IRC Logs

*******************************************************************
[03:01] sub_pop ([email protected]) left irc: "Client exiting"
[03:09] harshy ([email protected]) left irc: Remote closed the connection
[03:12] <ds-work> I doubt it
[03:12] <Company> you doubt what?
[03:12] <ds-work> there have been remote root exploits that have been left unpatched for long time
[03:12] <ds-work> in the Hurd
[03:13] <ds-work> unless it's been fixed, there's a local root exploit that's been around since 1996
[03:13] <Company> you might want to read #gnome-hackers btw (including some backlog)
[03:16] <ds-work> about gtk?
[03:16] <Company> yeah
[03:16] <Company> and about their ideas about releases
[03:17] <Company> because that's what they'll demand from us, too
[03:18] sub_pop ([email protected]) joined #gstreamer.
[03:24] <ds-work> I agree 100% about time-based releases
[03:25] <ds-work> btw, if we want 0.10 on a 6-month schedule, we need to start working on 0.9 as soon as 0.8 is out
[03:25] <ds-work> at least, you and I do
[03:26] <Company> i dunno
[03:26] <Company> we need to work on elements
[03:26] <ds-work> work on elements in 0.8, work on the core in 0.9
[03:26] <Company> like autopluggers, basic classes and so on
[03:27] <Company> i dunno what i'll be working on, but i know that a 6-month cycle is a bit hard to get all of this done
[03:27] <ds-work> ideally, I'd like to not change the API significantly in 0.9, so that elements can be developed in parallel for 0.8 and 0.9
[03:28] <Company> i don't think the API will change much for either apps or plugins
[03:28] <ds-work> yeah
[03:28] <ds-work> imo, that's the best bet for the next release cycle
[03:29] <ds-work> either that, or keep the core essentially frozen for 6 months
[03:29] <ds-work> because we really need a bunch of time to work on elements
[03:29] <ds-work> without the aggrevation that happened in 0.7's release cycle
[03:30] <Company> we should work on elements
[03:30] <Company> and freeze the core
[03:30] <ds-work> speaking of which, I should probably go home and work on the new autoplugger
[03:30] <Company> so we have a better idea what we miss in the core
[03:31] <ds-work> I'd be ok with that, too
[03:31] <Company> we'll probably get really frustrated with something after 3 months anyway and then work on the core again ;)
[03:31] <ds-work> but I think there's going to be a serious need to fix, say, clocking
[03:32] <Company> and threading
[03:32] <Company> and state changes
[03:32] <Company> but currently i want my media played
[03:32] <ds-work> I already have a long list of stuff in docs/random/ds/0.9-suggested-changes
[03:32] <Company> qtdemux and asfdemux are worse than that...
[03:33] <Company> yeah, i've seen this file mentioned, but not looked at it
[03:33] <ds-work> hrm, maybe I should fix qtdemux tonight instead
[03:34] <Company> i'll continue to just hack on what i want to btw
[03:34] <Company> so i have no clue what i'll be working on next...
[03:35] <ds-work> that's essentially what I do.  But when I have a list of goals written down, I tend to think more about those problems, and naturally start thinking about solving those problems
[03:36] <Company> there are more things i have solved in my head but not done than there are things that are unsolved
[03:36] <Company> i suffer from the "i have an idea" desease
[03:37] <Company> like GstTreeScheduler
[03:38] Action: ds-work suffers from "I have a vague idea" disease
[03:39] <Company> most of the caps stuff was done beautifully
[03:39] <Company> so i can't complain
[03:39] <Company> i'm going to bed now btw
[03:39] <ds-work> ok, I'm going home
[03:39] <ds-work> later
[03:40] <Company> gnight
[03:40] Company ([email protected]) left irc: Remote closed the connection
[03:42] mxpxpod ([email protected]) joined #gstreamer.
[03:56] sub_pop ([email protected]) left irc: "Client exiting"
[03:59] sub_pop ([email protected]) joined #gstreamer.
[04:05] thaytan ([email protected]) left irc: Read error: 113 (No route to host)
[04:08] thaytan ([email protected]) joined #gstreamer.
[04:17] thomasvs ([email protected]) left irc: Read error: 60 (Operation timed out)
[04:26] thomasvs ([email protected]) joined #gstreamer.
[04:29] thaytan ([email protected]) left irc: Read error: 113 (No route to host)
[04:30] thaytan ([email protected]) joined #gstreamer.
[04:31] <bitshifter> is there any way to link an oggdemux element to a (compatible) follow-up element that doesn't involve connecting to 'new-pad' or using spider?
[04:35] jcsston ([email protected]) joined #gstreamer.
[04:51] thaytan ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[05:19] <ds> bitshifter: no
[05:19] <bitshifter> k, thanks
[05:30] thaytan ([email protected]) joined #gstreamer.
[05:39] Marsupilami23 ([email protected]) joined #gstreamer.
[06:22] tlg ([email protected]) left irc: "Client exiting"
[06:49] spyder482 ([email protected]) left irc: "Leaving"
[06:51] <bluejay> bitshifter: I managed to run oggdemux ! vorbisdec ! float2intnew
[07:17] harshy ([email protected]) joined #gstreamer.
[07:19] ChrisHJW ([email protected]) joined #gstreamer.
[07:19] harshy ([email protected]) left irc: Client Quit
[07:29] harshy ([email protected]) joined #gstreamer.
[07:29] Nick change: mathrick|sleep -> mathrick
[07:29] <mathrick> mornin
[07:56] ChrisHJW ([email protected]) left irc: 
[08:07] thomasvs ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:28] harshy ([email protected]) left irc: "I quit for now"
[08:56] Company ([email protected]) joined #gstreamer.
[08:57] mathrick ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[09:01] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[09:07] somex1 ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[09:24] kmaraas ([email protected]) left irc: "Leaving"
[09:39] Company ([email protected]) left irc: Remote closed the connection
[10:02] sxpert ([email protected]) left irc: "Client exiting"
[10:05] harshy ([email protected]) joined #gstreamer.
[10:17] pmfp ([email protected]) joined #gstreamer.
[10:17] pmfp_ ([email protected]) joined #gstreamer.
[10:17] pmfp ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[10:17] Nick change: pmfp_ -> pmfp
[10:19] markey ([email protected]) joined #gstreamer.
[10:25] markey ([email protected]) left irc: "bbl"
[10:34] Shoragan ([email protected]) joined #gstreamer.
[10:40] wheels ([email protected]) left irc: "work"
[11:07] dolphy ([email protected]) joined #gstreamer.
[11:09] <dolphy> morning
[11:10] <pmfp> mornin'
[11:36] thaytan ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[11:42] thaytan ([email protected]) joined #gstreamer.
[11:44] ain ([email protected]) joined #gstreamer.
[12:09] <pmfp> where did the spider documentation go?
[12:13] Rotty ([email protected]) joined #gstreamer.
[12:17] <pmfp> hmm... heh, ok, that was harder said than done :)
[12:17] <pmfp> I thought you had to explicitly add factories to a list in the spider
[12:17] <pmfp> damn this rocks :))
[12:27] alley_cat ([email protected]) left irc: "May the Source be with you!"
[12:32] thaytan ([email protected]) left irc: Remote closed the connection
[12:36] thaytan ([email protected]) joined #gstreamer.
[13:17] wingo ([email protected]) joined #gstreamer.
[13:25] <wingo> hey all.
[13:25] <wingo> ain: i just wrote in some things to make audioconvert do intfloat conversion, took your oneton and made it do float, and wrote a new interleaver --
[13:26] <thaytan> hi, wingo!
[13:26] <wingo> I'm going to put my revisions to oneton into a new plugin in gst/interleave, called deinterleave
[13:26] <wingo> then we can deprecate int2float, float2int, and oneton
[13:26] <wingo> ain: does that sound fine to you?
