[00:06:55] *** prophile has quit IRC
[00:41:08] *** jvalenzu has joined #openal
[01:15:32] *** dra has joined #openal
[01:16:02] *** jvalenzu has quit IRC
[01:22:00] *** KittyCat has quit IRC
[01:26:29] *** P__ has quit IRC
[01:28:16] *** jvalenzu has joined #openal
[02:51:40] *** jvalenzu has quit IRC
[03:08:15] *** LtJax has quit IRC
[03:10:04] *** KittyCat has joined #openal
[03:10:08] *** ChanServ sets mode: +v KittyCat
[06:32:09] *** jvalenzu has joined #openal
[06:34:09] *** ChanServ sets mode: +v jvalenzu
[07:59:01] *** mattn2|home has joined #openal
[07:59:02] *** ChanServ sets mode: +v mattn2|home
[08:32:16] *** jvalenzu has quit IRC
[09:12:36] * KittyCat is away: sleep
[09:26:02] *** echelog has joined #openal
[09:26:03] *** ChanServ sets mode: +v echelog
[09:35:45] *** dra has quit IRC
[10:35:45] *** P__ has joined #openal
[10:35:45] *** ChanServ sets mode: +v P__
[12:42:21] *** LtJax has joined #openal
[12:42:22] *** ChanServ sets mode: +v LtJax
[13:29:45] *** mattn2|home has quit IRC
[13:58:25] *** mattn2|home has joined #openal
[13:58:25] *** ChanServ sets mode: +v mattn2|home
[14:03:27] *** LtJax has quit IRC
[15:39:53] *** jvalenzu has joined #openal
[17:10:02] *** jvalenzu has quit IRC
[17:11:48] *** mohij has joined #openal
[17:11:49] *** ChanServ sets mode: +v mohij
[17:18:01] *** P__ has quit IRC
[17:20:59] *** mattn2|home has quit IRC
[17:28:44] *** mattn2|home has joined #openal
[17:28:45] *** ChanServ sets mode: +v mattn2|home
[17:33:50] *** mattn2|home has quit IRC
[17:36:21] *** mattn2|home has joined #openal
[17:36:21] *** ChanServ sets mode: +v mattn2|home
[17:41:36] *** jvalenzu has joined #openal
[19:00:49] *** jvalenzu has quit IRC
[19:03:21] *** dra has joined #openal
[19:47:24] *** jvalenzu has joined #openal
[20:08:53] *** P__ has joined #openal
[20:08:53] *** ChanServ sets mode: +v P__
[20:16:47] *** P2__ has joined #openal
[20:18:56] * KittyCat is back.
[20:20:17] *** P2__ has quit IRC
[21:34:56] <P__> if i want to make a software renderer for audio, what I should use on Windows, the MCI ?
[21:37:52] <KittyCat> if you want to make a software renderer, you just have to do the mixing yourself
[21:39:33] <KittyCat> whether you output to mci or dsound doesn't matter
[21:41:55] <P__> what do you think about making a software renderer ? it looks like it's going popular
[21:42:34] <P__> i think fmod does that to have similar results on all platforms they support
[21:43:16] <KittyCat> the sample implementation is software
[21:43:31] <KittyCat> and the windows version has the Generic Software device, which is also software
[21:44:32] <P__> that's right
[21:44:58] <P__> but it's not abstracted in the way usually followed by other library, where you can make a new renderer easily isn't it ?
[21:45:39] <KittyCat> the sample implementation is easy to add backends to
[21:46:52] <P__> there is even a backend for window ?
[21:47:25] <P__> i see a file in the structure in the backend dir which let me think that
[21:47:41] <KittyCat> yeah. there's a waveout one, and I believe the dsound one I added
[21:48:25] <P__> btw, why it has been choosen to write it in C ?
[21:48:49] <KittyCat> probably because it's more portable.
[21:49:57] <P__> ok
[21:50:36] <P__> since everything from creative is closed source, that would be interesting to run this sample implementation on windows
[21:50:45] <P__> i'm sure someone already tried it
[21:51:53] <KittyCat> I've done it in wine, and it worked as you'd expect
[21:51:56] <P__> some backends look really simple, thank you for pointing me in that direction
[21:52:19] <P__> do you know if someone already used openal on consoles ?
[21:52:41] <KittyCat> no idea
[21:52:56] <P__> apart Xbox of course
[21:53:44] <P__> but in fact, the software renderer in windows in the installer provided by Creative is using the mmsystem implementation that is inside the sample implementation ?
[21:53:53] *** jvalenzu has quit IRC
[21:54:15] <KittyCat> no, it's using its own
[21:54:38] <P__> ...
[21:54:39] <KittyCat> the svn version of OpenAL includes OpenAL-Windows, which is the slightly dated code used in windows
[21:54:56] <P__> the code from Creative ?
[21:55:11] <KittyCat> an older version
[21:55:18] <KittyCat> hasn't been updated in a while, I don't think
[21:55:25] <P__> why have they began it open source and then close it..
