From: IRC <wt...@us...> - 2003-04-15 05:42:40
|
******************************************************************* [03:24] <Company> ds: you can now have fun with audioconvert - i expect it to break on anything not x86 ;) [03:24] <ds> I'll play with it on powerpc [03:25] <ds> I wouldn't recommend doing rate conversion [03:28] <Company> i want to, because we have some stupid soundcards that don't support 44100 hz with 2 channels [03:29] <Company> iain's [03:29] <Company> and i thought it would be nice to do this with an all-in-1 plugin [03:30] <Company> so that i.e. the player can insert an audioconvert directly after spider and play any audio on any soundcard [03:31] <ds> Company: we have audioscale [03:31] <ds> although it's a simple wrapper around libaudioresample [03:33] <Company> yeah, i know [03:33] <Company> maybe i'll give it a property to diable that stuff, but it isn't hard to add anyway [03:40] Marsupilami23 (~Mar...@15...) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!" [03:46] Marsupilami23 (~Mar...@15...) joined #gstreamer. [03:46] Marsupilami23 (~Mar...@15...) left irc: Client Quit [04:05] Marsupilami23 (~Mar...@15...) joined #gstreamer. [04:17] hadley (~hadley@65.112.179.156) joined #gstreamer. [04:19] Company (~Co...@p5...) left irc: Remote closed the connection [04:34] Marsupilami23 (~Mar...@15...) left irc: Remote closed the connection [04:35] Marsupilami23 (~Mar...@15...) joined #gstreamer. [04:42] <taaz> woot! [04:42] Action: taaz <- now a debian developer ;) [04:43] <walters> sweet! congratulations! [04:43] <ds> 2*woot! [04:44] <walters> taaz: how's it feel? :) [04:45] <taaz> grr... hopefully i can dig through enough work so i can actually get around to fixing the rather high bug count on gst packages... maybe tuesday i'll have time [04:46] <walters> sorry for increasing it :/ [04:46] <taaz> walters: email address count ++ [04:46] <walters> taaz: hehe. i looove collecting email addresses :) [04:46] <taaz> did 0.6.1 get released yet? [04:47] <walters> taaz: i forget, are you dlehn@d.o now? [04:47] <taaz> i haven't had time to read this channel last few days... [04:47] <walters> taaz: not yet i don't think [04:47] <taaz> walters: that's the theory, i should send myself some email ;) [04:48] <walters> taaz: btw, if you log in at db.debian.org you can change all your settings [04:48] <walters> taaz: be sure to put in your lat/long :) [04:48] <taaz> i'm not sure where i can get an accurate ICBM location [04:49] <taaz> i should borrow a GPS unit from someone i guess [04:49] <walters> mmm, i forget how i did it. [04:50] <taaz> well.. i can get close. local GIS info is online. but i know nothing about lat/long stuff so i'm not sure how accurate the numbers i read off that are [04:51] <taaz> a few of the gst deb bugs are related to gstreamer-runtime (gst-register/compprep) requireing a scheduler, and it doesn't depend on gstreamer-core so things break somehow [04:51] <taaz> i'm not sure why that is the case though [04:51] <taaz> it used to work without that dep.. i'm not sure what changed [04:52] <taaz> anyone know? does gst-register really require a scheduler? [04:57] <taaz> heh... took 815 days to become a dd ;) [05:00] <taaz> i was just wondering if those bugs were something that could be fixed in 0.6.1. no time next couple days to look into it though... [05:03] <ds> A FEW [05:03] Action: ds is surprised by the caps [05:18] Marsupilami23 (~Mar...@15...) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!" [05:30] Marsupilami23 (~Mar...@15...) joined #gstreamer. [05:34] chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) left irc: "leaving" [05:53] Action: Marsupilami23 is away: Going to the gym to work off most of my stress! THANKS ashmodai!!! [05:58] chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) joined #gstreamer. [06:14] seth__ (~hadley@65.112.179.174) joined #gstreamer. [06:16] hadley (~hadley@65.112.179.156) left irc: Read error: 113 (No route to host) [06:34] Action: Marsupilami23 is back (gone 00:41:03) [06:41] Marsupilami23 (~Mar...@15...) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!" [07:11] ChrisHJW (v-167@12.160.18.11) joined #gstreamer. [07:12] ChrisHJW (v-167@12.160.18.11) left irc: Read error: 104 (Connection reset by peer) [07:12] seth__ (~hadley@65.112.179.174) left irc: "Client exiting" [07:26] ChrisHJW (v-167@12.160.18.11) joined #gstreamer. [07:26] ChrisHJW (v-167@12.160.18.11) left irc: Read error: 104 (Connection reset by peer) [07:29] ChrHJW_log (~cwi...@p5...) joined #gstreamer. [07:42] Marsupilami24 (~Mar...@15...) joined #gstreamer. [07:48] Marsupilami24 (~Mar...@15...) left irc: "Client Exiting" [07:53] Marsupilami23 (~Mar...@15...) joined #gstreamer. [07:55] BBB (~rb...@01...) joined #gstreamer. [08:00] BBB (~rb...@01...) left irc: "Client Exiting" [08:14] RebelHeart (~sagar@61.11.23.199) joined #gstreamer. [08:15] RebelHeart (~sagar@61.11.23.199) left #gstreamer. [09:21] andy_t (~cha...@pp...) joined #gstreamer. [09:25] walters (walters@216.226.142.129) left irc: "I like core dumps" [09:37] chrisime (~chrisime@pD9E8C9FE.dip0.t-ipconnect.de) left irc: "Client exiting" [09:42] thomasvz (~th...@71...) left irc: Read error: 113 (No route to host) [09:56] harobed (~ha...@ha...) left irc: Read error: 104 (Connection reset by peer) [09:56] harobed (~ha...@ha...) joined #gstreamer. [10:08] thomasvz (~th...@ca...) joined #gstreamer. [10:08] Nick change: thomasvz -> thomasvs [10:20] BBB (~rb...@ph...) joined #gstreamer. [10:27] dolphy (~dolphy@213.170.47.89) joined #gstreamer. [10:30] <dolphy> morning [10:33] <harshy> whoa gcc broke in sid [10:45] <BBB> why am I not surprised? :) [10:46] Action: dolphy is building prerelease [10:48] <harshy> guess gdm dosnt get fixed today [10:56] Titanium (~giorgio@213.170.47.89) joined #gstreamer. [11:00] harshy (~ha...@dh...) got netsplit. [11:00] agonyzer (~ago...@12...) got netsplit. [11:00] tsiar ([m8r...@dh...) got netsplit. [11:00] Vakor (ms...@c1...) got netsplit. [11:01] harshy (~ha...@dh...) returned to #gstreamer. [11:01] Vakor (ms...@c1...) returned to #gstreamer. [11:01] tsiar ([m8r...@dh...) returned to #gstreamer. [11:01] agonyzer (~ago...@12...) returned to #gstreamer. [11:24] vishnu (~joshua@61.11.23.205) joined #gstreamer. [11:26] pb_ (~pb...@ds...) joined #gstreamer. [11:28] <vishnu> CVS audioconvert doesn't compile [11:34] Company (~as...@rz...) joined #gstreamer. [11:38] <dolphy> vishnu: repeat what you just said [11:38] <dolphy> :) [11:38] <BBB> CVS HEAD or 0_6? [11:39] Action: Company starts building 0_6 cvs on solaris... [11:39] <vishnu> CVS HEAD [11:39] <vishnu> that 0.7 thingy [11:39] <vishnu> yo BBB, i'm going to fix ffmpeg [11:39] <vishnu> tell me how to do it [11:39] Action: vishnu is ready for pain [11:40] <BBB> uhm... fix the mimetypes, implement capsnego, etc...? [11:40] <BBB> those things totally don't exist in there, afaik [11:40] <vishnu> yah [11:40] <vishnu> i need to do play->encoding with exact start & stop points [11:42] <Company> good luck [11:42] <Company> the framework is there, we just miss the plugins to do that ;) [11:43] <vishnu> Company: yah, we need ffmpeg to work [11:43] andy_t (~cha...