[13:27] <wingo> howdy, thaytan :)
[13:27] <thaytan> I don't know that anyone else is awake
[13:27] <thaytan> it's been pretty quiet in here
[13:27] <wingo> it seems that way ;)
[13:28] <wingo> hm, they're not paying me on time and I have like 5 namibian dollars in my pocket. that's about US$0.80.
[13:28] <thaytan> hey, that's about AU 1$ :)
[13:29] <thaytan> I'm pleased, because that's the strongest our dollar has been in years
[13:31] <wingo> i guess the low us dollar is good for industry in the us, but it makes travellers suffer :/
[13:31] <wingo> combine that with high rand prices -- whoa is this CNN or something? ;)
[13:32] <thaytan> heh
[13:32] <thaytan> it's a pity it's not the euro... an boost in currency exchange in time for our trip in April would be welcome ;)
[13:36] <ain> wingo: sure
[13:41] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[13:45] sublett ([email protected]) joined #gstreamer.
[13:54] ensonic ([email protected]) joined #gstreamer.
[13:54] <ensonic> hihi
[14:01] <wingo> yo
[14:24] alley_cat ([email protected]) joined #gstreamer.
[14:26] spyder482 ([email protected]) joined #gstreamer.
[14:38] thaytan ([email protected]) left irc: Read error: 113 (No route to host)
[14:39] thaytan ([email protected]) joined #gstreamer.
[14:43] Action: ensonic is just reading the pwg for cmpletness and is wondering if it wouldn't be better to have one guide describing concepts (elements,pads,caps) and one for the plugins
[14:47] <wingo> check the app development guide for that.
[14:49] <ensonic> wingo: was i was trying to say, that reading the pwg is useful, even if one does not want to write a plugin ;-) (and vice versa)
[15:17] teuf ([email protected]) joined #gstreamer.
[15:17] <teuf> hi
[15:17] wheels ([email protected]) joined #gstreamer.
[15:18] <wheels> where are the list archives now?
[15:21] <wingo> dunno, yo
[15:21] <ensonic> can anyone enlighten me how I tell gst-launch to do something like 'sinesrc freq=100' & 'sinesrc freq=200' -> adder -> ...
[15:22] Action: wheels assumes that they've ended up somewhere under FD.o...
[15:22] <wingo> sinesrc freq=100 ! adder ! osssink sinesrc freq=100 ! adder0.sink1
[15:25] <ensonic> wingo: not really is there a delimiter required between osssink and the new sinesrc?
[15:26] <wingo> nope
[15:26] <ensonic> seems the adder is broken
[15:26] <ensonic> gst-launch sinesrc freq=500 ! adder ! audioconvert ! esdsink fails already
[15:29] ibsteveog ([email protected]) joined #gstreamer.
[15:29] markey ([email protected]) joined #gstreamer.
[15:33] mxpxpod ([email protected]) left irc: "Guns kill people as much as spoons made Rosie O'Donnell fat.."
[15:35] <wingo> could be, i haven't touched it in a while
[15:38] <ensonic> wingo: in gst-editor it is just silence and I get "internal error: push on pad adder0.sink0 but it has no chainhandler"
[15:39] <wingo> dunno bro, hit the code :)
[15:40] <ensonic> wingo: i will, just tried if maybe someone says 'ohh that, just do this and that then' ;-)
[15:40] markey ([email protected]) left irc: "leaving"
[15:42] alley_cat ([email protected]) left irc: "May the Source be with you!"
[15:44] <wingo> aight, i'm out.
[15:44] wingo ([email protected]) left irc: "I like core dumps"
[15:45] <KoRnouille> hi. When I create a pipeline : pipeline = gst_pipeline_new ("pipeline");
[15:45] <KoRnouille>   is "pipeline" the name given to the pipe ?
[15:46] <KoRnouille> the pipeline inside the parenthese
[15:47] <ensonic> KoRnouille: yes, you can pass any name
[15:47] <KoRnouille> alright... it's juste the designation
[15:47] <ensonic> gst_pipeline_new("gst is cool") would work too
[15:47] <KoRnouille> :)
[15:48] <KoRnouille> I have a C question : GstElementFactory *factory; this is going to decalre a variable called factory, right ? what's the * for ?
[15:50] <bluejay> KoRnouille: the 'factory' variable _points_ to a memory block that holds a GstElementFactory.
[15:50] <bluejay> KoRnouille: * means pointer
[15:50] <KoRnouille> woah... I don't have this concept yet...
[15:50] <bluejay> KoRnouille: You need a good C book.
[15:51] <KoRnouille> I know... i ordered one yesterday... i should receive it next week
[15:51] <KoRnouille> ;)
[15:51] <ensonic> KoRnouille: 'int i' creates a thing called 'i' that is of type 'int'
[15:51] <KoRnouille> right, I know what
[15:51] <ensonic> KoRnouille: 'int *i' creates a pointer to such a thing
[15:51] <ensonic> KoRnouille: like a reference
[15:51] <KoRnouille> hummm...
[15:51] <KoRnouille> ok
[15:51] <KoRnouille> im not sure I get the use of it...
[15:52] <ensonic> KoRnouille: it is useful so that one not pases things around, but just tell where it is
[15:52] <bluejay> KoRnouille: the memory for 'int i' is located in the current stack frame. So when the current function ends, that i is gone.
[15:52] <ensonic> KoRnouille: like if you would ask something and I just tell you 'look at http://www....' then the URL is a pointer to the information
[15:53] <KoRnouille> I see
[15:54] <bluejay> KoRnouille: int*i stores the address of some memory somewhere else. You have to manually allocate it (g_malloc, g_object_new, etc), but what the address (URL-ish) that you then store in i can be given to other people
[15:54] <bluejay> s/what//
[15:56] <KoRnouille> hummm...
[15:56] <KoRnouille> i guess Ill have to practice
[15:57] <teuf> KoRnouille: looks like you need to read a C book before looking too much into gstreamer ;)
[15:57] <KoRnouille> I understand what means a pointer... but I can hardly understand the use
[15:57] <KoRnouille> yeah teuf
[15:57] <KoRnouille> :)
[15:57] <KoRnouille> im proud, i was able to display my 4 screens full screen correctly !!
[15:57] <KoRnouille> that's more a gtk issue
[15:57] <KoRnouille> more than a gstreamer
[15:58] <KoRnouille> i just need to find out how to hide the cursor
[15:58] <KoRnouille> right now, I'm trying to understand why I get an mpeg2dec error when reading an mpegfile...
[15:58] <KoRnouille> so I'm rewriting the code just like in the gstreamer API
[15:58] alley_cat ([email protected]) joined #gstreamer.
[15:59] <KoRnouille> when I have this : filesrc = gst_element_factory_make ("filesrc", "disk_source");
[15:59] <KoRnouille> I get : "filesrc =" is the variable
[16:00] <KoRnouille> gst_element_factory_make is the function that's going to make my object which is going to be assigned to the variable
[16:00] <KoRnouille> and what's in () are arguments for the gst_element_factory_make to play with
[16:00] <KoRnouille> right ?
[16:01] <teuf> yep
[16:01] <KoRnouille> so that argument filesrc is to tell gst_element_factory_make which element to make... and "disk_source" is juste the name given to it ? just like ("pipeline") ?
[16:02] <teuf> the filesrc variable will point to the actual object,
[16:02] <KoRnouille> ok
[16:02] <KoRnouille> what's the find function for ?
[16:02] <KoRnouille> factory = gst_element_factory_find ("mpeg2dec"); ?
[16:02] <KoRnouille> what's the use ?
[16:02] <teuf> KoRnouille: it's probably explained in the api doc...
[16:03] <KoRnouille> just make an gst-inspect, check if you have the codec, and write it directly without making a find, no ?
[16:03] <KoRnouille> yeah, the API says it's to find the codec
[16:05] <KoRnouille> oooh...
[16:05] <KoRnouille> i just figured out...
[16:05] <KoRnouille> that make the program compatible with other configurations. Not everyone it meant to have mpeg2dec I guess...
