[00:24:19] *** juanmabc has quit IRC
[00:39:25] *** prophile has quit IRC
[01:09:09] *** mattelacchiato has joined #openal
[01:09:09] *** ChanServ sets mode: +v mattelacchiato
[01:09:16] <mattelacchiato> hi KittyCat
[01:09:23] <KittyCat> hi
[01:09:41] <mattelacchiato> have you tested oal-soft with pulseaudio?
[01:10:14] <KittyCat> not personally, but things I hear say it's not too good with the alsa plugin
[01:10:23] <KittyCat> might have better luck with the oss plugin
[01:10:54] <mattelacchiato> yes, but oss sucks-> only one playback in the whole system
[01:11:24] <KittyCat> pulseaudio's oss plugin lets multiple apps work together
[01:12:02] <KittyCat> but you're best off turning off pulseaudio, and using alsa's dmix device
[01:12:53] <mattelacchiato> hmmm, okay
[02:02:18] *** mattelacchiato has quit IRC
[02:18:49] *** Walt has quit IRC
[02:20:51] <qknight> KittyCat: i've managed to make a git push
[02:20:56] <qknight> KittyCat: still i can't checkout it
[02:21:43] <qknight> KittyCat: you should have told mattelacchiato to use my pulseaudio openal extension ;_)
[02:23:28] <KittyCat> didn't know there was anything publicly available yet..
[02:24:27] <qknight> KittyCat: never mind ;-)
[02:27:18] *** Walt has joined #openal
[02:29:05] <qknight> KittyCat: git clone http://lastlog.de/git/openal-soft-pulseaudio/
[02:29:38] <qknight> KittyCat: it's no official release since the files are just pushed in with no comments
[02:56:58] <qknight> KittyCat: you can now use that repo to build a pulseaudio client
[02:57:01] <qknight> KittyCat: wana try?
[02:59:33] <KittyCat> I cheked it out. though I can't build it since I don't have pulseaudio
[03:01:59] <Alam_Debian> could you put your name in author config?
[03:02:30] <Alam_Debian> Author:     Your Name Comes Here <you at yourdomain dot example.com> <- funny name
[03:16:22] <qknight> sure i can
[03:16:30] <qknight> this isn't meant to be official yet
[03:16:38] <qknight> i'm just setting up my git repo
[03:16:46] <qknight> and i'm very new to git so it took quite a time
[03:16:57] <qknight> but now as i start to understand it ...
[03:17:02] <qknight> will do that tomorrow ;-)
[03:17:08] <qknight> good night ;)
[03:48:26] * KittyCat is away: sleep
[08:01:19] *** jvalenzu has quit IRC
[13:40:59] *** Walt has quit IRC
[14:16:55] <qknight> KittyCat: good news
[14:17:00] <qknight> KittyCat: my public repo is nearly down
[14:17:03] <qknight> done ;-)
[14:17:07] <qknight> not down
[14:54:03] *** mattelacchiato has joined #openal
[14:54:04] *** ChanServ sets mode: +v mattelacchiato
[14:54:10] <mattelacchiato> hi KittyCat
[14:54:40] <qknight> mattelacchiato: hey!
[14:54:47] <qknight> mattelacchiato: you can test my pulseaudio backend for openal-soft
[14:54:50] <mattelacchiato> hi qknight
[14:54:55] <qknight> mattelacchiato: KittyCat didn't know that yesterday
[14:54:58] <qknight> mattelacchiato: wana try?
[14:55:03] <mattelacchiato> yeah!
[14:55:18] <qknight> mattelacchiato: ok, give me 10minutes to upload my current git clone
[14:55:35] <qknight> be warned: the latency is about 1-2 sec right now
[14:55:40] <mattelacchiato> uff
[14:56:06] <qknight> can be tuned and will be autotuned when glitch_free is in the official pulseaudio release
[14:56:41] <mattelacchiato> hmm, eigentlich hatte ich jetzt auch erstmal pulse abgestellt, damit ich hier weiter programmieren kann
[14:57:20] 
[14:58:07] <mattelacchiato> qknight: oder habe ich dich falsch verstanden?
[14:58:36] <qknight> mattelacchiato: nein hast du nicht. die latenz ist gerade sehr hoch
[14:58:43] <qknight> mattelacchiato: du willst es zum spielen?
[14:59:32] 
[15:00:05] <mattelacchiato> mein problem ist derzeit nur, dass ich oal-soft nicht dazu bewegen kann, alsa zu nutzen
[15:01:07] <mattelacchiato> wenn ich die default-rc nutze (alsa,oss,... in der treiber-reihenfolge), nimmt der immer oss. wenn ich nur alsa angebe, will er gar nicht laden...
