRC logs

wim.taymans at chello.be wim.taymans at chello.be
Tue Mar 27 23:07:46 CEST 2001

[06:45] <taaz> i do not understand how oms works
[06:45] <omega_> heh
[06:45] <taaz> i don't know how to stick gst in there
[06:46] <taaz> maybe if i explain this on the livid lists people will stop emailing me random questions ;)
[06:46] <omega_> are livid people asking gstreamer questions?
[06:47] <taaz> no
[06:47] <omega_> what questions are they asking you then?
[06:47] <taaz> do you get people emailing you personally rather than the mailing list?
[06:47] <omega_> sometimes
[06:47] <taaz> i dunno... today i got a "where do i find divx source code for linux" from someone in china with poor english
[06:48] <taaz> usually its just follow ups to my "go look in the wiki" form letter response ;)
[06:48] <omega_> you'd think that if they know about 'divx' they would know where to look...
[06:48] <taaz> yeah.. i said projectmayo or google 
[06:50] <omega_> hmmm, I have some french mail, with the word GRATUIT in it.  spam.
[06:51] <taaz> i'm getting an average of >10 spam emails a day now... its nuts
[06:51] <omega_> ouch
[06:51] <ajmitch> that's quite a few, i only get 2 or 3
[06:52] <taaz> about 40% are about how to spam or spaming products or spam email lists or how to make $10000 a week with spam...
[06:52] <omega_> heh
[06:53] <matth> spam spam spam spam... spam spam-spammmmm spammity-spam!
[06:53] <omega_> <splot>
[06:53] Topic changed on #gstreamer by ChanServ!s at ChanServ: GStreamer: multimedia made easy (http://gstreamer.net)
[06:54] <taaz> ok, so what's the best way to force gst down oms's throat? ;)
[06:54] <omega_> um...
[06:55] <omega_> well, obviously, getting a pipeline playing a DVD straight through is the first major trick
[06:56] <taaz> well yeah...
[06:56] <taaz> but it's not that easy
[06:57] <omega_> ok, then we need a list of things that don't work
[06:57] <taaz> oms has type detection code that instantiates decaps (ie, stream parsers) and codecs in some way i don't understand
[06:58] <omega_> is that really relevant?
[06:58] <taaz> i dunno ;)
[06:58] <omega_> I doubt it
[06:59] <taaz> hmm...
[06:59] <taaz> ok, i guess i don't understand gstreamer either ;)
[06:59] <omega_> um
[07:00] <taaz> how is it possible to push data from oms into a gst pipeline?
[07:00] <omega_> you don't
[07:00] <omega_> rather, define 'push data from oms'
[07:02] <taaz> hmm....
[07:03] <taaz> i suppose i'm confused where data production happens
[07:04] <taaz> not sure how to craete an oms src
[07:04] <omega_> if you move oms over to gst, all the data is handled by gst.  none of the oms code is left, except for the ifo stuff 
[07:04] <taaz> cause it has its own threads and stuff going on now
[07:04] <omega_> you ditch all of that
[07:04] <taaz> well, i'm trying to do this in stages ;)
[07:04] <omega_> you can't, without rewriting it again
[07:04] <ajmitch> hehe
[07:05] <omega_> the architecture of oms replicates gstreamer, but you can't just take half of it, you have to move all the data handling at once
[07:05] <taaz> you sure?
[07:05] <omega_> quite
[07:05] <omega_> how would you suggest the split be done?
[07:06] <taaz> chainsaw.
[07:06] <omega_> um, yeah
[07:06] <omega_> the whole idea is that we already have all the elements one needs to play a DVD
[07:07] <ajmitch> taaz: lots & lots of blood?
[07:07] <omega_> what's missing is the higher-level structures: IFO, menu, etc.
[07:07] <ajmitch> bits flying everywhere?
[07:07] <taaz> i was kinda hoping i could just tap into where the mpeg2 stream gets split up and instead push it into a gst pipeline
[07:07] <omega_> why would you want to go through all that work?
[07:07] <taaz> i thought it would be less work
[07:08] <omega_> I doubt it
[07:10] <taaz> does gst support transport streams?
[07:10] <taaz> i know dent was trying to support some mpeg2 tv broadcasting stuff
[07:10] <omega_> no, though I do have plenty of test streams
[07:11] <omega_> I tried to get some code working, and didn't get very far ;-(
[07:13] <taaz> are there still thread issues in gstmediaplay?
[07:13] <taaz> my little test vob locks up at around the same spot alot
[07:14] <omega_> probably, yes
[07:14] <omega_> incsched has threading fixes, but incsched isn't quite ready yet
[07:16] wtay-sleeping (wim at cable-195-162-214-198.upc.chello.be) left irc: Ping timeout for wtay-sleeping[cable-195-162-214-198.upc.chello.be]
[07:21] ajmitch (me at p47-max11.dun.ihug.co.nz) left irc: http://www.freedevelopers.net
[08:21] <matth> i'm off to sleep
[08:21] Nick change: matth -> matth-sleep
[08:24] wtay (wim at cable-195-162-214-198.upc.chello.be) joined #gstreamer.
[08:58] Nick change: taaz -> taazzzz
[09:04] Action: omega_ needs some sleep too
[09:04] omega_ (omega at omegacs.net) left irc: [x]chat
[09:27] steveb (chatzilla at 212186169160.chello.com) joined #gstreamer.
[09:28] <steveb> hi all
[09:46] steveb (chatzilla at 212186169160.chello.com) left irc: 
[09:48] steveb (chatzilla at 212186169160.chello.com) joined #gstreamer.
[11:18] ajmitch (me at p39-max6.dun.ihug.co.nz) joined #gstreamer.
[11:21] <steveb> hi
[11:21] <ajmitch> hi
[11:23] <ajmitch> wtay: you there?
[12:03] steveb (chatzilla at 212186169160.chello.com) got netsplit.
[12:03] steveb (chatzilla at 212186169160.chello.com) returned to #gstreamer.
[13:06] ajmitch (me at p39-max6.dun.ihug.co.nz) left irc: Read error to ajmitch[p39-max6.dun.ihug.co.nz]: EOF from client
[14:33] steveb (chatzilla at 212186169160.chello.com) left irc: ChatZilla 0.8 [Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; 0.8.1) Gecko/20010323]
[15:34] sienap (synap at ipc379c005.dial.wxs.nl) joined #gstreamer.
[15:34] <sienap> Hi all :)
[15:40] taazzzz (dlehn at left irc: Ping timeout for taazzzz[]
[15:40] taazzzz (dlehn at joined #gstreamer.
[16:12] Nick change: matth-sleep -> matth
[16:13] <sienap> he ;)
[16:13] <sienap> hi!