[16:08] <bluejay> I sent a suggestion to the list to allow "tracknum" as well as "tracknumber" in vorbis tags. Did it get through?
[16:09] <teuf> bluejay: why should tracknum be allowed ?
[16:09] <bluejay> teuf: I want it because I downloaded a Magnatune album that used it. But it does seem like a natural fallback
[16:10] <bluejay> teuf: If you think it's a bad idea, I'll just change the tags in the album; I just wanted to see what people thought.
[16:10] <teuf> bluejay: http://www.xiph.org/ogg/vorbis/doc/v-comment.html only mentions tracknumber...
[16:11] <KoRnouille> okay, last C question please... client->video_sink... what does that arrow means ? a link ? ->
[16:11] <teuf> KoRnouille: pointer dereference, it's equivalent to (*client).video_sink
[16:12] <teuf> where *client refers to the struct (object) pointed to by the pointer client
[16:12] <bluejay> teuf: thanks for that link, I'd been wondering if they were documented. So . . . you think I should complain to magnatune instead?
[16:12] <KoRnouille> aaah.
[16:13] <teuf> bluejay: if they encoded the vorbis file, yeah
[16:13] <bluejay> teuf: Thanks. :)
[16:16] <KoRnouille> gst_element_factory_find () returns GstElementFactory if found, NULL otherwise. Does it mean that the string "NULL" returns if not found ?
[16:16] <spyder482> anyone here know why the bedian glib2.0 dev package seems to be missing glibconfig.h?
[16:17] <teuf> spyder482: are you sure it's missing ? iirc it's hidden somewhere in lib/glib
[16:17] <spyder482> debian, not bedian ;)
[16:17] <spyder482> hmmm
[16:17] <teuf> KoRnouille: NULL is an "empty" pointer
[16:17] <spyder482> ok i'll look
[16:17] <KoRnouille> can you create a variable from an empty pointer ?
[16:18] <spyder482> ah
[16:18] <spyder482> found it
[16:18] <teuf> spyder482: dpkg -S glibconfig.h is your friend ;)
[16:18] <spyder482> :)
[16:18] <teuf> or dpkg -L libglib2.0-dev |grep glibconfig
[16:18] <spyder482> i'm still adjusting to debian ;)
[16:19] <spyder482> is there some easy way to include these files that are scattered everywhere or should i just copy them back to /usr/include ?
[16:20] <teuf> hm ? do not copy them to /usr/include ;)
[16:20] <teuf> including glib.h should be enough
[16:20] <teuf> (or glib/glib.h)
[16:20] <spyder482> well, then it complains of no glibconfig.h ;)
[16:20] <teuf> what is "it" ?
[16:20] <spyder482> gcc
[16:21] <spyder482> ;)
[16:21] <KoRnouille> factory = gst_element_factory_find ("mpeg2dec"); if mpeg2dec is not found, will the facotry variable will be assigned ? to test it, can I do a "if factory" to know if it exists or not ?
[16:21] <KoRnouille> something like that : if(!factory);
[16:21] <KoRnouille>     printf "mpeg2dec not found !";
[16:21] <KoRnouille> exit(1)
[16:21] <teuf> did you pass it the proper flags ?
[16:22] <spyder482> i guess i have to tell it to look in /usr/lib/glib-2.0/include also?
[16:22] <spyder482> i just prefer to not have to give 30 locations for headers...
[16:23] <teuf> spyder482: did you pas $(pkg-config --cflags --libs glib-2.0) to gcc ?
[16:23] <spyder482> AH!
[16:23] <spyder482> thanks :)
[16:24] ChrisHJW ([email protected]) joined #gstreamer.
[16:24] <teuf> spyder482: and I guess the glibconfig.h file went into /usr/lib because it's platform dependent
[16:29] thomasvs ([email protected]) joined #gstreamer.
[16:38] alley_cat ([email protected]) left irc: "May the Source be with you!"
[16:41] <KoRnouille> how do I write this in C ?
[16:41] <KoRnouille> video/x-raw-yuv,width=640,height=480
[16:42] <KoRnouille> what plugin am I calling ?
[16:43] <spyder482> char * blah = "video/x-raw-yuv"; int width = 640; int height = 480; <-- somthing like that...maybe using glib types
[16:43] <spyder482> unless you mean something totally different...
[16:43] yippi ([email protected]) joined #gstreamer.
[16:43] <KoRnouille> it's for resising a video
[16:44] <KoRnouille> videoscale only give the scaling method
[16:46] <KoRnouille> so someone told me to use the "... ! video/x-raw-yuv,width=640,height=480 ! ..." syntax for scaling.
[16:46] alley_cat ([email protected]) joined #gstreamer.
[16:47] <KoRnouille> are they caps ?
[16:48] <spyder482> teuf: how do you do the same for libxml2?
[16:48] <teuf> pkg-config --cflags --libs libxml2 or something like taht
[16:49] <teuf> $(pkg-config --cflags --libs libxml-2.0 glib-2.0)
[16:49] <spyder482> ah
[16:49] <teuf> the known libs are listed in /usr/lib/pkg-config
[16:49] <spyder482> the -2.0
[16:50] <spyder482> i tried with libxml2
[17:00] ibsteveog ([email protected]) left irc: "quitting"
[17:04] markey ([email protected]) joined #gstreamer.
[17:05] jcsston ([email protected]) joined #gstreamer.
[17:07] kwm ([email protected]) joined #gstreamer.
[17:07] mxpxpod ([email protected]) joined #gstreamer.
[17:08] markey ([email protected]) left irc: Client Quit
[17:18] Norman ([email protected]) joined #gstreamer.
[17:18] <Norman> hi!
[17:18] <Norman> anybody here ? - I just would like to ask a little question about gst
[17:18] <teuf> Norman: ask it, and wait until you get an answer (or get bored)
[17:21] <Norman> how do I record a movie from tv ( bttv / v4l ) to a avi file using the ( windows ) hyffuv video codec and no audio compression
[17:23] <Norman> I would like to know the correct gst-launch command - or is there info on this somewhere on the net ?
[17:27] <spyder482> what would be the easiest way to do 2pass normalization with gstreamer
[17:30] yippi ([email protected]) left irc: Remote closed the connection
[17:32] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[17:37] Norman ([email protected]) left irc: 
[17:38] kmaraas ([email protected]) joined #gstreamer.
[17:39] ensonic ([email protected]) left #gstreamer.
[17:40] yippi ([email protected]) joined #gstreamer.
[17:40] sub_pop ([email protected]) left irc: "Client exiting"
[17:48] markey ([email protected]) joined #gstreamer.
[17:53] thomasvs ([email protected]) left irc: Connection timed out
[17:56] <KoRnouille> when I do an filesrc ! mpegparse, what comes out of mpegparse ?
[17:57] <KoRnouille> is it some kinda "cat file.mpg" to cat the file at the right mpeg speed ?
[17:58] <spyder482> hmm, is gst_element_unref still used? because gcc can't find it...
[17:58] <KoRnouille> spyder482: yeah, I bumped into the same problem a few hours ago... but I didn't want to bother you guys thinking the problem came from me...
[17:59] <KoRnouille> :P
[17:59] <teuf> spyder482: doesn't seem to exist in my gstreamer head source tree
[17:59] <spyder482> how do you unref an element then?
[17:59] <spyder482> the guide says do this ;)
[17:59] <KoRnouille> I didn't unref it...
[17:59] <teuf> gst_object_unref I guess
[18:00] <spyder482> hmmm
[18:00] <spyder482> k
[18:00] alley_cat ([email protected]) left irc: "May the Source be with you!"
[18:00] <KoRnouille> for my case, it was just for playing around with the code... more for testing...
[18:01] teuf ([email protected]) left irc: "Client exiting"
[18:01] <spyder482> hmm, warns of incompatible pointer type...i'm unfamiliar with glib so how do i cast the element to a object?