[15:03:26] 
[15:03:33] <qknight> ja
[15:03:34] <qknight> verstehe
[15:03:44] <qknight> was willst du denn genau machen? einfach openal als backend einsetzen
[15:03:55] <qknight> und dann z.b. das spiel mit openal programmieren
[15:04:04] <qknight> dann geht das mit pulseaudio schon ganz gut
[15:04:25] <mattelacchiato> was meinst du mit openal als backend einsetzen, qknight?
[15:04:40] <qknight> mattelacchiato: und openal+alsa ist genau der grund warum ich mir die vielen stunden zeit genommen habe ein pulseaudio backend zu bauen
[15:04:51] <qknight> mattelacchiato: ok wozu willst du openal-soft nutzen?
[15:05:06] <mattelacchiato> ja, das spiel, was ich programmieren muss
[15:05:24] <qknight> ok dachte ich mir
[15:05:32] <qknight> mattelacchiato: und zum entwicken kann dr die latenz ja egal sein oder?
[15:06:55] 
[15:09:29] <mattelacchiato> du verstehst, qknight?
[15:10:03] <qknight> ja klar versthee ich das
[15:11:51] <qknight> sorry, also soweit ist es noch nicht
[15:12:23] <qknight> mattelacchiato: in deinem fall kannst du das sdl backend nehmen und versuchen ob du dmix irgendwie so hinbekommst
[15:12:25] 
[15:12:30] <qknight> oder mit openal-soft dmix reinmachen
[15:12:42] <qknight> habs aber damals auch nicht richtig hinbekommen und kann dir da folglich leider garnicht helfen
[15:12:56] <qknight> du musst halt alles ausmachen was alsa verwendet
[15:13:02] <qknight> dann geht es sicher
[15:13:13] 
[15:13:24] <mattelacchiato> jo, schon klar
[15:13:26] <qknight> du hast halt vermutlich eine recht schlechte soundkarte (ich tippe auf laptop/intel onboard im dekstop)
[15:14:00] <mattelacchiato> intel onboard, ja. schlecht==kein hardwaremixing, ja
[15:14:20] <qknight> ja du verstehst ja was ich meine ;-)
[15:14:36] <mattelacchiato> aber die frage ist weiterhing, wieso ich openal nicht dazu bringen kann, alsa zu nutzen!?!?
[15:15:08] <qknight> vielleicht hat ein prozess das alsa schon im betrieb
[15:15:14] <qknight> amarok/artsd z.b.
[15:15:17] <qknight> lsof hilft da witer
[15:15:22] <qknight> KittyCat: http://www.lastlog.de/wiki/index.php/Pulseaudio
[15:15:29] <qknight> KittyCat: i've uploaded it now
[15:16:39] <mattelacchiato> ja, amarok/artsd sind zwar an, aber das macht ja nix. deswegen will ich ja alsa nutzen
[15:16:48] <mattelacchiato> das lief ja auch schon mal..
[15:17:06] <qknight> ja, mag sein
[15:17:07] <mattelacchiato> selbst wenn ich alles ausmache, nimmt openal ja immernoch oss
[15:17:08] <qknight> ist halt alsa
[15:17:18] <qknight> aso jetzt verstehe ich dein problem erst
[15:17:21] * KittyCat is back.
[15:17:23] <qknight> hi KittyCat
[15:17:26] <mattelacchiato> hi KittyCat
[15:17:27] <KittyCat> hi
[15:17:43] <qknight> mattelacchiato: you need to set the right settings in ~/.openalrc
[15:17:52] <qknight> mattelacchiato: have you done that already? please nopaste your config
[15:18:07] <mattelacchiato> for now, it's the sample-rc
[15:18:14] <KittyCat> hmm, I'm not sure I'd agree with points 1 and 3 on that page..
[15:18:40] <KittyCat> the only reason it's platform dependant is simply because no one's made it for other platforms
[15:18:42] <qknight> KittyCat: 1. platform dependant (linux only)
[15:18:59] <KittyCat> the name may suggest it's linux-only, but there's no reason it can't work on other OSs
[15:19:13] <qknight> KittyCat: you mean alsa right?
[15:19:20] <KittyCat> yeah
[15:19:33] <qknight> KittyCat: ok, once a windows alsa driver exists i will change that
[15:19:35] <KittyCat> and, I'd say ALSA is the current defacto standard for linux audio, not the "future"
[15:19:53] <qknight> KittyCat: but since there are more than 20 drivers for linux and none for any other platform i state it as linux only
[15:20:35] <qknight> KittyCat: depends some think it's oss other think it's alsa
[15:21:02] <qknight> KittyCat: and it seems as if most toolkits do a high level abstraction because of thta problem
[15:21:19] <qknight> KittyCat: so it starts to be a problem of having backends since none likes to use it directly
[15:21:30] <qknight> KittyCat: phonon/xine-lib/...