[16:30] <matth> hi, sienap
[16:31] <sienap> matth how is stuff going ?
[16:37] Uraeus (pirch at joined #gstreamer.
[16:37] <Uraeus> allo
[16:41] <sienap> ur anus
[16:47] <Uraeus> Uraeus slams a 60 lb unix manual on sienap 's head. That gotta hurt... but then again... where there's no sense... there's no feeling!
[16:49] <matth> sienap: stuff's okay -- doing interruptable incremental scheduling sample code
[16:51] <sienap> he
[16:51] <sienap> :)
[17:05] sienap (synap at ipc379c005.dial.wxs.nl) left irc: latur
[18:06] <wtay> yo guys
[18:06] <matth> hey
[18:14] <Uraeus> howdy
[18:51] <Uraeus> wtay: there is still some oss dependancy around, gstparsewav.c: no sys/soundcard.h
[19:27] <wtay> Uraeus: ok thx , fixed
[19:29] <Uraeus> wtay: tell me when you have updated CVS and I will have Gman try compiling again
[19:30] <wtay> Uraeus: done before I replied :-)
[19:38] <Uraeus> thanks
[19:38] <Uraeus> see ya later, going home now
[19:38] Uraeus (pirch at left #gstreamer.
[19:45] iGN_ (ign at login1.simplemente.net) got netsplit.
[19:45] wtay (wim at cable-195-162-214-198.upc.chello.be) got netsplit.
[19:47] iGN_ (ign at login1.simplemente.net) returned to #gstreamer.
[19:47] wtay (wim at cable-195-162-214-198.upc.chello.be) returned to #gstreamer.
[19:51] omega_ (omega at omegacs.net) joined #gstreamer.
[19:53] <wtay> hi
[19:54] <omega_> yo
[19:54] <omega_> I leave for Boise in a little over two hours
[19:58] <wtay> ok
[19:58] <omega_> and we should have some prototype logos available today online
[19:58] <wtay> wow cool
[19:58] <omega_> no idea how we're gonna get 50-60 tshirts to Copenhagen by next friday, but we'll figure something out <g>
[19:59] <wtay> do you know what the numbers in the YUV format 4:2:0 mean?
[19:59] <omega_> Brock's suggestion of iron-on may be necessary for the first dozen for the Saturday meet <g>
[19:59] <omega_> 4:2:0 means that there are luma samples for every pixel..
[19:59] <omega_> and chroma samples for every 4th pixel, in a 2x2 grid
[19:59] <wtay> so what does the 0 mean?
[20:00] <omega_> 'ditto', afaict
[20:00] <wtay> but it's not the same as 4:2:2?
[20:00] <omega_> nope
[20:00] <omega_> I dunno, there's some odd way of doing it...
[20:00] <omega_> I haven't figured out the pattern
[20:00] <wtay> only halved along horiz axis it says...
[20:01] <wtay> me neither...
[20:01] <omega_> 4:2:2 has twice as many chroma samples, they're decimated horizontally but not vertically
[20:01] <omega_> see www.webartz.com/fourcc for definitions
[20:02] <wtay> will try to figure it out...
[20:03] <omega_> and don't forget 4:2:2:4 <g>
[20:03] <wtay> are the T-shirts going to have a GStreamer logo? <g>
[20:03] <omega_> yup, and probably a small rr logo, maybe on the sleeve or something
[20:03] <wtay> omega_: hehe
[20:03] <wtay> nice
[20:03] <omega_> unfortunately, you'll be asleep before I get close to seeing them today
[20:04] <omega_> it's 8pm right now, right?
[20:04] <wtay> yes
[20:04] <wtay> I'll see them in the morning I guess..
[20:05] <omega_> yeah
[20:05] <omega_> my clock changes next weekend I think, while I'm in Amsterdam
[20:05] <omega_> so I've got some funky time-switching goin' on... ;-(
[20:05] steveb (steveb at node1ee09.a2000.nl) joined #gstreamer.
[20:05] <omega_> yo
[20:06] <wtay> hi
[20:06] <omega_> steveb: you're in Amsterdam, right?
[20:06] <wtay> hmm I do you also have DST?
[20:06] <omega_> yes
[20:06] <steveb> ja
[20:07] <wtay> I thought it was a European thing only..
[20:07] <omega_> nope ;-(
[20:07] <omega_> steveb: any ideas where we could meet?
[20:08] <omega_> steveb: and where I should look for a hotel to stay at <g>
[20:09] <steveb> I could get back to you on hotels
[20:09] <omega_> doh.  search for amsterdam linux users group turned up the Portland LUG as the second hit.
[20:09] <omega_> wierd
[20:10] <steveb> do we need whiteboards? network?
[20:11] <omega_> white/black board minimum
[20:11] <omega_> network is nice, power is almost a requirement
[20:11] <steveb> so that rules out cafes
[20:11] <omega_> unless someone brings a large notepad (1m x 0.5m)
[20:12] <omega_> or if there's a hotel I could stay at that has a conference room we could use
[20:13] <steveb> i could look into that - my fiancee books rooms in Amsterdam for many corporate types
[20:13] <omega_> um, does that mean I'm a "corporate type" now? ;-)
[20:14] greg_ (greg at home.sente.pl) joined #gstreamer.
[20:14] <omega_> yo
[20:15] <steveb> well, i'm guessing you'll be at the lower end of the scale :)
[20:15] <omega_> um, ok <g>
[20:16] <steveb> but seriously, how much would you want to pay for the hotel?
[20:18] <omega_> <$100
[20:18] <steveb> and a budget to book a meeting room?
[20:18] <omega_> dunno
[20:18] <omega_> no idea what the range is, so have no basis for budgetting
[20:19] <steveb> ok, i'll get back to you in the next day or two
[20:19] <omega_> ok.  sooner the better, since it's Saturday day and night that this is happening <g>
[20:20] <steveb> oh crap! ok
[20:20] <omega_> didn't realize it was that soon? ;-)
[20:20] <steveb> it sneaks up on you when you're not the one thats traveling
[20:20] <omega_> heh
[20:22] <greg_> excuse me guys, but I've been at this channel for 3 times so far, and TWICE you were talking about hotels (first time "hotel where my father lived ?" or sth) what does it has to do with gstreamer ? ;-)
[20:22] <steveb> ha
[20:22] <omega_> hehehe, I'm heading to europe for two weeks on Friday, to do (at last count) three different GStreamer talks/meets
[20:23] <steveb> an encounter in meatspace
[20:24] <steveb> http://liga.osc.nl/ <-- the amsterdam LUG (with address of meeting space)
[20:24] <omega_> y
[20:24] <steveb> i know nothing about these guys
[20:24] <greg_> omega_: you're famous ! anyway guys: what about asf (and other f..! m$ formats) in gstreamer ?