[21:55:51] <KittyCat> because they added EAX2/EFX emulation to the software renderer, and they don't want anyone to see the code
[21:56:08] <P__> ok
[21:56:09] *** prophile has joined #openal
[21:56:10] *** ChanServ sets mode: +v prophile
[21:57:14] <P__> they better protect everything they can, there supremacy will end quite soon
[21:58:06] <KittyCat> are there newcomers in the consumer audio market?
[21:58:18] <KittyCat> that's really the only reason they're still around
[21:58:25] <KittyCat> no one's around to challenge them
[21:59:49] <KittyCat> consumer audio has been pretty stale because there's no competition to improve it
[22:00:16] <KittyCat> most audio you get is single-channel AC'97
[22:00:21] <P__> i haven't heard about any newscommer, but since the technology is more and more software, that could happen, no ?
[22:00:43] <KittyCat> creative is one of the few that gives you multiple channels (eg. hardware mixing) with hardware effects
[22:01:24] <KittyCat> and software isn't the place for audio, IMO
[22:01:42] <P__> more an more developper seem to think it is
[22:01:43] <KittyCat> it's fine as a fallback, but it should be done on dedicated hardware when possible
[22:02:18] <KittyCat> dsp effects aren't cheap. and physics cards haven't been catching on
[22:02:18] <P__> why ? I have been told there isn't somethign specific that can be accelerated by hardware more efficiently than a cpu
[22:02:36] <P__> why buying an additionnal card, better to upgrade a cpu, no ?
[22:02:50] <KittyCat> not that the cpu can't do it efficiently, but the cpu has other things to do, too
[22:02:53] <P__> it seems to be only marketting to sell audio cards that do hardware specific stuff
[22:02:58] <KittyCat> ai, physics, general processing, etc
[22:03:07] <P__> yes but it's like putting a small cpu near a big one
[22:03:44] <KittyCat> that's really all a gpu is, too
[22:03:51] <KittyCat> a gpu is just a dedicated cpu
[22:03:54] <P__> it's much more specific
[22:04:01] <KittyCat> just like an apu is a dedicated cpu
[22:04:52] *** LtJax has joined #openal
[22:04:53] *** ChanServ sets mode: +v LtJax
[22:05:00] <P__> what is an APU ?
[22:05:17] <P__> for graphic cards it's different, really
[22:05:21] <P__> it makes sense to have a GPU
[22:05:24] <KittyCat> audio processing unit
[22:05:27] <P__> with a bigger cpu it would be slower
[22:05:52] <P__> gpu memory is much faster, it's really something that has been wired together to do a task efficiently
[22:06:12] <P__> i'm not sure about what i'm telling, just to speak about it, it's a topic i have hear often about speaken
[22:06:53] <KittyCat> most of the things done in graphics rendering have audio equivilants. sound refraction, shadowing, occlussion, intesity, resampling, anti-aliasing, etc (plus others.. reverb, echo, chorus, etc).
[22:07:35] <P__> yes but isn't the amount of data needed to transfert much lower than for graphics ?
[22:07:46] <P__> i also thinks PPU is a joke
[22:07:52] <P__> well marketting
[22:08:12] <KittyCat> depends on how the data is constructed
[22:08:26] <KittyCat> 96khz sound would need more than 48khz
[22:08:40] <KittyCat> multiple samples would need more
[22:08:49] <KittyCat> plus oyu have the cost of the actual processing
[22:08:53] <P__> i have also read several discussion about that, people saying that audio hardware have a long road in front of them, and sound could be a domain to enhance a lot in game programming per instance
[22:09:05] <P__> and usually, when i say that, most developper answer : "what do you want more ?"
[22:09:26] <KittyCat> on top of the fact that OSs aren't really designed for the precise timing needed for audio. it'd be better to have dedicated hardware to handle that automagically, like we do with graphics.
[22:09:56] <P__> but for that there is a standard needed isn't it ?
[22:10:41] <P__> so far the main manufacturer is keeping its infos, and i have heard creative drivers on linux sucks
[22:10:53] <KittyCat> most commercial developers won't care. they only give what consumers ask for. and most consumers don't know better, so they don't ask.
[22:10:54] <P__> it looks like far from a standard
[22:11:15] <KittyCat> there really isn't a standard yet, which is a problem
[22:11:35] <KittyCat> there's standard apis (eg. openal), but no standard way for the effects
[22:11:57] <P__> that's explaining why fmod choosed to do it all in software
[22:12:11] <P__> one more reason, since they sell their lib for game dev on all platforms
[22:12:13] <KittyCat> not really
[22:12:33] <KittyCat> you can provide a general api for the effects, and offload it to hardware when you know you can
[22:12:49] <P__> but openal isn't doing that
[22:12:50] <KittyCat> but it'd have to program for each specific card that can handle it
[22:13:09] <P__> else i agree with you
[22:13:14] <KittyCat> openal on windows does do that
[22:13:29] <P__> i mean generally, there is no DSP chain like fmod does
[22:13:44] <P__> and i'm sure they use it for occlusion/obstruction internally as well
[22:13:58] <P__> *i would bet it, i mean.