[15:21:58] <qknight> KittyCat: ok, apart from this: what do you think about the code?
[15:22:57] *** jvalenzu has joined #openal
[15:23:23] * mattelacchiato is going for reboot
[15:23:28] *** mattelacchiato has quit IRC
[15:26:34] *** mattelacchiato has joined #openal
[15:26:34] *** ChanServ sets mode: +v mattelacchiato
[15:26:40] <mattelacchiato> re
[15:27:23] <qknight> hi
[15:27:45] <qknight> KittyCat: if alsa is so nice as you argue, why is mattelacchiato here then?
[15:28:19] <mattelacchiato> :-D
[15:28:45] <mattelacchiato> well everything was fine, since i installed kubuntu 8.04 with pulse :-D
[15:34:14] <KittyCat> seems (k)ubuntu making pulse the default has caused a few headaches for people
[15:34:58] <mattelacchiato> yeah, thats right
[15:35:29] <mattelacchiato> do you have again some minutes to help me, getting oal and alsa connected?
[15:35:56] <KittyCat> yeah
[15:36:55] <mattelacchiato> okay: using the default (or no) rc, oal want to use oss. when i set only alsa as driver, the startup fails, too
[15:37:27] <KittyCat> did openal soft build with alsa support?
[15:38:07] <mattelacchiato> hmmm. how can i test that?
[15:38:38] <mattelacchiato> or what have i to do when building openal-soft?
[15:38:40] <KittyCat> it would say when you ran cmake
[15:39:31] <KittyCat> you'll need the alsa-devel, or whatever it's called, package so openal soft can use the headers
[15:39:35] <qknight> KittyCat: changing to pulseaudio is not trivial but once done it's a nice thing to have
[15:40:08] <KittyCat> I just think they jumped the gun on it
[15:40:18] <KittyCat> they should've waited until pulseaudio's alsa plugin worked better
[15:40:19] <qknight> they juped the gun on it?
[15:40:41] <qknight> KittyCat: well you you always 'wait' things never get done ;-)
[15:40:42] <mattelacchiato> thats it
[15:40:56] <qknight> no
[15:41:02] <qknight> mattelacchiato: alsa headers aren't found
[15:41:05] <mattelacchiato> i'll try it with alsa headers
[15:41:06] <qknight> mattelacchiato: you have to check that
[15:41:08] <mattelacchiato> yes
[15:41:58] <KittyCat> qknight, not saying they shouldn't include pulseaudio. but making it the default was a mistake since it breaks a  lot of apps that use alsa
[15:42:14] <jvalenzu> alsa breaks a lot of apps that use alsa
[15:42:25] <KittyCat> :P
[15:42:29] <jvalenzu> if it goes a month without a major api change it's a miracle
[15:42:31] <qknight> jvalenzu: ;-)
[15:42:59] <KittyCat> alsa api has been stable since 1.0
[15:43:08] <jvalenzu> I'm sure it has
[15:43:14] <jvalenzu> I haven't looked at it for a while
[15:43:34] <qknight> KittyCat: can you give me a linux example program which uses openal sound input?
[15:43:37] <qknight> namely capture?
[15:43:46] <qknight> i think it's ut2004, isn't it?
[15:44:04] <KittyCat> that opens a capture device, yes. but I'm not sure of what extent it uses it
[15:44:05] <mattelacchiato> fyi: libasound2-dev is the wanted package called
[15:44:42] <mattelacchiato> and everything's fine! once again: thank you, KittyCat!
[15:45:00] <KittyCat> no problem :)
[15:46:55] <qknight> hehe ;-)
[15:49:10] <qknight> KittyCat: are you interested in testing the pulseaudio backend?
[15:49:25] <KittyCat> I don't have pulseaudio, so I can't
[15:50:39] <KittyCat> btw, I've been thinking about breaking out openal soft's mixer (where the backends do their main processing) into a seperate process, and letting apps talk to that process with IPC and shared memory (so multiple apps can share the same openal device with their own contexts)
[15:53:10] <qknight> what will be the main advantage?
[15:53:15] <qknight> less overhead?
[15:54:16] <KittyCat> less overhead when multiple apps use openal, being able to to renice the main mixing processing, being able to share a hardware voice, etc
[15:55:33] <qknight> ok, sounds nice but i don't think that there are so many usages of multiple openal instances
[15:55:45] <qknight> prove me wrong but openal is mainly used on games right?