[18:03] alley_cat ([email protected]) joined #gstreamer.
[18:03] <KoRnouille> is there a way I can make a search ?
[18:03] <KoRnouille> in the docs ?
[18:03] <KoRnouille> im looking for gst_bin_iterate
[18:03] <bitshifter> spyder482: GST_OBJECT(yourelement)  probably
[18:03] <KoRnouille> I guess I'll have to download the PDF
[18:03] <spyder482> k
[18:04] mxpxpod ([email protected]) left irc: "Leaving"
[18:07] <spyder482> oh, i'm stupid :P
[18:07] <spyder482> i had a typo in the plugin name
[18:08] <spyder482> so it was a NULL i was trying to unref ;)
[18:11] <KoRnouille> aaah. I guess I'm feeling the same way you are.
[18:11] <KoRnouille> :)
[18:12] <spyder482> yay! i can open files now ;)
[18:12] <KoRnouille> :D
[18:14] <spyder482> now i need to recompile the plugins to get the mad decoder built
[18:17] <KoRnouille> make sure you have id3tag also
[18:17] <KoRnouille> mine would build mad because id3tag was missing
[18:18] <KoRnouille> s/would/wouldn't/
[18:18] <spyder482> k thanks
[18:25] Rotty ([email protected]) left irc: Read error: 110 (Connection timed out)
[18:28] <KoRnouille> what's the syntax to ... ! mpegdemux video_%02d ! {queue ! mpeg2dec ...} ? I only want the video stream from the demux...
[18:30] <spyder482> E0 i think
[18:30] <KoRnouille> to create pads, do I have to be in GST_STATE_PAUSED ?
[18:30] <spyder482> hmmm
[18:30] <spyder482> maybe it's not in this case ;)
[18:30] <KoRnouille> E0 ?
[18:31] <spyder482> the video stream in an mpeg ps usually has an id of 0xE0
[18:31] <spyder482> but maybe that's not what you mean
[18:31] <KoRnouille> huh. I don't know how to do that
[18:31] <KoRnouille> yes, it's an mpeg PS tream
[18:32] <spyder482> i don't know how that plugin labels the pads
[18:32] markey ([email protected]) left irc: "leaving"
[18:32] <KoRnouille> k
[18:33] <KoRnouille> does that xml graphics builder writes the code for us behind when everything is connected together ?
[18:33] <KoRnouille> ;)
[18:47] <KoRnouille> warning : error scheduling group 0x8243a10 ---- What is this supposed to mean ?
[18:50] Rotty ([email protected]) joined #gstreamer.
[19:18] smoke ([email protected]) joined #gstreamer.
[19:18] <KoRnouille> I need a guru
[19:18] <bitshifter> I am observing something weird with oggdemux: If I do  gst-launch-0.7 filsrc location=ogg ! oggdemux ! md5sink  I get (as expected) a different hash than with   filesrc=ogg2 ! oggdemux ! md5sink
[19:18] <bitshifter> ogg1 and ogg2 only differ in that they have different vorbis comments
[19:18] <KoRnouille> Im' having a hardtime connecting my pads together
[19:18] <bitshifter> and oggdemux passes on the vorbis comments (as expected) to the md5sink
[19:19] <KoRnouille>   GstPad *srcpad, *sinkpad;
[19:19] <KoRnouille>   srcpad = gst_element_get_pad (video_demux, "video_0E");
[19:19] <KoRnouille>   sinkpad = gst_element_get_pad (mpg_decoder, "sink");
[19:19] <bitshifter> however, when I plug my own hash sink plugin after oggdemux (also caps 'any'), then oggdemux seems to filter out vorbis comments (and some other stuff)
[19:19] <bitshifter> and I would very much like to understand what is happening there
[19:20] smoke_ ([email protected]) joined #gstreamer.
[19:20] smoke_ ([email protected]) left irc: Client Quit
[19:20] markey ([email protected]) joined #gstreamer.
[19:21] <bitshifter> (ie. my own hash sink produces identical hashes for file 'ogg' and 'ogg2'), and if I dump the incoming buffers to file then the vorbis comments are also missing, while if I use gst-launch-0.7 filesrc location=ogg ! oggdemux ! filesink location=out.vorbis, then the vorbis comments are not missing
[19:22] <spyder482> hmm, what's the correct way to gst_element_factory_make an alsa sink element
[19:22] alley_cat ([email protected]) left irc: "May the Source be with you!"
[19:22] <ds-work> bitshifter: see if the two pipelines end up with different caps on oggdemux.src
[19:24] jcsston ([email protected]) joined #gstreamer.
[19:27] <KoRnouille> so is my syntax correct ?
[19:27] <spyder482> nm about the alsa sink thing
[19:28] <KoRnouille> we all want to ask something... :)
[19:28] <spyder482> video_0E?
[19:28] <KoRnouille> let me check spyder
[19:28] <spyder482> i got me app working
[19:28] <spyder482> it plays mp3 now ;)
[19:29] <spyder482> but i want to make it decode,normalize, and then encode to vorbis ;0
[19:31] <KoRnouille> i don't have the alsa
[19:32] <spyder482> it's gst_element_factory_make("alsasink", "play_audio");
[19:33] <KoRnouille> so it (video_demux, "video_0E"); looking right ?
[19:33] <bitshifter> ds-work: the gst-launch pipeline says the caps on oggdemux0.src are NULL. With my own pipeline, the debug output says 'gst_opt_scheduler_pad_link: NULL pad link between "oggdemux:serial_171xxxxxxx" and "hashsink:sink" (is that the right line?)
[19:33] <KoRnouille> s/it/is/
[19:36] <bitshifter> ds-work: hmm, in the oggdemux "new-pad" callback I get caps as "audio/x-vorbis" though
[19:37] alley_cat ([email protected]) joined #gstreamer.
[19:38] <bitshifter> ds-work:  and if I do  gst-launch-0.7  filesrc ! oggdemux ! audio/x-vorbis ! md5sink   I still get different hashes for both files
[19:38] <spyder482> KoRnouille: i think it would be E0 if it was labeled by that ID
[19:38] <spyder482> but it may not be that way
[19:38] <KoRnouille> i really don't know
[19:38] <KoRnouille> gst_element_link_many (udp_src, video_demux, mpg_decoder, colorspace, client->video_sink, NULL);
[19:38] <KoRnouille> this seems to be enought
[19:45] <taaz> hmm... --gst-debug-help prints some of this:
[19:45] <taaz> ** (process:10957): CRITICAL **: how to remove plugins?
[19:46] <taaz> was trying to play a .mov and get lots of:
[19:46] <taaz> ** (player.py:10993): WARNING **: flush event
[19:46] <taaz> anyone know offhand what that means?
[19:46] apoc__ ([email protected]) joined #gstreamer.
[19:47] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[19:51] jcsston ([email protected]) joined #gstreamer.
[19:51] <ds-work> taaz: rerun gst-register
[19:53] <KoRnouille> trying to set non fixed caps on pad movie_src:src, not allowed (movie_src is a udpsrc element)
[19:53] smoke ([email protected]) got netsplit.
[19:59] smoke ([email protected]) got lost in the net-split.
[20:00] sublett ([email protected]) left irc: "I like food, food is good!"
[20:03] apoc_ ([email protected]) left irc: Connection timed out
[20:03] <bitshifter> ds-work: any more ideas? :)
[20:04] <ds-work> no
[20:06] somex1 ([email protected]) joined #gstreamer.
[20:09] <bitshifter> maybe the first buffer that oggdemux gets and uses to determine the src pad type gets lost somehow?  I'm currently doing this:  setting up  filesrc ! oggdemux, then playing the pipelin until I get a 'new-pad' signal, then pausing the pipelin, then adding and linking my "hashsink", and set the pipeline to 'playing' again. Are there any additional measures I have to take to ensure the first buffer gets passed on?
[20:12] smoke ([email protected]) joined #gstreamer.