[20:25] <omega_> apparently avifile has done a lot of asf work, so we'll try to work from that
[20:25] <greg_> omega_: aviplay has some very basic support for this, and i've read on its site, that there is some unmaintained code to marse this format.
[20:25] <omega_> although from what I hear from wtay, the API isn't that friendly
[20:25] <omega_> steveb: liga site doesn't have much contact info
[20:26] <greg_> omega_: see ? we think similar way ;) Probably this doesnt help you much, but wanted to say,
[20:26] <greg_> that you're driving THE PROJECT i've been waiting for soo long. I hate going around all the Internet sites, just
[20:27] ajmitch (me at p27-max7.dun.ihug.co.nz) joined #gstreamer.
[20:27] <omega_> yo
[20:27] <greg_> to get lots of partiali usable programs just to watch some video files. Sth. I get under M$ win with one double-click.
[20:27] <ajmitch> hey
[20:27] <wtay> hi
[20:28] <omega_> yeah, gstmediaplay is getting pretty close to that
[20:28] <omega_> I did a CVS update and build, grabbed an FLI movie, and it *just* *played*.
[20:28] <ajmitch> wtay: what happened to logs being mailed? script broken?
[20:28] <greg_> omega_: is there any usable way to help you ? some bugreports ?
[20:28] <omega_> bugreports are good
[20:29] <omega_> the sf bug tracking mechanism is very underutilized...
[20:29] <wtay> ajmitch: seems like my sendmail is configured wrong...
[20:29] <wtay> ajmitch: and I don't know how to fix it...
[20:29] <ajmitch> wtay: hehe
[20:29] <greg_> omega_: but what can I say about bugreport: "This *.avi doesn't work" ???
[20:29] <omega_> the URL of the avi <g>
[20:30] <wtay> ajmitch: all mails basically baounce back to me with an error...
[20:30] <greg_> omega_: I thought you don't want such bugreports, If you do, I'll find some ;)
[20:30] <ajmitch> wtay: you could switch mailers
[20:31] <wtay> ajmitch: ?
[20:31] <omega_> the more failures we can actually personally reproduce, the faster we can get things fixed
[20:31] <greg_> omega_: anyway seeking from nevest CVS works pretty well.
[20:31] <omega_> hmmm, ok.  news to me <g>
[20:32] <omega_> today and tomorrow I hope to hash out the beginnings of the event system with matth and brentb, after we figure out the thread sync stuff
[20:32] <omega_> then I'll start implementing the event system, and we'll get closer to having *real* seeking
[20:32] <ajmitch> wtay: exim is meant to be good
[20:33] <greg_> omega_: ;)) you're a joker ;)) I've got simple feature request:
[20:33] sienap (synap at ipc379c13e.dial.wxs.nl) joined #gstreamer.
[20:33] <sienap> Hi all
[20:33] <wtay> hi
[20:33] <greg_> You're not narively english-speaking. Me too (I am from Poland). Hope you're planning some support for
[20:33] <wtay> hmm all apps segfault now... will need to restart for libc upgrade to kick in...
[20:34] <ajmitch> wtay: hehe
[20:34] <greg_> text displayed while playing file ? (tranlation texts) ?
[20:34] wtay (wim at cable-195-162-214-198.upc.chello.be) left #gstreamer (reboot :().
[20:35] <greg_> omega_: It would be a nice feature for gstmediaplay
[20:37] _gst_newt_ joined #gstreamer.
[20:38] wtay (wim at cable-195-162-214-198.upc.chello.be) joined #gstreamer.
[20:38] <greg_> omega_: Program may be in English - I don't care - But there are 2 or 3 formats of text files with translations, as supported by MDVDPlayer or VPlayer
[20:38] <omega_> hmmm, there's a bug report I just responded to that I didn't see an mail on
[20:38] <omega_> hmmm, ok
[20:38] <omega_> what you could do is add something to the gstreamer wiki GstWishList with links to the formats in question
[20:39] <wtay> ajmitch: can't do exim, it'll remove sendmail...
[20:39] <greg_> omega_: OK - If you prefer such way.
[20:39] <omega_> and send mail to the list as well, of course
[20:39] <ajmitch> wtay: what's the problem there?
[20:39] <omega_> it's just that the wiki provides a convenient permanent repository for the info, rather than searching through mail archives
[20:39] <wtay> ajmitch: I rely on it to dispatch my mail...
[20:40] <omega_> so a) add it to the wiki and b) send mail to the list with a heads-up (abstract of what you added to the wiki and where)
[20:40] <ajmitch> wtay: exim does the same
[20:41] <wtay> ajmitch: I'll backup my sendmail config then and give it a try...
[20:41] <ajmitch> wtay: exim has a small /usr/sbin/sendmail program that just calls exim ;)
[20:41] <omega_> wtay: 'ddent' might be interested in being our 'docs manager' person
[20:42] <wtay> omega_: oh good
[20:42] <omega_> if you can derive his email address from last night's logs, we can get in touch with him and see if he's still interested
[20:44] <wtay> will do that shortly...
[20:48] <ajmitch> bbl
[20:48] hadess (hadess at pc123-gui14.cable.ntl.com) joined #gstreamer.
[20:49] <omega_> yo
[20:49] <wtay> hi
[20:50] <sienap> hadess!
[20:50] <hadess> hi guys
[20:50] <sienap> how is the itunes clone going ?
[20:53] greg_ (greg at home.sente.pl) left irc: Ping timeout for greg_[home.sente.pl]
[20:53] <hadess> not to bad, been reading libcdaudio's api
[20:53] <hadess> not too bad even
[20:54] greg_ (greg at home.sente.pl) joined #gstreamer.
[20:54] <omega_> greg_: welcome back <g>
[21:02] <ajmitch> hi hadess
[21:02] <hadess> hi ajmitch
[21:04] <ajmitch> wtay: exim working?
[21:05] <wtay> ajmitch: sorta, I can receive mail, now have to check if I can also send...
[21:06] <ajmitch> wtay: good luck ;)
[21:07] <wtay> ajmitch: doesn't work :(
[21:09] <ajmitch> wtay: what happens?
[21:10] <wtay>  lookup of host "sentinel" failed in smarthost router
[21:10] <wtay> I think I solved that part though
[21:10] <wtay> whoo got mail back...
[21:10] <ajmitch> wtay: hmm, what does 'sentinel' refer to though?
[21:11] <wtay> ajmitch: my mail server, I added sentinel.upc.chello.be and it sorta worked now...