[15:55:50] <KittyCat> yeah
[15:56:06] <jvalenzu> yeah!
[15:56:08] * jvalenzu high fives
[15:56:18] <KittyCat> I could envision an app spawning an mplayer process to play music
[15:56:24] <KittyCat> or for a video cutscene
[15:56:33] <jvalenzu> mixing daemons have problems
[15:56:34] <jvalenzu> latency for one
[15:56:44] <jvalenzu> aren't there a thousand of them anyway?
[15:57:03] <KittyCat> I'm not too sure how I could handle it efficiently.
[15:57:16] <jvalenzu> unix domain sockets would probably be fine
[15:57:26] <KittyCat> but the main thing is because openal would be used for games, you don't want much latency, so you'd generally want to give openal a hardware voice
[15:57:36] <jvalenzu> I thought oss etc allowed opening the sound device multiple times anyway
[15:57:44] <KittyCat> but at the same time, you don't want to completely block other apps
[15:57:52] <KittyCat> only if the hardware can handle it
[15:58:22] <KittyCat> if the hardware has multiple voices, you can open it multiple times. but if it only has one, it'll only open once
[15:59:51] <KittyCat> I can take advantage of the fact that openal is sorta designed around offloading processing (to hardware) anyway
[16:00:12] <KittyCat> I'm just not sure how I could deal with all the locking it would need to do, without stalling processing
[16:00:55] <jvalenzu> openal should really be abstracted from the backend
[16:01:35] <KittyCat> it is
[16:01:38] <jvalenzu> I know
[16:01:42] <jvalenzu> I mean, it should continue to be
[16:02:37] <KittyCat> it would. the thing that would change is how context (source, listener) updates can be reflected in the mixing process
[16:10:28] <qknight> KittyCat: i love openal as it is right now
[16:15:24] <qknight> KittyCat: while phonon and xine-lib is a high level abstraction i would say that openal-soft is one as well
[16:15:48] <KittyCat> somewhat
[16:16:08] <qknight> KittyCat: and for most systems it's ok to use the local backends as alsa/oss/pulseaudio ...
[16:16:25] <qknight> because this is the low level abstraction 'above' hardware
[16:16:44] <qknight> which is needed since alsa has some problems in that regard (as i mentation already)
[16:16:59] <qknight> KittyCat: btw: did you write all this code which is in openal-soft?
[16:17:17] <KittyCat> being that it's for performance/gaming, being able to optionally get direct hardware voice access, and be able to tweak the mixing priority/niceness, would be good
[16:17:35] <KittyCat> not all of it. it's based off the old windows code. though I've changed quite a bit of it
[16:17:52] <qknight> i think that you did a great job!
[16:18:16] <KittyCat> thanks :)
[16:18:55] <qknight> yes direct hardware voice access is a nice thing for games
[16:19:02] <qknight> KittyCat: does that still work on windows vista?
[16:19:09] <qknight> i bet no since it uses only software mixing
[16:19:59] <KittyCat> I doubt it. unless you have hardware openal drivers
[16:21:53] <qknight> having hardware openal drivers means not using openal-soft
[16:21:58] <qknight> is that why it's called soft?
[16:22:03] <qknight> since all processing is done in software?
[16:22:28] <KittyCat> yeah
[16:22:52] <qknight> ok
[16:23:10] <qknight> i personally think that it is the way to go: do everything in software -> do not rely on any hardware feature
[16:23:19] <qknight> since this is easy to support across several platforms
[16:23:37] <KittyCat> I'd liken it to opengl.