@pp...) left irc: Read error: 113 (No route to host) [11:43] <Company> yeah [11:44] <Company> but i'll get audio going first [11:44] <dolphy> vishnu: Titanium might be interested to help you [11:44] <Company> because audio is more user visible in gnome right now [11:44] <vishnu> Company: yah, you do audio. gstreamer needs all the help it can get [11:45] <Company> gawd, i have elite 2nd semester leet linux guys sitting here :/ [11:46] <pb_> heh [11:48] <Company> they're bragging about thinkgeek t-shirts now [11:48] <vishnu> Company: look on the bright side, at least their not windows geeks [11:49] <Company> windows geeks always ffel inferior to linux people at least... [11:49] <Company> s/ffel/feel/ [11:49] <Company> "linux is no operating system, it's a culture" - omg [11:50] <Company> emacs vs vi now [11:51] <vishnu> Company: don't worry, give them a few years and they'll start writing device drivers [11:52] <Company> i sincerely hope they even know what "C" is [11:52] <Company> they're probably only knowing java, delphi or c# [11:53] <vishnu> Company: add daily pizza, flourecent light, mix well and wait 3 years [11:55] andy_t (~cha...@pp...) joined #gstreamer. [11:55] <Company> people talking about what they talk and in that tone aren't hackers [11:56] <Company> they might qualify as zealots, but not hackers [11:56] <Company> but maybe they'll improve [11:58] <Company> now they're complaining that finger doesn't work for hotmail [11:58] <Company> they're _really_ cool dudez [11:59] <Company> thomasvs: ping [12:10] foser (d0...@22...) joined #gstreamer. [12:16] <BBB> Company: is that at the university or so? [12:17] <BBB> you sound like... there's some kind oh high school group "look we've installed linux we're cool (mama, my <insert favourite game> doesn't work! help!)" dudes [12:17] <BBB> s/dudes/weirdos/ [12:19] <dolphy> BBB: i still have the building issues with libgstgconf [12:19] <Company> yes, it's at the uni [12:19] <dolphy> BBB: with 0.6.0.2 [12:19] Action: Company should be at some meeting since 5 minutes, stupid bugs [12:20] <BBB> what building issues? [12:20] walken (fo...@12...) joined #gstreamer. [12:20] <walken> boink [12:21] <BBB> welcome ;) [12:21] <walken> hehe [12:21] <dolphy> BBB: libtool tries to link with /usr/lib/libgstgconf-0.6 instead of /opt/gstreamer/lib/libgstgconf-0.6 [12:22] <BBB> erm... [12:22] <BBB> what's your $PATH? [12:22] <BBB> (echo $PATH) [12:23] <dolphy> dolphy@dolphy:~/workdir/gst-player$ echo $PATH [12:24] <dolphy> argh shit :) [12:24] <dolphy> it is /opt/gnome2/bin:/opt/gnome2/sbin:/usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/games:/opt/gstreamer/bin [12:24] <BBB> change it [12:24] <BBB> so that /opt/gstreamer/lib comes first [12:24] <BBB> oh [12:24] <BBB> erm [12:24] <BBB> LD_LIBRARY_PATH [12:24] <BBB> I need that one [12:24] <dolphy> lol :) [12:24] <BBB> not PATH [12:24] Action: BBB is weird [12:24] <BBB> I'm asleep [12:24] <BBB> apologies for that ;) [12:25] <dolphy> LD_LIBRARY_PATH is empty [12:25] <thomasvs> dolphy: try setting it to LD_LIBRARY_PATH=/opt/gstreamer/lib and rerun [12:25] <BBB> I was just going to say what thomasvs said, but nevermind then [12:25] <BBB> ;) [12:25] <Company> maybe it takes the wrong .pc file? [12:25] <Company> cyas [12:26] <thomasvs> no, it's a runtime linking issue I think [12:26] Company (~as...@rz...) left irc: Remote closed the connection [12:26] Action: BBB thinks so too [12:26] <dolphy> rerun the whole autogen ? [12:26] <BBB> no [12:26] <BBB> just export LD_LIBRARY_PATH=/opt/gstreamer/lib [12:26] <BBB> and then run ldd again [12:26] <dolphy> cause simply rerunning make gives the same : -ldl /usr/lib/libgstgconf-0.6.so [12:26] <BBB> and check if it's still bad [12:27] <BBB> or run make clean && make, to check that... [12:28] <pb_> dolphy: if it's putting "/usr/lib/libgstgconf-0.6.so" on the link command line, that doesn't sound like an LD_LIBRARY_PATH issue. [12:29] <BBB> well, if it's libtool - libtool tries to find the full path to the lib [12:29] <BBB> if there's no way it can find it... [12:29] <pb_> oh, does it? geez, I hate libtool so much. [12:31] <BBB> it probably has some use [12:31] <BBB> sometimes [12:31] <BBB> for some people [12:31] <BBB> in some cases [12:31] <BBB> etc. etc. [12:32] <dolphy> yup still the same problem [12:33] <BBB> and if you ldd the resulting file [12:33] <BBB> what does it find? [12:34] <dolphy> there is no resulting file [12:34] <dolphy> the gcc command is simply failing [12:34] <dolphy> because gst-gconf symbols are missing [12:34] <dolphy> the new methods we added [12:47] <thomasvs> dolphy: ok, so that's a different problem then, no ? paste the exact error [13:02] andy_t (~cha...@pp...) left #gstreamer. [13:13] <BBB> so what's failing? gst-plugins compile or gst-player compile? [13:15] <thomasvs> dolphy: can you log a full bug report please so we can look at it ? [13:17] gernot (~ge...@a2...) joined #gstreamer. [13:18] <dolphy> gst-player compile [13:18] <dolphy> the problem is in the fact that i have debian packages installed locally for 0.6.0 [13:19] <BBB> deinstall these first... having multiple versions of the same majorminor combination around is never gonna work [13:20] <BBB> (I definately wouldn't support it) [13:20] <thomasvs> no, that's not true [13:20] <thomasvs> it works [13:20] <thomasvs> and it should work [13:21] <thomasvs> dolphy: can you try building it against uninstalled ? [13:21] <thomasvs> ie, set PKG_CONFIG_PATH to include the uninstalled locations of gstreamer and gst-plugins [13:22] <dolphy> so PKG_CONFIG_PATH should be on the unpacked tarballs ? [13:22] <thomasvs> dolphy: hm, not sure, does the tarball have gstreamer-uninstalled.pc ? [13:22] <thomasvs> I know it works against CVS, just not sure against tarballs [13:24] <gernot> BBB: Got it to work yesterday, at least partially - www.geofront.se/snap.png :) [13:24] walken (fo...@12...) left irc: "l8r" [13:24] <BBB> cool [13:24] <BBB> :) [13:24] <dolphy> i m trying [13:24] <BBB> gernot: but you're using fakesink now? [13:24] <BBB> or did you already write a glsink? [13:24] <gernot> BBB: Right ! [13:25] Action: BBB hopes gernot will write a glsink [13:25] <gernot> BBB: There is no use in a glsink for the moment - what should be the purpose of it ? [13:25] <BBB> the same you're doing [13:25] <BBB> blit incoming video data onto a 3d object as a texture [13:25] <gernot> I have written a vtexsink, since I thought there were no callbacks for frame delivery [13:26] <gernot> sure, but that can be done in the existing framework [13:26] <gernot> just register a sink handoff [13:26] <gernot> Uploading the texture is impossible to do implicitly, or so complicated that it is easier to do it with a callback [13:27] <gernot> what I _could_ do is a kind of "OpenGL toy sink" [13:27] <gernot> when you have created interactivity support :-) [13:27] Action: BBB looks at taaz & co ;) [13:28] <gernot> I need to talk to wtay later on, 'cause ds and I have found out yesterday that the meaning of red,green,blue_mask is _reversed_ in comparison to SDL, AFAIK [13:29] <gernot> e.g. 0xff000000 designates the _last_ bytes in a pixel buffer: *(buf + 3) [13:29] <thomasvs> gernot: so, it seems you like gstreamer ? :) in it for the long run ? [13:29] <BBB> gernot: the first byte in an uint32 is the lsb on little endians [13:30] <BBB> so yes, 0xff0000 is the msb, so the last byte, so it's *(buf + 3) [13:30] <BBB> that's logical [13:30] <BBB> (on little endians) [13:30] <BBB> on big endians, 0xff000000 is the first byte (still