[21:12] <wtay> running script for FD now...
[21:14] <steveb> wtay: can you tell me what this bit is doing in gst_osssink_negotiate...?
[21:14] <steveb>   // have we got caps?
[21:14] <steveb>   else if (*caps) {
[21:14] <steveb>     if (osssink->fd == -1) 
[21:14] <steveb>       return GST_PAD_NEGOTIATE_FAIL;
[21:14] Action: ajmitch hopes that some of the stuff said in logs doesn't get thru ;)
[21:14] sienap (synap at ipc379c13e.dial.wxs.nl) left irc: sienap has no reason
[21:15] <wtay> steveb: if the osssink is not opened, the negotiation fails (cannot connect to peer pad)
[21:15] <wtay> ajmitch: nothing happens...
[21:16] <ajmitch> wtay: not mailed out?
[21:16] <wtay> ajmitch: mailed out and no error received...
[21:16] ChiefHighwater (floods at temple-baptist.com) joined #gstreamer.
[21:16] <ajmitch> wtay: that could be goos
[21:16] <ajmitch> good
[21:16] <ajmitch> ello ChiefHighwater
[21:16] <ChiefHighwater> Ello 8-]
[21:16] <wtay> hi ChiefHighwater
[21:17] <ChiefHighwater> everybody gearin up for the weekend?
[21:17] <omega_> ChiefHighwater: I leave for Boise in 1hr
[21:17] <ChiefHighwater> k...i thought you might already be there
[21:17] <wtay> ChiefHighwater: gearing up for GUADEC
[21:17] <omega_> nope, didn't get things planned quick enough
[21:17] <wtay> got my vacation card signed...
[21:18] Action: wtay signed it himself <g>
[21:18] <ChiefHighwater> omega_:hehe...back wed?
[21:18] <omega_> yup
[21:18] <ChiefHighwater> wtay:that makes it easy 8-]
[21:18] <ChiefHighwater> omega_:I'm making you a pile for you to take home
[21:19] <ChiefHighwater> someday
[21:19] <omega_> heh
[21:19] <omega_> kb/mouse
[21:19] <omega_> what else?
[21:19] <ChiefHighwater> so far that and shades
[21:19] greg_ (greg at home.sente.pl) left irc: Ping timeout for greg_[home.sente.pl]
[21:19] <omega_> heh
[21:19] <omega_> and a minidv tape box?
[21:19] <ajmitch> bye
[21:19] <omega_> l8r
[21:19] <ChiefHighwater> havent seen that yet, but the box is still very full
[21:19] <omega_> ok
[21:19] <ChiefHighwater> l8rz ajmitch
[21:20] <wtay> cya ajmitch
[21:20] <ChiefHighwater> can you mail me walmart url @ work address?
[21:20] greg_ (greg at home.sente.pl) joined #gstreamer.
[21:20] ajmitch (me at p27-max7.dun.ihug.co.nz) left irc: http://www.freedevelopers.net
[21:21] <ChiefHighwater> ah, there it is
[21:21] <ChiefHighwater> nm
[21:21] <omega_> also theonly.mp3
[21:21] <ChiefHighwater> kewl
[21:21] <omega_> got cut off, something's wrong with the DV transfer (that one's from the cam)
[21:23] <ChiefHighwater> fun
[21:23] <omega_> but you can hear his setup routine on that one <g>
[21:23] <ChiefHighwater> 8-]
[21:25] <ChiefHighwater> when do i get the video that goes with it? 8-]
[21:25] <omega_> um, later <g>
[21:27] <ChiefHighwater> doot doo doot, doot doo doot doo 8-]
[21:34] <steveb> wtay:  I think I found a problem in osssink - gst_osssink_open_audio isn't being called before gst_osssink_negotiate so nego fails
[21:34] <omega_> wtay: can nego happen in NULL state?
[21:34] <wtay> it shouldn't
[21:34] <steveb> oh
[21:34] <omega_> sounds like that's probably what's happening
[21:34] <wtay> it'll only happen when a peer pad has caps before actually running
[21:35] <omega_> else the nego function needs to silently succeed when connecting with no open sound card, then force a renego when it really does get opened
[21:35] <wtay> steveb: what elements are you using?
[21:35] <steveb> i'm still debugging my sinesrc
[21:36] <steveb> so I shouldn't set caps until in READY state?
[21:36] <wtay> steveb: yup
[21:36] <wtay> or the audiosink should accept anything in NULL state
[21:36] <omega_> so what's the general sequence for connecting things in NULL?
[21:37] <omega_> brb, I need to start packing
[21:37] <wtay> caps negotiation is happening when one of the pads has caps
[21:37] <wtay> regardless the state
[21:38] <wtay> or we could open the /dev/dsp device when doing caps nego...
[21:39] <hadess> i just thought that the same kind of "probbing" could be used to detect if a CD is present
[21:39] <wtay> steveb: I would suggest to set the caps in the _get function for now
[21:40] <wtay> hadess: probably, or at init time...
[21:40] <hadess> wtay: in soundbox, i will need to probe the device every 10 seconds or so, i need a way to know if there is a CD in the drive
[21:41] <omega_> I wish there was a less *stupid* way of waiting for a CD
[21:41] <wtay> hadess: I would suggest letting the app get a disk_present property then
[21:41] <omega_> since the CD drive *does* fire a notification event to the kernel, why can't an app listen for that event?
[21:41] <ChiefHighwater> omega_:no tape box, but a red hat t-shirt
[21:41] <omega_> eh?
[21:41] <omega_> is ChiefHighwater cleaning his office???
[21:42] <hadess> omega_: well, yeah, can do that i guess
[21:42] <hadess> wtay: forget about what i said =)
[21:42] <ChiefHighwater> omega_:a gnome t-shirt (white)
[21:42] <ChiefHighwater> in the box
[21:42] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[21:42] <omega_> heh, but is there some way for the kernel to notify a user app??
[21:42] <Uraeus> ho
[21:42] <wtay> hadess: I don't think the app can listen for a CD insert event...
[21:43] <wtay> howdy Uraeus
[21:43] <Uraeus> what about the method GNOME uses? magicdev or whatever?
[21:43] <hadess> omega_, wtay: not really right now, but it's Linux only
[21:43] <omega_> magicdev just keeps probing
[21:44] <hadess> Uraeus: magicdev checks with the kernel every x seconds, and it produces a lot of kernel messages
[21:44] <hadess> Uraeus: it doesn't actually probe the drive, which makes it quite fast
[21:44] <omega_> which is why I turn magicdev off with a shotgun whenever I see it running
[21:44] <Uraeus> ok, I was just trying to set a new record in joining channel/participating in debate :)
[21:44] <hadess> omega_: there must be a way to do that more cleanly
[21:45] Action: omega_ will be back in a few minutes
[21:45] <wtay> yeah, me too.