[20:14] <bitshifter> grrr, and I get segfaults whenever I set the pipeline to READY or NULL (from PAUSED) before setting it to PLAYING again
[20:15] Action: bitshifter cvs updates
[20:17] <smoke> how do i fix the missing bar keys in kernel 2.6 ?
[20:18] <smoke> and first of all why are they missing?
[20:28] Jara[zZ] ([email protected]) got netsplit.
[20:28] KoRnouille ([email protected]) got netsplit.
[20:28] ds-work ([email protected]) got netsplit.
[20:28] herzi ([email protected]) got netsplit.
[20:31] teuf ([email protected]) joined #gstreamer.
[20:34] ds-work ([email protected]) got lost in the net-split.
[20:34] herzi ([email protected]) got lost in the net-split.
[20:34] KoRnouille ([email protected]) got lost in the net-split.
[20:34] Jara[zZ] ([email protected]) got lost in the net-split.
[20:36] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[20:38] Jara[zZ] ([email protected]) joined #gstreamer.
[20:38] KoRnouille ([email protected]) joined #gstreamer.
[20:38] ds-work ([email protected]) joined #gstreamer.
[20:38] herzi ([email protected]) joined #gstreamer.
[20:38] <spyder482> (process:8731): GStreamer-WARNING **: internal error: push on pad encoder:sink but it has no chainhandler
[20:38] <spyder482> what would this mean?
[20:38] <spyder482> it says this at the very last iteration
[20:41] BBB ([email protected]) joined #gstreamer.
[20:42] <spyder482> BBB!
[20:42] <spyder482> i have a question for you
[20:43] <teuf> hi BBB
[20:43] <spyder482> WARN          default( 8747) ebml-write.c(233):gst_ebml_write_element_size: NULL Invalid size, maximizing
[20:43] <spyder482> why do i get a bunch of these?
[20:45] <BBB> er....
[20:45] <BBB> depends on your file, I guess
[20:45] <BBB> but you really want to send me more than just the warning
[20:46] <BBB> ;)
[20:46] <spyder482> most likely a coding error ;)
[20:46] <spyder482> i'm trying to make an audio transcoder
[20:46] <BBB> use --gst-fatal-warnings to make it error our
[20:46] <BBB> then you can use a debugger to debug it
[20:47] <BBB> or send me the file (via bugzilla please) so I can look at it
[20:47] ds-work ([email protected]) got netsplit.
[20:47] KoRnouille ([email protected]) got netsplit.
[20:47] herzi ([email protected]) got netsplit.
[20:47] Jara[zZ] ([email protected]) got netsplit.
[20:47] Action: BBB almost has a new laptop screen
[20:47] herzi ([email protected]) returned to #gstreamer.
[20:47] <spyder482> it may not be connecting the pads..would that cause this?
[20:47] KoRnouille ([email protected]) returned to #gstreamer.
[20:48] <BBB> I'm not sure
[20:48] Action: BBB checks CVS
[20:49] <spyder482> (process:8787): GStreamer-WARNING **: internal error: push on pad encoder:sink but it has no chainhandler
[20:49] <spyder482> does this mean something isn't connected?
[20:49] <BBB> possibly
[20:49] <BBB> it can mean a lot of things
[20:49] hallibaby ([email protected]) joined #gstreamer.
[20:49] <BBB> your warning means an invalid size
[20:49] <BBB> or maybe my coding style isn't accepted by the win32 compiler
[20:49] <pmfp> If two threads are working on the same GList, if one thread uses g_list_next, will the change affect the other thread?
[20:50] hallibaby ([email protected]) left irc: Client Quit
[20:50] <BBB> pmfp: no
[20:50] <pmfp> BBB: great, thanks!
[20:50] <BBB> unless you write the result into a pointer used by both
[20:50] ds-work ([email protected]) returned to #gstreamer.
[20:50] Jara[zZ] ([email protected]) returned to #gstreamer.
[20:50] <BBB> but reading can be done by two threads
[20:51] <spyder482> BBB: not win32 ;)
[20:51] <spyder482> i'm working on linux ;0
[20:51] <BBB> oh!
[20:51] <BBB> hm
[20:51] <BBB> in that case
[20:51] <BBB> what are you doing?
[20:51] <spyder482> lol
[20:51] <BBB> you shouldn't write elements of size 0
[20:51] <pmfp> BBB: the question was whether it would point to the same place in both threads if you use g_list_next() in one of them to change the index
[20:52] <BBB> GST_WARNING ("Invalid size, maximizing");
[20:52] <BBB> that line in ebml-write.c
[20:52] <BBB> add %d, size somewhere over there
[20:52] <BBB> or add something so that it works with elements of size 0
[20:52] <spyder482> BBB: i'm trying to connect filesrc->mad decoder->mpegaudio encoder->matroskamux->filesink
[20:52] <BBB> I ususally don't accept those
[20:52] <BBB> because an element of size 0 is broken
[20:52] <BBB> but anyway
[20:52] <spyder482> i think maybe the muxer isn't connected...
[20:52] <BBB> maybe 0-size audio buffers...?
[20:52] Action: BBB doesn't know
[20:53] <BBB> no, this is different
[20:53] <spyder482> hmmm
[20:53] <BBB> (I think)
[20:53] <BBB> add some more debug information in the function that triggers the warning
[20:53] jcsston ([email protected]) joined #gstreamer.
[20:53] <spyder482> this is all i do:
[20:53] <spyder482> 	gst_element_link_pads(source, "src", decoder, "sink");
[20:53] <spyder482> 	gst_element_link_pads(decoder, "src", encoder, "sink");
[20:53] <spyder482> 	encOut = gst_element_get_pad(encoder, "src");
[20:53] <spyder482> 	muxIn = gst_element_get_request_pad(mux, "audio_%d");
[20:53] <spyder482> 	gst_pad_link(encOut,muxIn);
[20:53] <spyder482> 	gst_element_link_pads(mux, "src", dest, "sink");
[20:53] <spyder482> 	
[20:53] <spyder482> 	gst_element_set_state(pipeline, GST_STATE_PLAYING);
[20:54] <spyder482> the elements source,decoder,etc. are all setup above this
[20:54] <spyder482> then i have the loop to iterate the pipeline
[20:55] <spyder482> maybe the "audio_%d" isn't needed?
[20:55] smoke ([email protected]) left irc: Client Quit
[20:57] <BBB> why not just gst_element_link (filesrc, mad, enc, mux, sink, NULL);?
[20:57] <BBB> that'll work just as well
[20:57] smoke ([email protected]) joined #gstreamer.
[20:57] <spyder482> well, i started with that :P
[20:57] <BBB> oh
[20:57] <BBB> hm
[21:02] <spyder482> hmm, a bunch of "Sent 220 bits = 27 slots plus 4 220" lines...
[21:03] <spyder482> it seems to be muxing...
[21:04] jcsston` ([email protected]) joined #gstreamer.
[21:04] <spyder482> well, it wrote some matroska file ;)
[21:05] smoke ([email protected]) got netsplit.
[21:05] <spyder482> but it says samplerate is 0
[21:05] <spyder482> and it's only 380KB...which isn't much for a 192kbps mp3...
[21:06] <spyder482> it only worked when i connected the pads manually
[21:06] <spyder482> with the link_many it won't do that...
[21:06] <pmfp> is a GList NULL when it has reached the end?
[21:07] <teuf> pmfp: yep
[21:07] smoke ([email protected]) returned to #gstreamer.
[21:12] Zeenix ([email protected]) joined #gstreamer.
[21:13] <ds-work> Zeenix: fix massink!
[21:14] <Zeenix> hi
[21:14] <BBB> no
[21:14] <BBB> remove massink
[21:14] <ds-work> it's been removed
[21:14] <Zeenix> why? it did'nt work?
[21:15] <BBB> it accepted audio/x-wav
[21:15] <BBB> and mas isn't what it promised it'd be
[21:15] <BBB> so...