msb) [13:30] <gernot> thomasvs: well, it is a _clearly_ technically superior solution to my TexMPEG, see www.geofront.se/products_software_texmpeg.php [13:31] <gernot> thomasvs: If it works stably, I will even incorporate MPEG Z/Alpha support into it in the long run [13:31] <gernot> thomasvs: But what I need you developers to do is to support us application developers and fix the remaining bugs now that the API is so good [13:31] <gernot> thomasvs: and not get bored with us ;) [13:31] <gernot> thomasvs: _not_ that this has _happened_ here [13:32] <gernot> thomasvs: But I just know from my own experience that there is always a danger for that [13:32] <BBB> we're not bored yet [13:32] <BBB> this is far too much fun ;) [13:32] <BBB> and in the long run, you'll be able to do v4lsrc ! glsink too :) [13:32] <gernot> I can tell you that you can KICK ASS with this stuff ! [13:33] <gernot> BBB: yepp, no prob whatsoever ! ;) [13:33] <BBB> (watch TV on a GL object, what more could one wish for?) [13:33] <gernot> We can do that already _today_, if the v4lsrc works ;) [13:33] <BBB> v4lsrc works [13:33] <BBB> it just has some synchronization issues with sound [13:33] <BBB> working on that :) [13:33] <BBB> for direct display, it just works [13:33] <BBB> it's recording that's not perfect yet [13:34] <BBB> (recording-to-disk/stream) [13:36] <gernot> The most important thing is that you start to be able to kick OpenML's and DigitalMedia SDK's butt, too ! [13:36] <gernot> Know those 2 API's ? [13:36] <gernot> DM SDK is SGI's long-established standard for their own hardware [13:37] <BBB> never heard of before [13:37] <BBB> does gstreamer run on irix? [13:37] <BBB> can you make it run on irix? ;) [13:37] <gernot> They made a half-hearted attempt to release it once as open software, but then they let everybody down on it [13:38] <gernot> OpenML is supposed to extend OpenGL so that it can handle video hardware, but then OpenGL 2.0 came, and there is a slight confusion about things for the moment. [13:38] <gernot> www.khronos.org [13:38] <gernot> Well, I think I _can_ make it on IRIX, but I was always afraid that the scheduler stuff would mix things up [13:38] <gernot> that one is written in assembler, right ? [13:38] <BBB> I think so [13:39] <BBB> but we've got gthread schedulers too [13:39] <gernot> all assembler would have to be skipped [13:39] <BBB> both in 0.6.0.2 and in HEAD [13:39] <gernot> right - if it is so, then I can give it a shot ... does it need gcc ? [13:39] <gernot> guess so, right ? [13:40] <gernot> need to compile up all the required libraries - very busy with video texture stuff for the moment, but I'll keep it in mind :-) [13:40] <BBB> yes [13:40] <BBB> or maybe cc works too [13:40] <BBB> that's just glib2 and libxml2 [13:40] <gernot> there is gcc on IRIX, too, but I don't what's the most recent version that IRIX delivers in packages [13:40] <gernot> *checking* [13:40] <BBB> I mean, we only care about core portability [13:40] <BBB> plugins will come on the long run [13:41] <gernot> sure [13:41] <gernot> hpv ~\> gcc -v [13:41] <gernot> Reading specs from /usr/freeware/lib/gcc-lib/mips-sgi-irix6.2/2.95.2/specs [13:41] <gernot> gcc version 2.95.2 19991024 (release) [13:41] <gernot> matrix /disk/media/home/gernot\> gcc -v [13:41] <gernot> Reading specs from /usr/freeware/lib/gcc-lib/mips-sgi-irix6.5/3.0.4/specs [13:41] <gernot> Configured with: ../configure --prefix=/usr/freeware --enable-version-specific-runtime-libs --disable-shared --enable-threads --enable-haifa [13:41] <gernot> Thread model: single [13:41] <gernot> gcc version 3.0.4 [13:41] <gernot> okej, so they are quite recent ... [13:42] <gernot> prob is, that I won't access to any Onyx now that I move to Austria - anyone can offer me one ? ;) ;) [13:42] <gernot> Well, I _do_ have access, but only remotely [13:44] <BBB> onyx? [13:44] <BBB> and why are you moving back to austria? [13:45] <gernot> SGI Onyx, the bigger babies, see http://www.rug.nl/rc/hpcv/visualisation/ [13:45] <gernot> you've been there ;) [13:45] <gernot> Ars Electronica Center has offered me a nice job, working with _3D video textures_ on _clusters_ with _big projectors_, using _GStreamer_ [13:46] <gernot> (tell me when you've recovered again ;) ) [13:46] Action: BBB slams [13:46] <BBB> can I come too? [13:46] <gernot> Sure ! [13:47] <BBB> cool! [13:47] <gernot> Not for working there, yet, but be assured that if this expands, you are one of the first guys I will propose :) [13:47] <BBB> haha [13:47] <BBB> I'd better find a job in my own industry ;) [13:47] <gernot> Naaah, why that ? ;) [13:48] <BBB> once gstreamer can do videorecording fine, I'll propose using gstreamer at my current job too [13:49] <gernot> ah ! What are you working with? [13:52] <BBB> delta solutions, we do video archving of all dutch tv channels and stream that to interested clients [13:53] <BBB> we currently use ffmpegrec, I'm converting lavrec to do that job better ;) [13:53] <BBB> and in the long run, I might use gstreamer instea [13:54] <gernot> niiiice ;) [13:54] <gernot> (and what does that have to do with Biology ? ;) ) [13:54] <gernot> (I see some indication that Ronald gets stuck in computer science) [13:58] Action: BBB too [13:58] <BBB> bad thing :P [14:01] ChrisHJW (~christian@12.160.18.10) joined #gstreamer. [14:02] ChrisHJW (~christian@12.160.18.10) left irc: Read error: 104 (Connection reset by peer) [14:03] ChrisHJW (v-167@12.160.18.11) joined #gstreamer. [14:05] <gernot> Why that ? The most important thing is to do something you really burn for, that's why go for higher education [14:05] <gernot> I mean, I can not really claim that it was the _education_ here that offered me access to this job [14:06] <gernot> but it made me stand out and gave me time to concentrate on what I was good at [14:06] <BBB> I don't know what I want, I really don't [14:06] <BBB> I've always figured I didn't want to do computing science because it's too easy [14:06] <pb_> ha :-) [14:07] <BBB> education here in the netherlands is either too theoretical and high-level-language (pascal, perl, java) or you just don't learn a single thing [14:07] <BBB> there's nothing like low-level (kernel? unix? c? asm?) programming education here at a university level [14:07] <BBB> if that existed, my whole life would be so much easier ;) [14:08] <gernot> I had some ... realtime operating systems and their programming [14:08] <gernot> but it was only 4 weekhours [14:08] <gernot> 4 points, I just remember you had the same system [14:09] <gernot> "because it's too easy": maybe because you're _good at it_ ? ;) [14:09] <gernot> man, you can earn _money with that_ ! ;) [14:09] <BBB> maybe... but that doesn't defeat the point :P [14:09] <BBB> it's not a challenge [14:09] <vishnu> yah, programming is boring [14:09] <BBB> generally, yes [14:10] <vishnu> but nobody else is fixing the ffmpeg encoders so i guess i have to [14:10] <gernot> Sure, you can see it like that, too, but there is _enough_ abstract and advanced work to do - see GStreamer :) [14:10] Company (~as...