[21:45] <Uraeus> omega_: will you send me that mail?
[21:45] <hadess> Uraeus: but i think the first code i will write will be ripped from magicdev
[21:45] <wtay> does anyone want to start a discussion about video properties? :-)
[21:45] Action: steveb woops
[21:46] <steveb> wtay: that worked
[21:46] <wtay> great
[21:46] Action: hadess wants Xv and DGA in the videosink
[21:46] <Uraeus> hadess: don't Linux Mandrake have something else (not magicdev) how does that work?
[21:46] mrrazz (mrrazz at asd-tel-ap01-d11-137.dial.freesurf.nl) joined #gstreamer.
[21:46] <wtay> yo
[21:46] <Uraeus> hi mrrazz
[21:46] <mrrazz> hi all
[21:46] <hadess> Uraeus: i don't know about that, but it's worth checking, thanks
[21:47] Action: Uraeus is a fountain of knowledge :)
[21:47] <hadess> wtay: i have a stupid question... why not basing the videosink on SDL ?
[21:47] <hadess> wtay: or use SDL for the videosink
[21:47] <wtay> hadess: because that would be another videosink :-)
[21:48] <Uraeus> sdlsink?
[21:48] <hadess> wtay: heh, ok, but there's already a lot of code in SDL for video playing (Xv scaling/overlay, DGA, etc.)
[21:48] <wtay> yup
[21:48] Action: wtay is going to pay his taxes... :(
[21:49] <hadess> maybe a coding idea for me then =)
[21:49] Action: hadess adds SDL to the long list of needed libraries for soundbox
[21:50] <Uraeus> hadess: you will make a sdlsink for Gstreamer?
[21:50] <hadess> Uraeus: at some point, and if the videosink wtay is writing doesn't meet the app's needs
[21:53] Action: hadess is away: food
[21:56] zaheer (zaheer at tnt-16-343.easynet.co.uk) joined #gstreamer.
[21:56] <zaheer> yo
[21:56] <zaheer> ill brb
[21:58] <omega_> hadess: keep in mind that SDL doesn't play well with other apps as far as embedding the display into widgets
[21:59] Action: omega_ would much much rather see a collection of a few well-thought-out videosinks for the main output methods
[21:59] <omega_> X, Xv, DGA, fbdev
[21:59] <omega_> add other specialized ones later
[21:59] <zaheer> back
[21:59] Action: omega_ leaves in 20min for Boise
[22:00] <wtay> omega_: are X and Xv separate videosinks IYO?
[22:00] <omega_> hrm, probably not
[22:00] <omega_> and DGA is probably part of the main X sink too
[22:01] <wtay> I currently set up a plain X window and draw X/Xv images into it...
[22:01] <wtay> DGA is something else..
[22:04] <omega_> I have to go now, so I can do the rest of my packing
[22:04] <zaheer> theres also SVGALib if anyone uses that anymore :)
[22:04] <omega_> ick
[22:05] <omega_> see you in a couple hours
[22:05] <zaheer> cya
[22:05] omega_ (omega at omegacs.net) left irc: [x]chat
[22:05] greg_ (greg at home.sente.pl) left irc: Ping timeout for greg_[home.sente.pl]
[22:06] Action: zaheer wonders how long the IBM jdk1.3 download will take...
[22:07] <zaheer> 94 mins seems a bit excessive over 52kbits/sec
[22:08] <zaheer> tho it is 30MB :(
[22:08] <hadess> zaheer: i can d/l for you if you want, and post it on my website
[22:09] <zaheer> hadess: i d/led it to university earlier today
[22:09] <zaheer> i forgot to copy it on my laptop be4 going home :)
[22:09] <zaheer> bit stupid i know :P
[22:10] Action: zaheer needs to use an XSLT prog tonight...
[22:10] greg_ (greg at home.sente.pl) joined #gstreamer.
[22:12] <wtay> zaheer: no Amsterdam for me...
[22:12] <zaheer> i heard, wtay...
[22:13] <zaheer> i even helped you beat me at snooker by screwing up my heel while playing golf :P
[22:13] <wtay> bad?
[22:13] <zaheer> its very sore, i will say
[22:14] <zaheer> it'll take time to heal...
[22:14] <zaheer> better heal by sunday, coz im playing a tournament then
[22:14] <wtay> hmm yes
[22:14] <wtay> I would have beaten you anyway <g>
[22:14] greg_ (greg at home.sente.pl) left irc: Ping timeout for greg_[home.sente.pl]
[22:15] <zaheer> withdrawing = lose :)
[22:15] <wtay> I guess..
[22:15] <zaheer> 6 hr drive each way or total?
[22:15] <wtay> total
[22:16] <wtay> my GF doesn't like cars at all, so..
[22:16] <zaheer> only a 3hr drive then :)
[22:16] <wtay> 3+3=6 <g>
[22:16] Action: zaheer hands wtay a ferrari maranello....
[22:16] <wtay> I got a very slow car too...
[22:17] <wtay> R5 :)
[22:17] greg_ (greg at home.sente.pl) joined #gstreamer.
[22:18] <zaheer> it'll be about 2 hrs each way for me...1 hr at airport other hr in plane
[22:18] <matth> doh!  i sent an email to gst-devel when i meant to send it to erik...
[22:18] Action: matth is ashamed
[22:18] <zaheer> we lose a whiteboard too :)
[22:18] <wtay> matth: hehe, wanna read :)
[22:18] <zaheer> matth: you didnt incriminate yourself did ya?
[22:19] <wtay> secret plans...
[22:20] <zaheer> thread sync stuff :)
[22:20] <matth> nah -- i would have offered a bit more explanation up front going to gst-devel...
[22:20] <wtay> nice stuff
[22:20] <matth> questions / comments are welcome...
[22:21] <matth> i had hoped to catch omega before he left for the plane :-(
[22:21] Action: wtay is yelling at his e-banking proggy..
[22:21] <zaheer> he said he'll be back in a cpla hours
[22:22] <matth> yup: this would have been something for him to look at on the plane
[22:22] Action: zaheer smiles
[22:22] <zaheer> you sure he wont be looking at the chick in 8B rather than some thread sync code? :)
[22:23] steveb (steveb at node1ee09.a2000.nl) left irc: [x]chat
[22:23] <wtay> zaheer: don't tell me you would prefer a chick instead of thread code :-)
[22:24] <matth> zaheer: probably not many interesting chicks on this flight ;-)
[22:24] Action: zaheer forgot this is all logged to gst-devel so will say "no comment"
[22:25] <wtay> if the mail script would work...