[16:23:43] <KittyCat> be able to do simple things in software
[16:23:44] <qknight> and since mose hardwares don't support it anyway (you told me already that this is not the case, still i don't understand it)
[16:23:58] <KittyCat> and have the options of all kinds of crazy capbilities with hardware
[16:24:14] <qknight> ah ok, now i understnad
[16:24:25] <qknight> so you would like to have a hardware vendor which supports openal in hardware
[16:24:40] <KittyCat> more than just creative, yeah
[16:25:06] <qknight> yes, that would be nice
[16:25:25] <qknight> but since cpu prize drops so much i don't see the need anymore
[16:25:36] <qknight> all new systems are likely to have 4cpus
[16:25:43] <qknight> which can do that
[16:25:52] <KittyCat> extra cpus/cores aren't the answer to everything
[16:26:15] <KittyCat> the gpu field is flourishing despite all the extra cpu power we're getting
[16:27:00] <qknight> KittyCat: that's true for normal computers but the ps3 is different
[16:27:13] <KittyCat> audio is another thing that would be best done off the main board
[16:27:27] <KittyCat> not taking up precious cache, or motherboard RAM
[16:27:35] <qknight> oh wrong
[16:27:36] <qknight> not ps3
[16:27:44] <KittyCat> ps3's different
[16:27:53] <KittyCat> it's a completely new architecture
[16:27:58] <qknight> ok, then it as ps3
[16:28:09] <qknight> and i bet ps3 is the way to go in future
[16:28:42] <KittyCat> not for desktop PCs, I don't think. the x86 is far too engrained to be able to replace
[16:28:58] <qknight> yes, maybe ;-)
[16:29:00] <qknight> we'll see ;-))
[16:29:17] <KittyCat> it could be done, but it'd have to be wicked cool, with complete x86 compatibility
[16:29:33] <qknight> KittyCat: but once intel starts to put it's own 3d processing untis insted of external gpus as ati/nvidia that wil change
[16:29:53] <KittyCat> Intel's Itanium tanked because it broke x86 compatibility
[16:30:23] <jvalenzu> 99% of processing is better on the cpu
[16:30:50] <jvalenzu> and the 1% that is better with dedicated hardware you can't use anyway because most sound hardware sucks anyway
[16:31:21] <jvalenzu> most of the magic with (say) Aurreal 3d was done in the driver anyway
[16:32:10] <KittyCat> hardware would have a better time applying effects and doing the actual mixing (channel conversion, resampling, etc) than the CPU
[16:32:38] <jvalenzu> mixing
[16:32:51] <jvalenzu> mixing is raelly well suited to general purpose SIMD processors
[16:32:54] <KittyCat> and being able to have the non-streaming audio stored seperately so it doesn't have to run over the FSB would be a nice boost as well
[16:33:46] <qknight> yes i see the points with mixing applying effects
[16:33:50] <jvalenzu> how much memory do PC games reserve for non-streaming sounds anyway?
[16:33:54] <jvalenzu> DSP filters, sure
[16:34:03] <jvalenzu> but how many cards expose a (non-EAX) api for that anywy?
[16:34:30] <jvalenzu> look, obviously an SPU or DSP is going to be more efficient
[16:34:31] <KittyCat> dsound can work like that
[16:34:34] <qknight> i'm on linux so it's basically none
[16:34:41] <jvalenzu> but most sound hardware is total crap
[16:34:47] <jvalenzu> there's no market in advanced features
[16:34:54] <KittyCat> alsa is able to store stuff on hardware buffers
[16:35:10] <KittyCat> currently, yeah. the sound market is pretty crappy
[16:35:45] <KittyCat> not because it can't be a good market. but because of some players, who shall go nameless..
[16:42:23] <qknight> lol
[16:42:40] <qknight> i don't understand what you could mean by players which shall go nameless
[16:42:44] <qknight> but sounds like a lobby to me
[16:43:10] <KittyCat> companies in the sound market
[16:43:21] <KittyCat> particularly the consumer sound market
[16:43:25] <qknight> KittyCat: ok
[16:43:28] <qknight> i know who is meant
[16:43:46] <qknight> KittyCat: why isn't openal soft included in every standard linux distribution?
[16:44:09] <qknight> gentoo for instance doesn't even have a ebuild in portage
[16:44:19] <KittyCat> because people are afraid of what it could break
[16:44:31] <KittyCat> not all apps that were made for the SI will work with openal soft
[16:44:39] <qknight> ok name one which doesn't
[16:45:11] <KittyCat> some rely on extensions that aren't supported anymore. others may rely on bugs/improper behavior the SI allowed
[16:45:21] <qknight> and what will break? does the application stop working at all or does it just miss some features
[16:45:33] <qknight> KittyCat: ok
[16:45:39] <qknight> have to go now
[16:45:41] <qknight> nice talk
[16:45:46] <qknight> cya later ;-)
[16:45:47] <KittyCat> can be anything from missing sound effects, to no sound at all
[16:45:50] <KittyCat> cya
[17:33:45] *** prophile has joined #openal
[17:33:45] *** ChanServ sets mode: +v prophile
[18:05:31] *** Walt has joined #openal
[18:05:34] *** mattelacchiato has quit IRC
[20:24:05] *** juanmabc has joined #openal
[20:24:05] *** ChanServ sets mode: +v juanmabc
[21:07:58] *** Walt_ has joined #openal
[21:11:50] *** Walt has quit IRC
[21:37:55] *** prophile has quit IRC
[23:52:11] *** prophile has joined #openal
[23:53:05] *** ChanServ sets mode: +v prophile





top