@rz...) joined #gstreamer. [14:10] <gernot> Servus Company ! [14:10] <Company> hi [14:11] <vishnu> gernot: whatever, it's all simple to me -- "advanced" just means something you didn't understand yet [14:11] Action: vishnu yawns [14:13] <gernot> vishnu: *laughing* well, when you have _understood_ something, everything is easy ;) [14:13] <vishnu> yah :-( [14:13] <gernot> vishnu: it's only that only _we_ have the capability to quickly analyze and construct software, and that's what we are training and exercising here :-) [14:14] <vishnu> the chosen few .. what an honor :-P [14:14] <Company> hm, i'm still confused by that Makefile.am [14:14] <gernot> vishnu: naaah, not _that_ few [14:15] <gernot> vishnu: I mean more that what BBB considers as easy to him, is _hard_ (NP-hard ? nerd-joke ;) ) for 99% of the rest [14:15] <gernot> vishnu: which means that he should _use_ it to make his living, and by that I mean not _exploit_ others' ignorance [14:16] <Company> thomasvs: ping? [14:16] <gernot> vishnu: but get so much that he can concentrate on what he considers being fun in life, whatever that is is up to him [14:17] <BBB> going out, I guess :P [14:17] <vishnu> d'oh! [14:17] <gernot> eating in nice restaurants, maybe ? ;) [14:18] <gernot> vishnu: yes, of course d'oh, but not every nerd realizes that in the beginning :-) [14:18] Action: BBB will go to pinkpop and such [14:19] <vishnu> yah, get married, buy a house, etc. it's fun [14:19] <dolphy> vishnu: yep indeed :) [14:19] Action: gernot forbids BBB to go to evil pinkpop concerts, that's not good for him [14:20] <gernot> well, if you consider that _fun_ then do it ! [14:20] <gernot> but don't do it only because _society_ expects it from you ... [14:20] <dolphy> gernot: but even when something is simple it can be boring or interesting ... just depends in which goal you use it [14:20] <vishnu> is gst-editor borked in CVS HEAD? or is it just me? [14:20] <Company> vishnu: might be that it's borked for HEAD [14:21] <gernot> there are 1000 other possibilities to do with your life [14:21] <Company> vishnu: i think the release (and therefore the HEAD status) is for 0.6 [14:21] <vishnu> Company: yah .. ok [14:21] <gernot> dolphy: that's right, I never questioned that [14:21] <Company> though we haven't had many API changes in 0.6 => HEAD yet [14:21] <Company> should be easy to fix [14:22] <gernot> So people, what do you want ? A flag with a video on it, or a rotating sphere with a video on it ? :) [14:22] <Company> water that reflects video from a flag above it? [14:22] <Vakor> A rotating spherical flag with a video on it [14:23] <vishnu> Company: yah but gst-editor is crashing in gst_plugin_find_feature (for me) [14:23] <gernot> gnaaah, no special wishes here, my Matrox G400 can't handle that ;) [14:23] <BBB> gernot: pinkpop is one of the things I enjoy most... you remember krezip? they're pinkpop groups - that's my stuff ;) [14:23] <gernot> BBB: I suspected so ;) [14:23] Action: BBB is fighting with a good A/V sync system in v4lsrc [14:23] <Vakor> My G400 can't even handle normal 2d graphics well enough for my monitor, which is rather disappointing [14:23] <BBB> ^H^H^H^H^H^ I'm doing experiments in the lab here [14:24] <Company> vishnu: not good at all... [14:24] <Company> vishnu: please fix it ;) [14:24] <vishnu> Company: easy? uh, well, how? [14:24] <BBB> vishnu: did you forget to run gst-register? [14:24] <vishnu> BBB: i just checked by registry. it looks OK .. [14:24] <Company> if forgetting to run gst-regster crashes our apps, we have a loooong way to go ;) [14:24] <gernot> Ok, doing a flag for the moment [14:25] <Company> gah, autotools hackers are rare [14:26] <Company> should --prefix support relative paths? [14:26] <pb_> Don't think so, no. [14:26] <Company> ok, so i'm not trying it :) [14:27] <Vakor> Company: rule of thumb for dealing with autotools: any and all simplifying assumptions are not only acceptable, they're to be encouraged :-) [14:28] <Company> Vakor: i have a patch here that i need on solaris, but thomasvs ants to apply it only if he understands why the other stuff breaks [14:28] <Company> and i have _no_ idea [14:29] <Vakor> Company: what was the phase of the moon when it was testing on solaris? [14:29] <gernot> Oh yes ! One thing I forgot [14:29] <Company> dunno, what's the moon phase today? [14:29] <gernot> I need to be able to specify rgb_line_stride in colorspace [14:30] <Vakor> if it involves autotools, a slightly different phase of moon is generally sufficient to cause breakage :) [14:30] <gernot> that will say, that each line of pixels can be followed by some empty space [14:30] <gernot> only this way I can fill a video texture of e.g. 1024x1024 with a video rectangle of e.g. 320x240 [14:30] <Vakor> gernot: sounds reasonable, that's a fairly common thing to do. [14:31] <Company> but that's not only useful for RGB, isn't it? [14:31] <gernot> Vakor: can you fix that in some future ? Or only wtay ? [14:31] <Vakor> Company: it's... much more likely to be useful for RGB, but you're right, it's not _solely_ useful for that [14:31] <Vakor> gernot: me? Hah! I don't know much about gstreamer [14:32] <Company> so we should make that an optional property for video/raw caps then [14:32] <gernot> Vakor: Ok, overrated you ;) ... good, will ask wtay ! [14:32] <Company> ... some day [14:32] <Company> ... when we have nothing else to do [14:32] <gernot> Company: yepp, don't hurry [14:32] <gernot> with texture subimage downloads, this is not necessary, either [14:32] <gernot> it just doesn't work on my G400 card [14:32] <gernot> (bug in the driver) [14:33] <Company> fix the driver... [14:33] <Vakor> gernot: I'm just interested in gstreamer, and am planning on writing some plugins for it one day, when I have more time [14:33] Action: Company remebers when he wanted to fix something in the radeon driver [14:33] <gernot> Company: I'm no _masochist_ :-} [14:34] <Vakor> Company: it's common to want it for other reasons as well (mostly so that your in-memory representation is efficient w.r.t. things like alignment) [14:34] <gernot> Company: you heard about the Xfree86 split ? One reason was that there was no virtually _no_ doc ... [14:34] <gernot> ;) [14:34] <vishnu> cool, a new valgrind was released on 7 April [14:34] <Vakor> vishnu: yes. That reminds me, I have some bug reports on it to send off... [14:35] <Vakor> (I can't understand the relevent bits of the code well enough to figure out why it's breaking on my test cases) [14:37] hisdudeness (uy...@rz...) joined #gstreamer. [14:39] <Company> german people always have the weirdest names on IRC [14:40] <Vakor> Koreans are generally weirder. [14:40] <Company> yes, but koreans don't use ascii natively, so they're excused [14:57] md` (~illuminat@pD9533EDA.dip.t-dialin.net) joined #gstreamer. [15:00] <dolphy> BBB: can you review diffs between cvs head of libgstplay and 0_6 please ? [15:00] <dolphy> BBB: i ve made a change requested by Company to the pipeline structure [15:07] hisdudeness (uy...@rz...) left irc: "Leaving" [15:10] vishnu (~joshua@61.11.23.205) left irc: Read error: 110 (Connection timed out) [15:10] <Company> thomasvs: found the reason - i wasn't using gnu make, but suns make [15:11] <Company> thomasvs: gnu make uses the makefile just fine [15:12] <thomasvs> Company: ah, ok [15:12] <thomasvs> Company: yeah, I think we have gnu make as a requirement [15:18] Company (~as...