[22:25] <mrrazz> Date: Tue, 27 Mar 2001 21:32:13 +0200
[22:25] <mrrazz> From: wim.taymans at chello.be
[22:25] <mrrazz> Tue Mar 27 22:33:58 CEST 2001
[22:25] <mrrazz> damn :)
[22:25] <zaheer> CEST or CET?
[22:25] <wtay> CET
[22:26] <mrrazz> hmm sorry
[22:26] <wtay> ?
[22:26] <mrrazz> there's a difference of a hour:)
[22:26] <mrrazz> not 30 seconds 
[22:26] <wtay> mrrazz: DST remember? <g>
[22:26] <zaheer> yep
[22:26] <zaheer> +0200 is fine for CET at moment
[22:27] <mrrazz> wtay: :)
[22:27] <zaheer> brb
[22:28] sienap (synap at ipc379c06d.dial.wxs.nl) joined #gstreamer.
[22:28] <sienap> Hi all :)
[22:28] <mrrazz> oh btw .. i think i found a little problem with cothreading elements with only 1 pad  
[22:29] <zaheer> hiya sienap
[22:29] <sienap> Zaheer!
[22:29] <sienap> mkrraazzz
[22:29] <mrrazz> or i don't understand it probably  :)
[22:29] <zaheer> wtay: which elem
[22:29] <zaheer> aargh
[22:29] <zaheer> wtay: which element should i "copy" regarding caps for the stereo2mono and ulaw filters?
[22:29] <mrrazz> well in this case with the xmms plugin 
[22:30] <sienap> mrrazz
[22:30] <sienap> wake up
[22:30] <sienap> and don't ignore me
[22:30] <sienap> you gay person :
[22:30] <mrrazz> the gst_bin_iterate_func always sets the flags of cothreading elements to COTHREADING_STOPPING
[22:31] <mrrazz> gay ?
[22:31] Action: zaheer reminds sienap that the newt is watching :)
[22:31] <mrrazz> big gay al ? 
[22:31] <sienap> ooh sorry :)
[22:31] <sienap> i hate newt ;)
[22:31] <sienap> can impress my feelings freely now :)
[22:31] <zaheer> gay - happy
[22:31] <sienap> hej mrrazz look at query
[22:31] matth (math at qwest.dsplinux.net) left irc: Ping timeout for matth[qwest.dsplinux.net]
[22:31] <wtay> mrrazz: the cothread stopping is normal
[22:32] <mrrazz> wtay: why ? 
[22:32] <mrrazz> what's the reason for it
[22:32] <wtay> zaheer: testsuite/capsnego/converter2.c
[22:32] <wtay> mrrazz: a way for the element to be able to stop when instructed to do so
[22:33] <zaheer> wtay: thx...., once done should i commit my elements?
[22:33] <wtay> zaheer: sure
[22:33] <zaheer> ok
[22:34] <mrrazz> wtay: yeah, but it will always stop 
[22:34] <wtay> zaheer: care for an agrument about the standard video properties? <g>
[22:34] <mrrazz> then
[22:35] <wtay> mrrazz: sure, but the managing element will call it again
[22:35] <zaheer> wtay: my video knowledge is "sh*t" so i dont think i should open my mouth there :)
[22:35] <wtay> good, in a sense :)
[22:36] <sienap> wtay how is xvideosink going ?
[22:36] <wtay> mrrazz: also be aware that the loopfunc is started over again, so no initialisation in the loopfunc
[22:36] <wtay> sienap: trying to do the caps stuff now
[22:36] <sienap> he neat..
[22:36] BBB (BBB at ucu-105-116.ucu.uu.nl) joined #gstreamer.
[22:36] <wtay> yo
[22:36] <sienap> mrrazz care to look in the query ?
[22:36] <zaheer> as far as I am concerned development wise, audio will be the only media i will be interested in working on :)
[22:36] <sienap> hej BBB
[22:36] <BBB> hey all :)
[22:36] <mrrazz> wtay: oooh ok .. didn't knew that :)
[22:36] <BBB> how late is the party saturday?
[22:37] Action: BBB also needs to know something...
[22:37] <zaheer> BBB: I want to know also :)
[22:37] <BBB> if I have a pointer-to-char, how do I unref the char?
[22:37] <sienap> hej BBB
[22:37] <sienap> mede hollander ;)
[22:37] <wtay> mrrazz: I noticed you did that too in the flic plugin..
[22:37] <BBB> hoi sienap :)
[22:37] <mrrazz> wtay: indeed
[22:37] <BBB> ook-nl-er
[22:37] <sienap> he as mrrazz
[22:37] <sienap> as wtay :)
[22:37] <wtay> BBB: unref the char?
[22:37] <sienap> stikt hier van de nederlanders ;)
[22:37] <BBB> wtay: yes
[22:37] <wtay> ik ben geen nederlander!
[22:37] <BBB> wtay: unref (char); doesn't work
[22:38] <BBB> unref() cannot be found, the compiler complaisn :(
[22:38] <wtay> BBB: what's unref?
[22:38] <BBB> wtay: to free the memory
[22:38] <zaheer> unref as in decrement the reference count?
[22:38] <wtay> you can't free a char...
[22:38] <BBB> aaaaaaaaah
[22:38] <wtay> a char is just like an int
[22:38] <BBB> but I make a million of them
[22:38] <wtay> where?
[22:38] <BBB> in my scene detector :)
[22:38] <wtay> not with alloc(1) I hope <g>
[22:39] <sienap> wtay he belg dan :)
[22:39] <zaheer> a pointer to a char could be pointing to a massive array of char's
[22:39] <BBB> if I have more than like 2000 frames the memory will get fscked up - 100% swap and mem and a crash
[22:39] <BBB> zaheer: no, it's a guchar of a jpeg image/gdk-pixbuf
[22:39] <zaheer> whereby you just wanna g_free(ptr) when finished with the ptr
[22:39] <BBB> that is a bunch of 384x288x3 chars :)
[22:40] <BBB> per image
[22:40] <sienap> in this case
[22:40] <BBB> g_free?
[22:40] <BBB> kewl
[22:40] <BBB> let me try that
[22:40] <wtay> not on the char!
[22:40] <wtay> on the buffer instead
[22:40] <sienap> try memprof for finding mem leaks
[22:40] <zaheer> do a g_free(ptr) where ptr is the pointer to the guchar of the first pixel
[22:41] <BBB> oops
[22:41] <BBB> core dump :}
[22:41] Action: BBB is doing something wrong
[22:41] <zaheer> and see if it segfaults :)
[22:41] <mrrazz> wtay: i should use a sort of state checking then ? 