[21:15] <spyder482> BBB: your muxer doesn't connect without requesting an audio pad and then connecting manually
[21:15] <BBB> oh
[21:15] <BBB> ohwell...
[21:15] <ds-work> I don't mind having it around, as long as it works
[21:15] <BBB> maybe you're right
[21:15] <spyder482> :)
[21:16] <ds-work> but I know for certain that it doesn't work
[21:16] <Zeenix> yes, but is'nt MAS officially been accepted as the future part of X?
[21:16] <BBB> gotta work
[21:16] <BBB> bbl
[21:16] <BBB> zeenix: well, that's out of ignorance
[21:17] BBB ([email protected]) left #gstreamer.
[21:17] <ds-work> Zeenix: no
[21:19] smoke ([email protected]) got netsplit.
[21:21] wheels ([email protected]) left irc: Read error: 110 (Connection timed out)
[21:23] <Zeenix> ds-work: then just throw it of
[21:23] <Zeenix> s/of/off
[21:23] foser ([email protected]) joined #gstreamer.
[21:24] <ds-work> ok, it's officially gone then
[21:25] smoke ([email protected]) got lost in the net-split.
[21:29] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[21:31] mxpxpod ([email protected]) joined #gstreamer.
[21:41] Jara[zZ] ([email protected]) got netsplit.
[21:41] Jara[zZ] ([email protected]) returned to #gstreamer.
[22:00] ZigZag ([email protected]) joined #gstreamer.
[22:00] <ZigZag> where can I get alsasink?
[22:01] <bitshifter> ah, indeed I somehow got oggdemux to swallow the first packet in my own pipeline. If I seek to 0 before playing the pipeline 'properly', I also get the vorbis comments delivered to my own hash sink
[22:03] Rotty ([email protected]) left irc: Read error: 60 (Operation timed out)
[22:10] Rotty ([email protected]) joined #gstreamer.
[22:10] <ds-work> ZigZag: gst-plugins
[22:11] jcsston` ([email protected]) left irc: Read error: 110 (Connection timed out)
[22:12] ZigZag ([email protected]) left irc: "Bye bye, folks!"
[22:16] kwm ([email protected]) left irc: "Leaving"
[22:17] jimmy_dean_ ([email protected]) joined #gstreamer.
[22:19] jimmy_dean ([email protected]) left irc: Read error: 110 (Connection timed out)
[22:24] Edulix ([email protected]) joined #gstreamer.
[22:25] <bitshifter> hmm, I've got a seek event (bytes format). If I send that to my hashsink with gst_element_send_event(), it works fine in some pipelines (e.g. filesrc ! mp3parse ! sink), but it doesn't work in an   filesrc ! oggdemux ! sink  pipeline. Oggdemux just 'swallows' the event instead of sending it on to the previous element.
[22:25] <bitshifter> however, when I Send it to oggdemux directly, _get_random_pad() will pick oggdemux's sink pad, and it will be sent on to filesrc, and everything works just fine
[22:26] <bitshifter> because I construct the pipeline dynamically with different elements, I don't really want to special case this unless absolutely necessary.  Is oggdemux supposed to swallow the upstream seek event on its src pad?  Or am I just doing it the wrong way?
[22:29] <ds-work> uh, you hash events?
[22:29] <bitshifter> no, I don't hash events
[22:29] <bitshifter> I hash data, but I want to do a seek to 0 before I start hashing
[22:30] <bitshifter> because oggdemux seems to swallow the first buffer somehow if I don't do that
[22:30] <ds-work> hmmm, odd
[22:30] <ds-work> it could be that oggdemux is pushing a buffer to the pad before you link it
[22:31] <bitshifter> I see
[22:32] <bitshifter> I am actually not linking it in the 'new-pad' callback, but only after gst_bin_iterate() returns
[22:33] <bitshifter> still, I don't understand why oggdemux drops my upstream seek event that arrives on its source pad
[22:35] <bitshifter> maybe because it's in GST_FORMAT_BYTES? 
[22:35] <ds-work> that's why, then
[22:36] <ds-work> oggdemux pushes a buffer in between firing the signal and the end of the iteration
[22:40] <ds-work> it should still seek correctly, though
[22:41] sbw_ibiblio ([email protected]) joined #gstreamer.
[22:42] ChrisHJW ([email protected]) left irc: Connection timed out
[22:46] edulix_ ([email protected]) joined #gstreamer.
[22:46] <bitshifter> from the source code, it appears that oggdemux just drops upstream events if they are not in GST_FORMAT_DEFAULT
[22:47] <bitshifter> line 333
[22:48] Nick change: edulix_ -> Edu|ix
[22:48] apoc_ ([email protected]) joined #gstreamer.
[22:50] <taaz> ds-work: register didn't fix it
[22:50] Action: ds-work disbelieves
[22:51] Edulix ([email protected]) left irc: Client Quit
[23:04] apoc__ ([email protected]) left irc: Connection timed out
[23:04] wheels ([email protected]) joined #gstreamer.
[23:14] Rotty ([email protected]) left irc: Remote closed the connection
[23:26] ibsteveog ([email protected]) joined #gstreamer.
[23:27] <ibsteveog> anyone know where I can find specs on what exactly the YUV420P format is?
[23:29] <bitshifter> google most likely ;)
[23:30] Edu|ix ([email protected]) left irc: Client Quit
[23:31] <spyder482> hmmm, is it fourcc.org that has the big yuv explanation?
[23:32] <ibsteveog> (the problem is that i can't seem to get ffmpeg to read it.. but I dunno where to go for help with ffmpeg)
[23:32] <ibsteveog> and i've been at fourcc.org
[23:32] <ibsteveog> i've scoured the page, but its mostly just fourcc codes...  when what I want is to know what yuv 420P is (which isn't showing up anywhere on that site)
[23:32] <spyder482> 420P would be like this IIRC:
[23:32] <spyder482> YY
[23:33] Last message repeated 1 time(s).
[23:33] <spyder482> UV
[23:33] <ibsteveog> so its packed, not planar?
[23:33] <spyder482> where the top is the full resolution
[23:33] <spyder482> no, it's planar
[23:33] <spyder482> the four Y's represent the whole Y plane
[23:33] <ibsteveog> i know..
[23:33] <ibsteveog> i got the 420 part
[23:33] <spyder482> the U and V are each 1/4 the res of the Y
[23:33] <ds-work> that's I420
[23:34] <ibsteveog> i'm worried about byte ordering and the possibility of 10 bits per component (like YV12 does)
[23:34] <spyder482> hmmm, I420 = YV12?
[23:34] <ibsteveog> which is I420?
[23:35] <ds-work> what you're calling YUV420P is probably what gstreamer calls 'I420'
[23:36] Action: ibsteveog sighs
[23:36] Action: ibsteveog quotes someone... "thats the great thing about standards... there are so many to choose from"
[23:37] <ibsteveog> so what is gstreamer's version of yuv 420P?
[23:37] <spyder482> he said probably I420
[23:37] <ds-work> it's not gstreamer's version, it's a standard fourcc value
[23:37] <spyder482> I420 isn't interleaved?
[23:37] <ds-work> no
[23:38] <spyder482> the I just is funny ;)
[23:38] <ds-work> the I is for Intel
[23:39] <spyder482> ah
[23:40] <spyder482> it's YV12 with the U and V swapped?
[23:40] <ibsteveog> ok... so i'm trying to pump this into ffmpeg... and it obviously doesnt like the YYYYUV planar format... but it was ok with YYUV planar (which i thought was 422P?)
[23:41] <ibsteveog> (just to be clear, i'm putting all Ys together, then all Us, then all Vs)
[23:41] <ibsteveog> ds-work, do you know where to find these standards?