@rz...) left irc: Remote closed the connection [15:25] ChrisHJW (v-167@12.160.18.11) left irc: Read error: 110 (Connection timed out) [15:28] vishnu (~joshua@61.11.23.205) joined #gstreamer. [15:29] <vishnu> this is so frustrating [15:30] <BBB> dolphy: my email is still broken [15:30] <BBB> dolphy: can't do much :( [15:31] <dolphy> BBB: want the link ? [15:31] <dolphy> http://cvs.sf.net/cgi-bin/viewcvs.cgi/gstreamer/gst-plugins/gst-libs/gst/play/play.c.diff?r1=1.15&r2=1.16 [15:31] <dolphy> http://cvs.sf.net/cgi-bin/viewcvs.cgi/gstreamer/gst-plugins/gst-libs/gst/play/playpipelines.c.diff?r1=1.10&r2=1.11 [15:32] <vishnu> whoever is responsibilty for bug 110757 should be taken out and shot [15:35] <BBB> long live the cross-references source ;) [15:40] ^iain^ (~ia...@us...) joined #gstreamer. [15:40] <dolphy> vishnu: your patch seem correct [15:55] vishnu (~joshua@61.11.23.205) left irc: Read error: 110 (Connection timed out) [16:18] vishnu (~joshua@61.11.23.205) joined #gstreamer. [16:18] <vishnu> hello? [16:19] <vishnu> has anyone tested gst-launch with mpeg video pipelines? (CVS HEAD) [16:19] <BBB> no, will do after 0.6.1 is kicked out [16:20] <vishnu> BBB: well, it doesn't work. file ./grammar.y: line 407 (gst_parse_perform_link): assertion failed: (GST_IS_ELEMENT (src)) [16:25] <BBB> as I said, I'll help after 0.6.1 is out [16:25] <BBB> I don't mean to be rude, but all my attention has gone to 0.6.1 lately... we'll never get that out if we don't work on it fulltime for a while [16:25] <vishnu> BBB: ok ... [16:26] vishnu (~joshua@61.11.23.205) left irc: "Client Exiting" [16:28] <dolphy> you vexed him :) [16:33] <BBB> oops [16:35] Company (~Company@pD958B6BC.dip.t-dialin.net) joined #gstreamer. [16:35] <gernot> dolphy: vexed ? [16:44] <dolphy> hmm maybe that s a "frenchenglishized" word :) [16:47] <^iain^> annoyed [16:48] <gernot> dolphy: you're from Canada ? [16:48] <gernot> So ! [16:49] <gernot> www.geofront.se/snap_flag.png and www.geofront.se/snap_flag2.png to make our GStreamer core developers happy ! :) [16:49] <^iain^> nice... [16:49] <^iain^> it worked then? [16:49] <dolphy> gernot: nope french guy living in spain [16:50] <^iain^> gernot; if only you could make an mpg of it all :) [16:50] <gernot> yepp ! But I still have to copy the texture content once to get it into texture alignment [16:50] <gernot> well, you don't want to know the frame rate ;) [16:50] <^iain^> ah, heh [16:51] <gernot> I really have to buy myself a GForce, my Matrox G400 _sucks_ - but I need it for my TV-tuner, that's why I just bid for a Matrox G200 PCI [16:51] <^iain^> yeah g400 isn't great for 3d stuff [16:52] <gernot> dolphy: Oh, that's as uncommon as an Austrian in Sweden, I guess ;) [16:53] <gernot> good, mission accomplished for today ! Time for some sun ;) [16:53] <dolphy> gernot: exactly :) [16:53] <gernot> Swedish summer is great, light until 8-9 o'clock, maximum 11 ;) [16:56] <Titanium> Is there in gstreamer anything for managing RTP protocols? [16:56] <gernot> only udpsink and udpsrc, I think [16:58] <Titanium> ok [16:59] <Titanium> How can I send video and audio via internet without lose lot of information? [16:59] <Titanium> I have problems with udp datagrams, to send the video [17:03] <BBB> what do you guys think of http://bugzilla.gnome.org/showattachment.cgi?attach_id=15712 for v4lsrc A/V sync? [17:18] ChrHJW_log (~cwi...@p5...) left irc: Read error: 110 (Connection timed out) [17:20] <Titanium> BBB: Does it able to get the fps properly? [17:21] <BBB> if you use use_fixed_fps=false, it calculates the fps from all data received [17:21] <BBB> it doesn't drop/insert frames [17:22] <BBB> there's some horrible issues with that, but it's the best I can think of for now [17:23] <Titanium> ok, thanks [17:23] <BBB> by default, use_fixed_fps set to TRUE, because you get more reliable timestamps with that [17:23] <Titanium> In other hand, do you know who did the trp pluggin? [17:23] <BBB> zeenix [17:24] <Titanium> Thanks [17:24] Action: BBB still needs to adapt v4lmjpegsrc/v4l2src with these fixes [17:24] <BBB> anyway, I'll do that later [17:24] <BBB> tonight or so [17:24] Action: BBB goes home no [17:24] <BBB> later! [17:24] BBB (~rb...@ph...) left irc: "Client Exiting" [17:24] <Titanium> ok [17:25] <Titanium> bye [17:27] steveb_ (~steveb@202.0.44.177) joined #gstreamer. [17:41] steveb__ (~steveb@202.0.44.244) left irc: Read error: 110 (Connection timed out) [17:42] gernot (~ge...@a2...) left irc: Remote closed the connection [17:43] Uraeus (~csc...@7h...) joined #gstreamer. [17:45] <Company> hm [17:45] <Uraeus> hey ho [17:45] <^iain^> hey [17:45] <Company> we don't have any color -> black/white converter, do we? [17:45] <Uraeus> don't think so [17:45] Action: Company is impressing his brother with the player :) [17:45] <Uraeus> unless one of the effecttv plugins can do that [17:50] Marsupilami23 (~Mar...@15...) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!" [17:51] <^iain^> Company; how do I test your audioconverter? [17:52] Nick change: wtay-zZz -> wtay [17:52] <wtay> yo [17:53] Action: wtay has ideas to implement caps_nego failed signals + possibility to recover from failure [17:53] <foser> hmm reminds me, Uraeus gonna send you the gentoo blah for the release page [17:55] <foser> i noticed soundjuicer does checks for available plugins .. are there plans to completely split up plugins (would make things easier on the gentoo side) [18:02] <Uraeus> foser: ok :) [18:02] <Uraeus> foser: not sure what we plan at this point, I noticed that Red Hat bundled everything into one heap [18:02] <foser> whats the standard encoding level of the gstreamer vorbisenc ? [18:03] <Uraeus> foser: no idea, wtay probably now [18:04] <Uraeus> foser: lets make a deal, I find out what the default encoding level for vorbis is for you, and in return you make a aiffencoder for gstreamer, ok? ;) [18:04] <foser> Uraeus: we do too, but its possible to turn things on and off based on useflags (and we do not build everything, would be an enormous amount of deps ;)) .. but if an app wants a certain plugin we cant make sure it is there in the current situation [18:05] <foser> hehe i have no idea what format aiff is in the first place ;) [18:05] <Uraeus> well I guess getting more apps to do like sound-juicer and check for plugins would be good idea [18:05] <Uraeus> foser: afaik so is aiff the sound formats used by SGI on Irix, kinda like their version of M$ wav or Sun au [18:06] <wtay> foser: according to gst-inspect vorbisenc, the quality is set to 0.3 [18:06] <Uraeus> wtay: what are the differences between wav, au and aiff? quality? compression? [18:06] <wtay> foser: bitrate and max-bitrate are set to -1 (so libvorbis picks its default I guess) [18:07] <foser> right.. would that be 7 [18:07] <foser> on the vorbis scale.. [18:07] <wtay> Uraeus: they are all containers [18:07] <foser> not sure what soundjuicer is doing right now ;) [18:07] <wtay> foser: the scale goes from 1 to 10 [18:07] <Uraeus> wtay: but they all tend to contain uncompressed audio right? [18:07] <foser> yes.. thats why [18:08] <wtay> Uraeus: mostly uncompressed and ulaw, but anything is possible [18:08] <foser> i suppose 0.6.1 is anytime now ? dont wanna add soundjuicer to gnome with the comment encoding problem right now [18:09] <Company> ^iain^: you use incompatible formats and try plugging :) [18:09] <Uraeus> wtay: so an uncompressed & ulaw sound file would be the same size using aiff, au and wav right? [18:09] <Company> ^iain^: it was mostly meant for people having incompatibilities between audio plugins (like big/little endian etc) [18:10] <wtay> Uraeus: yes, roughly, just the headers are of different size [18:11] <wtay> aiff/avi/mov are all members of the iff/riff family [18:11] <Company> ^iain^: or you could get problems when people rund around with 8khz mono unsigned 8bit audio and somebody has a volume elemnt somewhere inbetween [18:11] <Uraeus> wtay: but wav and au is not iff/riff? [18:11] <wtay> au is very simple with a simple header [18:11] <wtay> Uraeus: wav is like avi [18:12] <wtay> Uraeus: wav is also riff/iff [18:12] sxpert_avignon (~sxpert@AMarseille-206-1-15-246.abo.wanadoo.fr) left irc: Read error: 110 (Connection timed out) [18:13] <wtay> wav is in fact very much like avi (can share the same code) [18:13] <Company> * wtay has ideas to implement caps_nego failed signals + possibility to recover from failure <-- that was something i need before i touch the autoplugger again, so gogogo :) [18:13] <wtay> Company: that is a good idea :) [18:13] <wtay> you really need it for a good autoplugger that doesn't resort to the spideridentity stuff there is now [18:13] sxpert_avignon (~sxpert@AMarseille-206-1-19-19.abo.wanadoo.fr) joined #gstreamer. [18:13] <Company> wtay: and we need to come up with a better way of doing seeking in clocking... [18:14] <Company> wtay: how would you do it without such elements? [18:14] <wtay> Company: yes, I think it's not good to try to resync with a discont event [18:14] <wtay> Company: just connect, try different elements on capsnego failure [18:15] <Company> wtay: but you have to look for events and caps nego so you can automatically replug on new caps [18:15] <wtay> the hard part is to try to work towards the target caps [18:15] <Company> well, actually you don't... [18:16] <Company> if you can listen to caps failures, then that's not a problem [18:16] <thomasvs> aiff is apple interchange file format [18:16] <wtay> just react to capsnego failures will work so much better than the current solution [18:16] <Company> though there might be intended caps failures (colorspace trying out for example) [18:16] <wtay> Company: yes... [18:17] <thomasvs> foser: the "comment encoding problem" ? [18:17] vishnu (~joshua@61.11.23.205) joined #gstreamer. [18:17] <Company> thomasvs: gstreamer puts comments in nonconformat format into the metadata part of a vorbis [18:17] <foser> thomasvs: is said to be fixed in 0.6.1 .. mentioned in sound juicer readme [18:18] <Company> thomasvs: i fixed it for 0.6.1 [18:18] <thomasvs> Company: bug number ? [18:18] <foser> there we have the man [18:18] <thomasvs> woo-woo [18:18] <thomasvs> Build of gst-plugins-0.6.0.2 succeeded. [18:18] <Uraeus> hi vishnu [18:19] <vishnu> Uraeus: hi [18:19] <Uraeus> vishnu: so Company broke gst-launch with his latest updates to head? [18:19] <wtay> Company: we probably need something to indicate that after this try_set_caps the plugin is going to be out of options [18:19] <vishnu> Uraeus: looks like he didn't test very many pipelines [18:20] <Company> thomasvs: http://bugzilla.gnome.org/show_bug.cgi?id=110544 [18:20] <Uraeus> vishnu: heh, well but between us we test enough to fix it before release I think :) [18:20] <Company> wtay: sounds good - gst_pad_failed_caps (pad), which would at the moment just call gst_element_error (parent, "could not set caps") [18:21] <Company> would be a useful API addition right now... [18:22] <wtay> we don't need a return code to instruct the plugin to try again [18:23] <thomasvs> Company: why did you change the names of the properties in that patch ? [18:23] <vishnu> Uraeus: regressions like this are really annoying [18:24] <thomasvs> vishnu: company did send a mail to the list, and I questioned onlist about it, so... [18:24] <thomasvs> vishnu: you could have joined in when you saw that discussion [18:24] <Uraeus> vishnu: they are, but when we run software from head they will happen from time to time and I don't think gstreamer cvs is worse than others [18:24] <thomasvs> vishnu: if you care about those regressions, help us set up the testsuite [18:24] Action: vishnu eats his words [18:24] <^iain^> wtay; the opt scheduler now crashes marlin when its finished loading. the basicomega one works fine [18:25] Nick change: wtay -> wtay-food [18:25] <wtay-food> brb [18:25] <vishnu> iain: yah, gst-editor shows the same behavior (CVS HEAD) [18:25] <Uraeus> wtay-food: no fries!! they are bad for you [18:25] <thomasvs> vishnu: damn you, thought I had you :) [18:25] <Company> thomasvs: i changed the names of the properties to reflect the vorbis naming [18:25] <^iain^> vishnu; glad I'm not alone this time [18:26] <vishnu> thomasvs: hey, cut me some slack. i'm working on finished my house and i fixed some bugs today too [18:26] <vishnu> finishing [18:26] <thomasvs> vishnu: I am cutting you slack, I was just hoping we were finally able to suckering you back in :) [18:26] <Company> thomasvs: there's a standard on tag naming for vorbis somewhere [18:27] <thomasvs> Company: can you show me that standard ? Last time I researched it the vorbis guys were adamant on it being free form instead of policied [18:29] <Company> thomasvs: http://www.xiph.org/ogg/vorbis/doc/v-comment.html [18:29] <Company> vishnu: what did i break in gst-launch? [18:30] <vishnu> Company: try the pipelines in the manual page or in my bugzilla report [18:31] <vishnu> Company: MPEG1 playback doesn't work, for one [18:31] <vishnu> Company: is there different magic for connecting dynamic pads? [18:31] <thomasvs> Company: cool, thanks for that link [18:32] <Company> vishnu: put a dot before video%02d [18:32] <Company> that should work [18:32] md` (~illuminat@pD9533EDA.dip.t-dialin.net) left irc: "Client exiting" [18:32] <vishnu> Company: i'm pretty sure i tried that. i'll try again ... [18:33] <Company> but the bug needs to stay open, the correct error would have to be "no such element video_%02d [18:33] md` (~illuminat@pD9533EDA.dip.t-dialin.net) joined #gstreamer. [18:34] <Company> though video_%02d is no valid element name - elements may not have % in them [18:34] <vishnu> do i need the "!src" thing or is "!" ok? [18:34] Action: Company is thinking aloud here [18:34] <Company> ! should be ok [18:35] <vishnu> then the manual page needs to be fixed [18:35] <Company> you specify links by <elementname>.<pad> [18:35] <Company> the man page isn't updated yet [18:35] <Company> look at docs/random/company/gstparse please [18:36] <Company> i thought that document was a bit long for a man page ... [18:37] <Company> somebody should shorten it, i always write docs too long - if i write them [18:38] Titanium (~giorgio@213.170.47.89) left irc: Remote closed the connection [18:39] <vishnu> hrm [18:41] <Uraeus> vishnu: that memory corruption patch, is it also valid for 0.6 branch? or is it a head only thing? [18:41] gernot (~ge...