[22:41] <zaheer> who does the g_new or g_alloc ?
[22:41] <wtay> mrrazz: when?
[22:41] <mrrazz> wtay: the same that's normally used in chainsfunc and so
[22:41] <BBB> zaheer: nobody :P
[22:41] <BBB> dammit
[22:42] <mrrazz> wtay: in a loopfunc 
[22:42] <wtay> mrrazz: you don't normally do state checking in the chainfunc...
[22:42] <zaheer> then you dont need to free it....
[22:42] <wtay> mrrazz: only the COTHREAD stopping
[22:42] <BBB> zaheer: I do.... memory is getting very full very quickly if I don't
[22:42] <sienap> mem leak for sure :)
[22:42] <sienap> sure you are not loosing pointers to the magic black hole ?
[22:42] <zaheer> BBB: you using a gtk-pixbuf object?
[22:43] <mrrazz> wtay: well with state checking i mean .. to know where you're at the moment
[22:43] <mrrazz> wtay: like, reading the header, parsing chunks
[22:43] <BBB> zaheer: that's a part - I already gdk_pixbuf_unref() it
[22:43] <wtay> mrrazz: if you use cothreads you don't have to..
[22:43] <sienap> BBB remallocing a pointer will not free the old one
[22:43] <sienap> however you won't be able to access it anymore
[22:43] <BBB> sienap: I noticed :P
[22:43] <BBB> but I am freeing it before it gets re-allocated
[22:43] <sienap> you just do a new mem locating
[22:44] <sienap> the old one stays reserved ofcourse
[22:44] <sienap> not that you can use it anymore..
[22:44] <sienap> BBB strange..
[22:44] <wtay> mrrazz: vorbisdec is a good example of cothread use.
[22:44] <sienap> so mem is going to expand every frame you load ?
[22:45] <mrrazz> wtay: ok i'm going to take a look at it :)
[22:45] <BBB> sienap: yups
[22:45] <sienap> he have fun mrrazz :)
[22:45] <zaheer> BBB: im just checking it out for ya
[22:45] <BBB> let me try a somewhat different architecture :)
[22:45] <sienap> bbb then you are losing mem somewhere for sure :)
[22:46] <BBB> yups
[22:46] <BBB> I know
[22:46] <BBB> if I try to run it, it goes well for the first few 100 frames - but the rest loads mem really quickly - a few percent per second
[22:46] <BBB> so I need to Ctrl-C it or crash :)
[22:47] <zaheer> wtay: do we have a filter converting big endian to little endian and vice versa?
[22:47] <BBB> imagine a movie of 2 minutes = 120x25 = 3000 frames :(
[22:47] <wtay> zaheer: nope
[22:47] <sienap> BBB try using memprof
[22:47] <wtay> zaheer: is that worth a filter?
[22:47] <sienap> if you can find out where the losses are
[22:47] <zaheer> wtay: ill do that too then...
[22:48] <BBB> what is memprof?
[22:48] <sienap> a program
[22:48] <zaheer> wtay: you saying every filter should deal with both types?
[22:48] <BBB> sienap: doh!
[22:48] <sienap> (gnome program)
[22:48] <sienap> really handy :)
[22:48] <BBB> where do I get it?
[22:48] <sienap> ehm he ;)
[22:48] <sienap> think it is at the ximian download section
[22:48] <BBB> moeilijke vragen he? :P
[22:48] <wtay> zaheer: uhm no I guess... it's just a silly filter...
[22:48] <sienap> mwha :)
[22:48] <zaheer> BBB: you use gdk_pixmap_new ?
[22:49] <sienap> going to get some more drink and then sleep..
[22:49] <zaheer> wtay: i know...maybe an inline in the core should do it....if its needed
[22:49] <BBB> zaheer: yes
[22:50] <BBB> hey kewl!
[22:50] <BBB> I solved it!
[22:50] ajmitch (me at p9-max3.dun.ihug.co.nz) joined #gstreamer.
[22:50] <ajmitch> hi
[22:50] <BBB> two extra unref()s and it's okay now :)
[22:50] <BBB> I think
[22:50] <BBB> hey ajmitch
[22:51] <wtay> hi ajmitch
[22:51] Action: BBB is running scene detection on a movie of about 20 minutes now
[22:51] <zaheer> BBB: that shouldnt be needed....
[22:51] <BBB> let's see how that works
[22:51] <BBB> but it rocks :P
[22:51] <zaheer> BBB: something must be increasing refcount without unreffing
[22:51] <sienap> hop... now have a bunch of snacks >:) let's satisfy this evil hunger
[22:52] <BBB> Sysinfo: Linux 2.4.2 | CPU: Pentium II (Deschutes) @ 400 mhz | Res: 1024x768 pixels /w 24 bit gfx | Mem usage: 44.1/124.0 mb (35%) | Swap usage: 55.7/125.5 mb (44%) | Disk usage (/dev/hda): 8.5/12.7 gb (66%) | Local Time: 10:57pm Uptime: 11:27 hours | Users: 3 Load: 0.81 | X-chat: 1.6.4
[22:52] <BBB> 35% mem use
[22:52] <BBB> that's okay :)
[22:52] <sienap>  11:01pm  up 17 days,  4:59,  5 users,  load average: 0.00, 0.00, 0.00
[22:52] <ajmitch> BBB: damned good
[22:52] <sienap> ooh this is my dumb machine
[22:52] <sienap> other machine is 83 days :)
[22:52] <ajmitch>  08:58:29 up  9:45,  8 users,  load average: 3.99, 3.65, 3.36
[22:52] <ajmitch> ;)
[22:52] <BBB> my server is only 56 days :P
[22:52] <BBB> I installed BSD on it 56 days ago :P
[22:52] <ajmitch> *never* run strace on a running X server!
[22:52] <BBB> heheheheheheh
[22:52] <sienap> he my server is 17 days and my workstation 83
[22:53] <sienap> my modem sometimes act strange
[22:53] <sienap> and then i try to hotswap it what verry often results in a hang :)
[22:53] <ajmitch> my server was about 40 days or so until i restarted it last night
[22:53] <BBB> ?
[22:53] <BBB> ajmitch: dumb you, never restart a server
[22:53] <sienap> he yeah i am strange
[22:53] <BBB> :)
[22:53] <sienap> i hotswap hardware :)
[22:53] <sienap> he servers with high uptime often means bad maintance :)
[22:54] Action: BBB is at 2500 frames now - still without problems
[22:54] <ajmitch> BBB: um, i sorta had to, to go from 2.2 to 2.4
[22:54] <BBB> 2500 = 100 seconds = 2 minutes :(
[22:54] <BBB> the speed of my scene detector is sooooooooooooooooos slooooooooooooooooooooooooow
[22:54] <sienap> don't know
[22:54] <ajmitch> BBB: server here==dumb box with hard drive & net card
[22:54] <BBB> this movie is 20 minutes!!!!!