[23:42] <ds-work> I don't know of any common planar formats that have U and V only vertically subsampled
[23:42] <ds-work> ibsteveog: fourcc.org
[23:42] <ibsteveog> hmm.. and yes, you were right about that example from spyder being I420P
[23:42] <ibsteveog> err, fourcc I420 (planar)
[23:43] <ibsteveog> so i still need to know what 420P is...
[23:43] <spyder482> ibsteveog: the way this works is you have all the Y's, the first line of the U then the first line of the V,U,V,U,V etc.
[23:44] <ibsteveog> ok, i'll give that a shot
[23:45] <spyder482> I think i'ts U then V
[23:45] <spyder482> maybe V then U
[23:45] <ds-work> btw, videotestsrc has programmatic definitions of most of the common YUV and RGB formats
[23:45] jcsston ([email protected]) joined #gstreamer.
[23:45] <spyder482> ibsteveog: it's laid out like i said
[23:45] <spyder482> YY
[23:45] Last message repeated 1 time(s).
[23:45] <spyder482> UV
[23:46] <spyder482> where all the Y is first, and there's 4 times the data
[23:46] <spyder482> 1 byt per pixel
[23:46] <spyder482> then the other 2 planes are stacked horizontally
[23:46] <spyder482> with 1/4 the number of pixels
[23:46] <spyder482> so half width
[23:50] <ibsteveog> wahoo!
[23:50] <ibsteveog> thanks a bunch
[23:50] <ibsteveog> a format that works
[23:51] <ibsteveog> although i haven't checked the color to make sure that U and V are in the right places...
[23:51] <spyder482> :)
[23:52] <spyder482> when it's interlaced...it's worse
[23:53] <spyder482> you have to use the first line of the U and V for the first line of Y but then use the second UV lines for the second line
[23:53] <spyder482> then the third line uses the first UV again and the fourht uses the secon UV
[23:53] <spyder482> real fun ;0
[23:54] <spyder482> [sigh]
[23:54] <spyder482> no matter what i do i can't get this app to actually do anything but play an mp3
[23:56] <ibsteveog> at least you are producing something =)
[23:56] <ibsteveog> and due to stupid testing on my part, i now have to revert to figure out what actually worked...
[23:57] walters ([email protected]) joined #gstreamer.
[23:57] <spyder482> well, i think BBB's matroska muxer is just broken...
[23:58] <spyder482> is there an ogg muxer for gstreamer?
[23:58] <spyder482> like to write vorbis ogg files?
[23:58] <spyder482> ah
[00:00] --- Sat Feb 21 2004
[00:01] teuf ([email protected]) left irc: Remote closed the connection
[00:02] <ibsteveog> ok, so with color its obvious that the UV part is not right...
[00:02] <spyder482> :)
[00:03] <spyder482> fourcc.org makes it sound like it's U then V
[00:03] <spyder482> yeah YV12 is V then U
[00:04] <ibsteveog> hmm... well, my colors are stretched towards the center of image (vertical stretching only)
[00:04] <spyder482> stretched?
[00:04] <ibsteveog> if they were just made taller.. or shorter.. i would add or drop vs... but stretched toward the center?
[00:05] <ibsteveog> yes, as in the closer they are to the top or bottom, the closer they are to where they should be, but items that should be 1/4 down the image are nearly at the center...
[00:05] <ibsteveog> meanwhile, some color that should be 3/4 down the image is just above the center
[00:05] <ibsteveog> and stretches nearly all the way back down
[00:05] markey ([email protected]) left irc: "nite"
[00:05] <spyder482> hrmmm
[00:05] <spyder482> can you gimme a pic?
[00:06] <ibsteveog> a single frame avi ok?
[00:06] <ibsteveog> mplayer test.avi -loop 0
[00:06] <spyder482> k
[00:07] <spyder482> hmm
[00:08] <spyder482> try again
[00:09] <spyder482> why does dcc never work when you want it to...
[00:09] <ibsteveog> one last try
[00:10] <spyder482> :/
[00:10] <spyder482> gotta http server?
[00:11] <ibsteveog> yup, just uploaded it to http://webspace.utexas.edu/~ogletree/
[00:11] <ibsteveog> ignore the other files =p
[00:12] <ibsteveog> the format on that one was a full Y plane, then full 1/2 U, then 1/2 V
[00:13] <spyder482> the planes are overlapped
[00:13] <spyder482> ermmm...
[00:13] <spyder482> hmm
[00:13] <spyder482> 1/2 U....
[00:13] <spyder482> how do you mean
[00:13] <spyder482> as in w/2 * h/2
[00:14] <spyder482> meaning 1/4 U?
[00:14] <ibsteveog> 1 U per 2 pixels
[00:14] <ibsteveog> so if there are 4 pixels.. i put 4 ys, 2 us, 2 vs
[00:14] <spyder482> oh
[00:14] <spyder482> that's wrong
[00:14] <ibsteveog> i also tried 4 ys, 1u 1v 1u 1v
[00:14] <ibsteveog> and 4ys 1u 1v
[00:14] <ibsteveog> i gotta retry them all again just to be certain
[00:15] <spyder482> ok, assume you have a 8x8 image, it would look like this:
[00:15] <spyder482> YYYYYYYY
[00:15] Last message repeated 7 time(s).
[00:15] <spyder482> UUUUVVVV
[00:15] Last message repeated 3 time(s).
[00:15] <spyder482> sorry for the flood
[00:16] <ibsteveog> oh geez.. i was alternating UV within each line..
[00:16] <ibsteveog> sigh
[00:16] <spyder482> :)
[00:16] <spyder482> that would be interleaved
[00:17] <spyder482> planar puts all in full squares
[00:19] <spyder482> \o/ I finally got something else to work
[00:19] <spyder482> mp3->ogg
[00:21] <ibsteveog> sadly, now the color is just really stretched vertically
[00:21] <spyder482> ?
[00:21] <spyder482> pic again? :)
[00:22] <ibsteveog> ok, i just replaced the one on the server... re-download it..
[00:23] <ibsteveog> and thats in the same format you just spammed :-)
[00:24] <spyder482> hmmm, you downsampled the U and V to 1/4 res?
[00:25] Action: ibsteveog checks
[00:25] Action: ibsteveog gets confused b/c of converting from UYVY
[00:26] <spyder482> UYVY is half horzontal res but not vertical IIRC
[00:26] <spyder482> so you have to reduce vertical res
[00:29] <spyder482> lmao @ stainless_steel_warning
[00:29] <ibsteveog> well, that seems to be backwards
[00:29] Zeenix ([email protected]) left irc: "The whole problem with the world is that fools and fanatics are always so certain of themselves, but wiser people so full of 
[00:29] <ibsteveog> yeah.. clubbers is not work safe tho
[00:29] <spyder482> not at work :)
[00:30] <spyder482> that's hilarious...
[00:30] <spyder482> funny what you find sometimes ;0
[00:30] <ibsteveog> hehe...
[00:31] <ibsteveog> well, we're back to that first stretching look again
[00:31] <spyder482> ?
[00:31] <ibsteveog> i put the first width/2 Us.. drop width/2...
[00:31] <ibsteveog> then width/2 Vs... then drop width/2 Vs...
[00:31] <spyder482> ummm
[00:31] <ibsteveog> i believe that would drop a line if its half-res horizontal
[00:32] <spyder482> i'm not sure that's smart but it shouldn't be too bad off i think
[00:32] <ibsteveog> where width=original image width
[00:32] <ibsteveog> well, i'm not interested in making it a smooth image.. just a correctly represented one
[00:32] <ibsteveog> once i find the right format, i'll worry about the rest =)
[00:33] <spyder482> update the avi?
[00:34] <ibsteveog> done
[00:35] <spyder482> hmmm
[00:35] <spyder482> seems you dropped too much
[00:36] <ibsteveog> hmm...
[00:36] <ibsteveog> i dropped an extra 2 lines
[00:36] <ibsteveog> and it actually lined up and was the right size
[00:36] <ibsteveog> just wrong colors
[00:36] <ibsteveog> want to see?