@a2...) joined #gstreamer. [18:42] <vishnu> Uraeus: dunno. ask wtay [18:43] <vishnu> Company: that dot makes a big difference. [18:43] Titanium (~giorgio@213.170.47.89) joined #gstreamer. [18:44] <Company> vishnu: it should. Pad references are required to have a dot with the new parsing [18:45] Nick change: Uraeus -> Ura_dinner [18:45] <Company> vishnu: but you may now put spaces between pad references and the ! [18:45] <vishnu> Company: ok, it's a good improvement. cool. now i can figure out why the pipeline doesn't work [18:46] vishnu (~joshua@61.11.23.205) left irc: "Client Exiting" [18:46] vishnu (~joshua@61.11.23.205) joined #gstreamer. [18:52] <vishnu> Company: does this work? ./gst-launch filesrc location=/local/aleader/ljr.mpg ! mpegdemux .audio_%02d ! mad ! alsasink [18:53] <vishnu> this works for me: ./gst-launch filesrc location=~/Shennay.mp3 ! mad ! alsasink [18:53] <vishnu> but the audio from mpeg isn't getting played [18:55] <vishnu> Company: ping? [18:57] <Company> vishnu: lemme check here [18:58] <Company> vishnu: you don't need to specify the padname [18:58] <Company> ./gst-launch filesrc location=/local/aleader/ljr.mpg ! mpegdemux ! mad ! alsasink should work [18:58] <vishnu> yah, it does. but why? [18:58] <Company> i don't know however why it doesn't take the padname... [18:59] <vishnu> Company: i can't get A/V pipelines to work, that's why i tried an audio only pipeline [18:59] <vishnu> Company: don't i need to specify the pad name for A/V? [18:59] <Company> vishnu: no, you don't [18:59] <vishnu> Company: cool [18:59] <vishnu> let's try it ... [19:00] <Company> vishnu: and you should specify .audio_00 because you don't want a request pad [19:00] <vishnu> Company: oh! [19:00] <vishnu> that makes more sense anyway [19:00] <Titanium> Is there any documentation about the rtp plugin? [19:01] ChrHJW_log (~cwi...@p5...) joined #gstreamer. [19:02] <vishnu> Company: wow, great work. [19:02] <vishnu> Company: now update the man page! [19:02] <Company> vishnu: thx [19:02] <Company> man page is on a todo - i still hope someone shortens the doc... [19:02] <vishnu> Company: audio_00 and video_00 work fine [19:03] <vishnu> Company: hey, just cut & paste from your write up if you're lazy [19:03] <vishnu> Company: just don't leave the man page in the poor shape it is now [19:03] <vishnu> Company: now it worse than useless -- misleading [19:03] <Company> vishnu: right [19:05] <vishnu> Company: is it possible to join two pipelines, such as inputs to avimux? do you have a syntax for that? [19:06] <Company> vishnu: the other way round [19:06] <vishnu> Company: how so? [19:06] <Company> videoin ! avimux ! filesink audioin ! avimux0. [19:07] <vishnu> Company: ok cool .. i'll try it later [19:07] <Company> hope it works as expected... [19:17] chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) left irc: Read error: 104 (Connection reset by peer) [19:18] <Company> anybody knows the switch for less to produce colored output? [19:18] <Company> i want to read a gst-mask log [19:21] chillywilly (~danielb@CPE-24-167-199-51.wi.rr.com) joined #gstreamer. [19:25] Action: dolphy is implementing totem media info popup in gst-player [19:30] foser (d0...@22...) left irc: Remote closed the connection [19:31] <dolphy> Company: i usually use more [19:33] thomasvs (~th...@ca...) left irc: Read error: 110 (Connection timed out) [19:40] <sxpert> dolphy: when is the september meeting ? (need dates for holiday put-down purposes) [19:41] vishnu (~joshua@61.11.23.205) left irc: "Client Exiting" [19:41] foser (d0...@22...) joined #gstreamer. [19:42] harshy (~ha...@dh...) left #gstreamer ("PART!"). [19:42] harshy (~ha...@dh...) joined #gstreamer. [19:43] <dolphy> sxpert : we should propose some dates [19:43] <dolphy> sxpert : i ll send a mail tonight to the mailing list [19:43] <sxpert> dolphy: ok... [19:43] pb_ (~pb...@ds...) left irc: "brb" [19:44] <sxpert> dolphy: http://hellxions.ath.cx/eigo/030.jpg [19:46] <dolphy> sxpert : this is a fake [19:46] <sxpert> dolphy: you think ? [19:47] <dolphy> sxpert 99% certain [19:47] <sxpert> dolphy: ah, pretty good photoshopping though [19:47] <dolphy> sxpert : i you look really close nobody is looking at them [19:47] <dolphy> sxpert: which is totally impossible [19:47] <dolphy> sxpert: even the guy is not looking at the girl [19:48] <sxpert> dolphy: ah, heh ;-) [19:48] <sxpert> dolphy: didn't think of that... that's why I decided to ask an expert in the field ;-) [19:50] <dolphy> sxpert : hehe :) but that's a good modification :) [19:51] <sxpert> dolphy: heh [19:59] <^iain^> it might not be a fake [19:59] thomasvs (~th...@71...) joined #gstreamer. [19:59] <^iain^> but there's nothing that implies she's got her pants down [20:00] <dolphy> ^iain^: how can this situation be real... [20:00] <^iain^> low riding jeans [20:00] <Company> you're all just saying that because you can't imagine doing that yourself :p [20:00] <sxpert> Company: you'd do that ? [20:00] <^iain^> I've seen people dancing like that all the time [20:00] <dolphy> ^iain^: if you look on the right there is one girl whose sight is in their direction [20:01] <dolphy> ^iain^: almost every human person would behave different when having that kind of scene in your vision area [20:01] <^iain^> well, is the assumption that they're having sex? [20:02] <^iain^> cos like I've said, dancing like that is quite common [20:02] <sxpert> ah... reminds me of art classes in high school... picture-reading was always fun ;-) [20:02] <dolphy> ^iain^: at least it look like that [20:03] <^iain^> more so in the US/Canada than Europe, but bump n grinding is fairly common [20:03] <^iain^> I don't think they're having sex, but I don't think it's a fake photo [20:04] <dolphy> so you think it's just a coincidence ? [20:04] <^iain^> yes [20:04] <dolphy> could be [20:04] <dolphy> we'll never know anyway :) [20:04] <^iain^> a combination of low jeans, the way they're dancing and the way the girl is looking up [20:04] <dolphy> i sincerly hope for that guy he was fucking her ;-) [20:05] <^iain^> maybe later...but in the club :) [20:05] <dolphy> why wait :) [20:05] pb_ (~pb...@pc...) joined #gstreamer. [20:05] <dolphy> ok time to ride back home [20:05] <dolphy> see you guys later [20:05] <Company> you can even see her jeans [20:05] dolphy (~dolphy@213.170.47.89) left irc: "Network down, IP Packets delivered via UPS" [20:06] <Company> bottom right, though it looks like a shadow [20:07] <sxpert> you gotta love the uraeus email about guadec 2003 ;-) [20:08] <Company> what mail? [20:09] Action: Company isn't subscribed to anything but gstreamer-devel [20:10] <sxpert> Company: it's in there [20:12] <Company> lol [20:13] <Company> tho i'm not sure if i'll go there, so i'll skip registering for now [20:15] Nick change: wtay-food -> wtay [20:15] <^iain^> hey wtay [20:15] <wtay> debian binutils 2.13.90.0.18-1.6 segfaults [20:15] <wtay> yo [20:20] <Company> lol, i didn't even know what spring loaded folders are, that is so obvious and there's a patent on it... [20:25] <^iain^> they rule... [20:26] <^iain^> I dunno what I'm going to do when I need to update Nautilus and my patch doesn't cleanly apply [20:26] <Company> fix the patch [20:26] <Company> isn't that obvious? [20:26] <^iain^> I don't know if I'll be bothered :) [20:26] <Company> is it that big? [20:27] <^iain^> quite big [20:27] <Company> why's that? you just need a signal handler for mouse_over or how it's called and you'Re done, no? [20:28] <^iain^> no [20:28] <^iain^> cos it uses CORBA [20:28] <Company> nautilus needs a disable-patents option :) [20:28] <Company> ah, ok, crap :) [20:29] <^iain^> I forget exactly how its done, but it passes calls back and forward over corba a fwe times [20:29] <Company> ./configure --disable-patented=USA,EU,Japan --disable-problematic=IslamicReligions,Scientology [20:30] ct (~ct...@13...) joined #gstreamer. [20:30] BBB (~rb...