[22:54] <sienap> but 100 seconds are NOT 2 minutes :)
[22:54] <BBB> sienap: :P
[22:54] <BBB> kind of
[22:54] <sienap> he only 1/3 away from it >:)
[22:55] <zaheer> BBB: what time you gonna be in Amsterdam?
[22:55] <sienap> in the last minute that is >:)
[22:55] <BBB> zaheer: I live close so I'll be there when you're there
[22:55] <sienap> he i am blaating to much again
[22:55] <BBB> just, one thing
[22:55] <BBB> I have a big party this friday
[22:55] <Uraeus> omega_?
[22:55] <BBB> it will be until 7 am saturday
[22:55] <BBB> so don't meet too early plz
[22:55] <BBB> :P
[22:55] <zaheer> Uraeus: he'll be back in a cpla hrs i think
[22:55] <Uraeus> zaheer: thanks
[22:56] <zaheer> BBB: last time i heard 12 noon till 6pm is what it was gonna be
[22:56] <sienap> zaheer if he isn't on the plane to the netherlands :)
[22:56] <sienap> and i think he may be :)
[22:56] <zaheer> sienap: he's going to Boise
[22:56] <sienap> ooh :)
[22:56] <sienap> that makes difference
[22:56] <zaheer> no idea where that is :)
[22:56] <BBB> I'm at 5000 frames now
[22:56] <sienap> he :)
[22:56] <BBB> hmz
[22:56] <BBB> long live slowness
[22:56] <BBB> this is worse than windoze
[22:57] <sienap> does it work property..
[22:57] <sienap> properly even :)
[22:57] <BBB> sienap: it works kind of.......
[22:57] <BBB> I tested it on some short movies and the results were okay
[22:57] <wtay> what does it do?
[22:57] matth (matth at qwest.dsplinux.net) joined #gstreamer.
[22:57] <BBB> it gets a movie and makes scene entries in an editlist
[22:57] <ajmitch> wtay: you never got the logs sent? ;)
[22:57] <wtay> oh cool
[22:57] <BBB> each entry (so each scene) will be represented my one image in my editor
[22:58] <zaheer> 55% complete only :( ....gonna go watch some telly in meantime
[22:58] <sienap> he cool
[22:58] <wtay> ajmitch: I sent them without errors :)
[22:58] <BBB> that looks really kewl :)
[22:58] <sienap> bbb :)
[22:58] <zaheer> bbs
[22:58] <sienap> bbb sounds really cool
[22:58] <BBB> if it works :P
[22:58] <BBB> I need to make it a little more complicated
[22:58] <sienap> we need some kind of gstreamer software map really soon :)
[22:58] <BBB> it is too simple now
[22:58] <BBB> if change > 20 then new scene
[22:58] <sienap> he yeah but it is a nice start :)
[22:59] <sienap> ehm change  can be everything
[22:59] <BBB> but this should be more like if change in change rises by 10 suddenly and then gets back to normal, then new scene
[22:59] <BBB> if it stays higher, no new scene
[22:59] <BBB> like that
[22:59] <sienap> define it a little :) will ya
[22:59] <BBB> change = difference between two sunsequent frames
[22:59] <BBB> :P
[22:59] <sienap> you compare different frames
[22:59] <BBB> subsequent
[22:59] <sienap> woow heavy :)
[22:59] <BBB> (opeenvolgend)
[22:59] <BBB> yes
[23:00] <sienap> how the heck are u doing that?
[23:00] <BBB> hehe, my secret :P
[23:00] <sienap> he GPL .. :)
[23:00] <BBB> it will be GPL'ed when it is ready :P
[23:00] <wtay> sienap: just substract the luma components...
[23:00] <sienap> he GPL the knowledge..
[23:00] <BBB> problem is, it is just tooooooooooooooooooooooooooooooo slow
[23:01] <BBB> I need to figure out how pinnacle or adobe is doing that
[23:01] <sienap> wtay aah ok now i realllly understand :)
[23:01] <BBB> they do the same thing but their stuff is ready for movies of 30 seoncds in like 10 seconds
[23:01] <wtay> BBB: basically an mpeg encoder does the same thing on the MBs
[23:01] <BBB> I use 30 seconds for 10 seconds of movie!!!
[23:01] <wtay> use MMX asm
[23:01] <BBB> it does, wtay?
[23:02] Action: BBB never looked at mpeg2enc yet
[23:02] <wtay> you are working in the YUV colorspace right?
[23:02] <BBB> no, RGB
[23:02] <BBB> or is it YUV?
[23:02] <wtay> ah ok
[23:02] <BBB> no
[23:02] <BBB> I think it's RGB
[23:02] <wtay> YUV is better suited I think
[23:02] <BBB> but that shouldn't make a difference
[23:02] <wtay> 3x speedup :)
[23:02] <BBB> the problem is, if I could compare jpegs...... it would be okay
[23:02] <BBB> but I cannot compare jpegs
[23:02] <BBB> because they're encoded
[23:03] <BBB> I need to compare unencoded frames for my app
[23:03] <wtay> yes
[23:03] <BBB> anybody have a clue on how to compare jpegs?
[23:03] <wtay> can try to work with the raw DCT coefs
[23:03] <wtay> but I dunno how that works...
[23:03] <BBB> you mean the raw jpeg data?
[23:03] <wtay> yes
[23:03] <BBB> I tried
[23:03] <BBB> no luck there :(
[23:03] <wtay> ok
[23:04] <BBB> it sees everything as different so all "change" values will be maximal
[23:04] <wtay> just decode to YUV and compare the Y planes will work
[23:04] <BBB> :(
[23:04] <BBB> true........
[23:04] <BBB> I shouldn't need to compare all......
[23:04] <wtay> you can use subsampling too
[23:04] <BBB> you mean only compare the first 100x100 pixels?
[23:04] <BBB> could do.....
[23:05] <wtay> look at mpeg2enc it does that
[23:05] <wtay> the never version does anyway
[23:05] <wtay> s/never/newer
[23:05] Action: BBB has the newer version :P
[23:05] <BBB> I'm supposed to have everythine newest because I help with the MJPEG stuff :)
[23:05] <ajmitch> wtay: the logs never made it onto the list here :(
[23:05] <wtay> they compare a 16x16 block by first comparing a 4x4 subsampled piece I think
[23:06] <wtay> ajmitch: dunno what went wrong I don't have an error message...

More information about the gstreamer-devel mailing list