[00:36] <spyder482> yeh
[00:37] <ibsteveog> its replaced
[00:37] <ibsteveog> the red is more blue than the blue is..
[00:37] <ibsteveog> and thats the most terrible shade of green i've ever seen
[00:38] <spyder482> hmmm
[00:38] <spyder482> something is definitely wrong
[00:39] <spyder482> you do see that the rows from U and V are supposed to alternate right?
[00:39] <spyder482> well, that's silly
[00:39] <spyder482> if you didn't this would be much worse ;)
[00:40] <ibsteveog> well, i'm out of time..
[00:40] <ibsteveog> i gotta go dancin tonigh
[00:40] <ibsteveog> (and its mardi gras!)
[00:40] <spyder482> :)
[00:40] <spyder482> where you live?
[00:40] <spyder482> for mardi gras i figure close to La ;)
[00:41] <spyder482> or maybe france
[00:41] <spyder482> haven't heard many people in france talk about it though
[00:41] <spyder482> oh
[00:41] <spyder482> duh ;)
[00:41] <spyder482> texas
[00:41] <ibsteveog> yes =)
[00:41] <spyder482> i'm in Louisiana
[00:41] <ibsteveog> hmm
[00:41] <spyder482> :)
[00:41] <ibsteveog> maybe i should swap UV order?
[00:41] <ibsteveog> ah, some friends went over to new orleans
[00:41] <spyder482> which order do you have now?
[00:42] <spyder482> YV12 is V then U
[00:42] <spyder482> I420 is U then V
[00:42] <ibsteveog> oddly enough, swapping V and U didnt change a thing
[00:42] <spyder482> any gstreamer person awake?
[00:43] Action: ibsteveog begins to suspect his sanity
[00:43] <spyder482> ibsteveog: you sure you are telling the codec the right csp?
[00:43] <ibsteveog> csp?
[00:43] <spyder482> colorspace
[00:43] <ibsteveog> well, its reading a .yuv file and complains if i change the extension
[00:43] <ibsteveog> and says its taking a yuv 420P 
[00:44] <ibsteveog> ffmpeg: Video: rawvideo, yuv420p, 1200x1024, 1.00 fps
[00:44] <spyder482> hmmm, what's the original's colorspace?
[00:44] <ibsteveog> uyvy
[00:44] <ibsteveog> i am playing with some C code to convert the UYVY to the 420p
[00:45] <ibsteveog> just so i can figure out the right format
[00:45] <spyder482> there's some code in xvid that's very good at colorspace conversion
[00:45] <spyder482> even sme highly optimized assembly routines
[00:45] <ibsteveog> yes, but i need to know the format for a different project
[00:45] <ibsteveog> s/yes/perhaps
[00:46] <spyder482> yeah
[00:46] <spyder482> you sure you have the right order for the planes?
[00:47] <spyder482> erm, pixels
[00:48] <ibsteveog> i know my source is UYVY and I am dumping it out in order to 3 files, one for each of Y, U, V
[00:48] <ibsteveog> the Y file is my final file..
[00:49] <ibsteveog> so i just dump a line of U, a line of V, skip a line of each, then dump a line of U, a line of V.. skip... repeat...
[00:49] <ibsteveog> that gets the colors in the right place, but with the wrong colors
[00:49] <spyder482> hmmm
[00:49] <ibsteveog> nothing else has gotten the placement right
[00:49] <spyder482> swapping didn't help?
[00:49] <ibsteveog> actually, i'm skipping 3 lines
[00:49] <ibsteveog> and swapping did not change the colors
[00:49] <spyder482> hmmm
[00:50] <spyder482> maybe it expects the planes to be separate
[00:50] <spyder482> try putting all U then all V
[00:50] <spyder482> this could explain the staggered look...
[00:51] <spyder482> something's funny about how every 3rd line is different
[00:52] <ibsteveog> i agree, those horizontal striations are suspicious
[00:52] <ibsteveog> but i gotta go...
[00:52] <ibsteveog> thanks for the patience and help!
[00:53] ibsteveog ([email protected]) left irc: "quitting"
[00:55] mathrick ([email protected]) joined #gstreamer.
[00:56] <mathrick> yo
[00:57] <ds-work> moo
[00:58] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[00:58] Action: mathrick is happy about buing himself copy of "Mastering regular expressions" :)
[00:58] <ds-work> heh
[00:59] <mathrick> s/buing/buying/
[01:00] yippi ([email protected]) left irc: "Client exiting"
[01:08] Action: spyder482 needs a book like that
[01:18] mariano ([email protected]) joined #gstreamer.
[01:18] <mariano> if I configure gstreamer with --disable-docs-build (to work around a -V being passed to jw...) for some reason at install time it complains that GStreamer was configured using the --enable-plugin-builddir option.
[01:19] <mariano> it is me?
[01:19] <thaytan> are you running autogen?
[01:20] <mariano> yes
[01:20] <mariano> FWIW, I'm building with jhbuild, and the --disable-docs-build is in module_autogenargs['gstreamer']
[01:20] <thaytan> autogen does --enable-plugin-builddir unless you tell it not to
[01:21] <thaytan> add --disable-plugin-builddir to that
[01:22] <mariano> maybe that should be added to the jhbuild module spec for gstreamer...
[01:22] Action: mathrick notes using autogen.sh generated build is bad idea eveb with --disable-plugin-builddir
[01:22] <mathrick> s/eveb/even/
[01:22] <thaytan> it is in the jhbuild moduleset
[01:22] <thaytan> but you've overridden it :)
[01:22] <mathrick> it's best to run ./configure afterwards
[01:23] <thaytan> jhbuild actually has autogenargs="-- --disable-plugin-builddir --disable-tests" as the default
[01:23] <mariano> oh, module_autogenargs does not add... I see
[01:23] <thaytan> mathrick: not much choice when he's using jhbuild
[01:23] Action: mariano should know better
[01:24] Action: mathrick wonders why gaim shows "Not authorized" as status message for buddies on Jabber
[01:25] <mathrick> thaytan: I don't know jhbuild, but I know that ./autogen.sh --disable-plugin-builddirs build crashes badly for me
[01:25] <thaytan> mathrick: it needs a -- --disable....
[01:25] <thaytan> note the extra -- first
[01:25] <spyder482> what does cutter do actually? divide audio on silence boundaries?
[01:26] <thaytan> everything after that marker is a configure flag
[01:26] <mathrick> thaytan: probably, it's long time since I last did that :)
[01:30] ChrisHJW ([email protected]) joined #gstreamer.
[01:40] <ds-work> spyder482: yes.  but it may not work correctly
[01:40] <thaytan> good morning ds-work
[01:41] <ds-work> good evening thaytan
[01:42] <spyder482> hmmm, what's the simplest way to do two-pass normalization of audio with gstreamer?
[01:42] ain ([email protected]) left irc: "the end"
[01:43] <spyder482> is there a plugin that will find the peak value or something?
[01:43] <ds-work> no, but that would be a good addition to level
[01:44] <spyder482> i may try to make one myself one dasy
[01:44] <spyder482> day
[01:44] <spyder482> right now i'm still learning how to use plugins...much less write them :)
[01:56] pmfp ([email protected]) left irc: "Leaving"
[01:56] foser ([email protected]) left irc: "[ I want to believe ]"
[02:01] jcsston ([email protected]) joined #gstreamer.
[02:18] Misirlou ([email protected]) left irc: "asdf"
[02:19] <mathrick> nite
[02:19] <thaytan> g'night
[02:19] Nick change: mathrick -> mathrick|sleep
[02:29] brett_ ([email protected]) joined #gstreamer.
[02:30] brett_ ([email protected]) left irc: Client Quit
[02:33] Nick change: jimmy_dean_ -> jimmy_dean
[02:37] sub_pop ([email protected]) joined #gstreamer.
[02:40] ChrisHJW ([email protected]) left irc: Read error: 110 (Connection timed out)


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click