@01...) joined #gstreamer. [20:30] <Makki> Company: What are spring loaded folders ? [20:30] <Makki> Company: url ? [20:32] <Company> Makki: http://www.macoptions.com/tips/os/sprgfolder.html [20:33] <Makki> Company: Oh, that! [20:33] <Makki> yeah, that's handy [20:33] <Makki> on occasion... [20:35] <Makki> does apple have a pattent on it ? [20:37] <ct> Company: what was your suggestion for crossfading? [20:40] <Company> ct: creating 2 pipelines with a custom crossfading element that pulls from pad 1 until there's only x frames left and then pulls from pad 2, merges the buffers and continues with pulling from pad 2 [20:40] <thomasvs> crossfading audio ? [20:40] <thomasvs> that should be done with adder [20:40] <thomasvs> the app should control crossfading, not the plugin [20:40] <thomasvs> (all MO) [20:41] <Company> yeah, that was what i meant [20:41] <Company> but you need an element to do the crossfading for you [20:44] <wtay> you also need to continue to pull the pad even if it's completely faded out, if not, the app needs to disable the sink pad or remove the crossfade [20:44] <wtay> (important for NLE) [20:44] <sxpert> wtay: is it possible to change the source of an src in mid-flight ? [20:45] <wtay> sxpert: no, you need to PAUSE (at least) [20:45] <wtay> currently [20:45] <sxpert> wtay: ah... crap ;-) [20:45] <Company> you can just disable the pad when it has sent eos [20:46] <wtay> Company: yes [20:47] <BBB> any critical bugs left in 0.6.0.2? [20:47] <Company> and you should send a signal so the app can connect a new element to the pad to get continuous crossfaded playback [20:47] <ct> thomasvs: yeah [20:47] <wtay> going to commit more advanced infrastructure to query the registry [20:48] <ct> Company: that's something i'd like lymric todo with gstplay [20:50] <BBB> I've got one issue with 0.6/0.7 crossing btw [20:50] <BBB> there's no separate registries [20:50] <BBB> which means you can only use one of the two at the same time [20:51] <Company> huh? [20:51] <Company> isn't the registry in /var/cache/gstreamer-0.x/registry.xml [20:51] <Company> ? [20:52] <BBB> there's no gst-register-0.x [20:52] <BBB> ;) [20:52] <Company> yeah, file a bug about it, somebody has to decide about that. [20:53] <Company> glib-mkenums and so on isn't versioned either btw... [20:53] <BBB> and besides that, ~/.gstreamer/ isn't versioned [20:58] <Company> that's a bug, too [20:58] <Company> it should be ~/.gstreamer/gstreamer-0.x imo [20:58] <Company> but that could be implemented right now :) [20:58] <Company> or would that be an ABI change? [21:02] <pb_> Nothing outside of gstreamer core knows where that directory is, right? If that's true, it wouldn't be an abi change. [21:02] <Company> the user might have a script using it [21:03] <wtay> if we assume the core+utils are upgraded at the same time and the user always reruns -register it's ok [21:10] <Company> shouldn't make install run gst-register? [21:11] <BBB> no [21:11] <thomasvs> Company: i've been told it shouldn't, in general [21:11] <BBB> RPMs do run gst-register [21:11] <Company> why not? [21:11] <BBB> but our docs say you need to run gst-register [21:11] <BBB> so... [21:11] <BBB> I don't consider it an ABI change [21:11] <thomasvs> I would also like to remind people who are be unsure that Dublin is not [21:11] <thomasvs> in Finland, nor are there many/cheap flights going from Helsinki to [21:11] <thomasvs> Dublin. Just so you know :) [21:11] <BBB> but I'd propose to leave that out for HEAD for now [21:11] <thomasvs> HAHAHAHA :) [21:12] <BBB> rofl :D [21:12] <BBB> where did you read that? [21:12] <Company> thomasvs: why shouldn't we run gst-register or libs in general run ldconfig? [21:12] <sxpert> that's what I was mentionning [21:12] <wtay> still trying to make me feel guilty it seems... :) [21:12] <Company> gstreamer-devel [21:12] <BBB> wtay: I wonder why... *glim* [21:12] Action: BBB still doesn't have email |:( [21:13] <sxpert> BBB: mail server b0rken ? [21:13] <BBB> I just sent them an email on how to configure sendmail [21:13] <sxpert> lol [21:13] <BBB> sxpert: a company is doing my mail, but they broke something... [21:13] <BBB> mail for everyone they're hosting is currently not working [21:13] <Company> BBB: i'm doing your mail? [21:14] <BBB> sort of ;) [21:14] Action: BBB will book a flight to oslo as soon as he's got money [21:15] <Company> and when you have money again after that you buy oslo-dublin? [21:15] <Company> and after that the return flights? [21:15] <sxpert> BBB: you mean, to helsinki ? [21:16] <wtay> helsinki is nice, you should all pass by :) [21:17] <^iain^> my friend is there, she loves it [21:17] <BBB> nah, s/oslo/dublin/ ;) [21:20] <ds-work> wtay: what should the caps be for 32-bit RGB video in RGBx memory order? [21:21] <BBB> better said - if you have 32bit RGB, what is the byte order? [21:21] <wtay> ds-work: big endian, r=0xFF000000, g=0x00FF0000, b=0x0000FF00, depth 32, bpp=24 [21:21] <BBB> i.e. which of the 4 bytes is RGBA? [21:22] <wtay> we don't have good caps for RGB yet, libcolorspace has a better way of representing it [21:23] <ds-work> wtay: is there any reason to use little endian for 32-bit RGB? [21:23] <BBB> it's easier for the host, I assume [21:23] <wtay> ds-work: no, it can be represented with an equivalent big endian notation [21:24] <wtay> not for 16 bits though [21:24] <ds-work> I'm adding RGB support to videotestsrc, and running into "issues" with colorspace [21:24] <sxpert> ds-work: well, it's simpler, as big-endians (such as the PPC) are able to read the other way too [21:24] <wtay> colorspace doesn't check the rgb masks at all when doing yuv conversions [21:24] <wtay> IIRC [21:24] <BBB> wtay: if you prefer another representation, please describe it in docs/random/mimetypes or so and I'll try to make it better in the next version of caps [21:25] <wtay> BBB: yeah [21:26] <BBB> afaik, on little endians it's also (in byte-per-byte) R-G-B-A, meaning that 32bit RGB has a red_mask of 0x000000ff, green_mask of 0x0000ff00 and a blue_mask of 0x00ff0000, right? [21:26] <ds-work> ok, I pretty much had my ducks in order with the RGB caps [21:26] <wtay> BBB: uhm, no [21:27] <BBB> why, first byte is LSB, so first byte (R) is 0x000000ff, right? [21:27] <wtay> the masks don't scale to 48 bits video [21:27] Action: BBB is confused now [21:27] <BBB> where's 48? [21:27] <wtay> BBB: you're supposed to be able to AND out the right color using the mask [21:28] <BBB> yes, that's what I'm doing, right? [21:28] Action: BBB is extremely confused now [21:28] <wtay> so, if you read the big endian guint32 in memory and you mask with 0xFF000000 you get the red color [21:28] <BBB> (* (guint32 *) GST_BUFFER_DATA(buffer)) & 0x000000ff = red pixel byte [21:29] <wtay> BBB: not if the data is big endian [21:29] <ds-work> 24/32 bit video is always big-endian [21:29] <BBB> I was talking about little endian [21:29] <ds-work> by definition [21:29] <BBB> ;) [21:29] <wtay> I would treat it like big endian and treat the masks as big endian too then [21:30] <BBB> ok, but this means that on little endian native RGB32, red_mask is 0x000000ff, and on big endian native RGB32, red_mask is 0xff000000 [21:30] <wtay> yes [21:31] <BBB> *pheew*, I still get it ;) [21:31] <wtay> that's why the endianness is also important [21:31] <wtay> what you do is read the data little endian and treat the ... [truncated message content] |