[22:14:37] <P__> don't take me wrong, i find openal fantastic
[22:14:47] <P__> just like to discuss this topic since i'm quite new with it
[22:15:26] <KittyCat> openal could handle an effects api. the challenge would be writing up a clear extension and getting people to implement the extension.
[22:15:46] <P__> that could be done via an extension ?
[22:15:50] <P__> no refactoring ?
[22:15:56] <KittyCat> just like opengl. it's made to be extended
[22:16:27] <P__> yes but usually to implement an effect you have to process a signal, with samples in openal you think you could do it ?
[22:16:39] <KittyCat> in fact, there are already extensions to do it.. that's what eax and efx are for. the only problem is, they're creative-proprietary
[22:16:46] <P__> i was thinking it was in some cases uploaded to the hardware, so no way to control it in realtime
[22:17:07] <KittyCat> if the hardware can do it, there's no reason not to let it
[22:17:16] <P__> yes but efx is hardware dependant, you don't have all effects in software, and no way to add your own effects
[22:17:31] <KittyCat> if the hardware can't do it, then you either don't get it, or software emulation is provded
[22:17:44] <P__> yes but that means that with fmod, you can do your own effects, so everything is done in software, that's a pro, but i understand the con
[22:18:04] <P__> software emulation isn't provided for everything but reverb and low pass
[22:18:18] <P__> on windows.
[22:18:35] <KittyCat> iirc, eax2 also includes chorus
[22:19:08] <P__> ok
[22:19:55] <P__> i guess if there have been no concrete attempt to this effect extension we are speaking about, there is a reason
[22:20:09] <P__> either what you said above, and maybe at the same time thedifficulty
[22:20:47] <KittyCat> the only reason creative's the only one with the extension is because creative's really th e only one trying with openal.
[22:20:54] <P__> but why ?
[22:21:10] <P__> why fmod hasn't started with openal
[22:21:13] <KittyCat> the sample implementation is designed to be bare. something light and simple that can be ported to any conceivable platform
[22:21:33] <P__> that's the impression i have when i look at the source code
[22:21:46] <KittyCat> the mac version... well, that's apple. bug them
[22:22:13] <P__> the thing about openal, is it doesn't look very active from outside
[22:22:28] <P__> per instance i have no idea who is the project leader, what is currently developped, etc...
[22:23:02] <P__> do you have this impression too or is it just me ?
[22:23:15] <KittyCat> I get that impression, too.
[22:23:31] <P__> i don't know how it was several years ago
[22:23:58] <KittyCat> there've been times I felt like forking the sample implementation simply because I either didn't know who was in charge, or things were just moving too slowly because the person in charge wasn't very active
[22:25:20] <P__> that's not a good thing
[22:25:42] <P__> that's probably also due to the split between the open source part and the close source one
[22:26:17] <KittyCat> nah. that was fairly recent
[22:26:40] <KittyCat> the main reason is because the main head guy (Sven) seems to have just disappeared
[22:26:51] <P__> since how many time ?
[22:27:10] <KittyCat> Jon Daly seems to be the only other person around, and he's scarce himself
[22:27:42] <KittyCat> for quite a while, I think. Sven was pretty scarce when I got into developing for it
[22:27:56] <KittyCat> Jason Daly, not Jon
[22:28:35] <P__> i read an answer from him today on the list, i would have given a different one and indeed 2 other people gave a different one
[22:29:19] <P__> well i can be wrong, it's hard to judge, but it doesn't look professionnal as few other libraries i'm working with. Not the result, but how the development is going
[22:29:46] <KittyCat> for the open source version, definitely
[22:30:00] <KittyCat> open source development is rarely professional
[22:30:27] <KittyCat> doesn't necesarilly mean it's a bad thing, of course
[22:30:50] <P__> few projects are, like Ogre3D
[22:31:05] <KittyCat> I prefer a more human face on things, instead of a high-brow corporate face.
[22:31:17] <P__> that's 2 extremes, there are in between
[22:31:27] <P__> i don't have the feeling Ogre 3D is a high brow corporate face
[22:31:42] <P__> but it's still looking profesionnal
[22:31:54] <P__> guidelines, updates, mangement, plans, etc..
[22:32:22] <KittyCat> true. not that openal couldn't be better, but as I said, most of the head people are scarce or missing
[22:32:57] <P__> i don't say there is no work just to be clear, but by looking at the code of the sample implementation, it doesn't look like years of development
[22:33:21] <P__> it's sure a lot of work, but doesn't look as big as other engines i looked at, especially math and algo side
[22:33:46] <LtJax> so what's the problem u think? is it not widespread enough to attract devs?
[22:34:11] <P__> that's a bit what made me think why not trying to build a software renderer, just as an exercice and see where it goes, but not to try to replace the thing, it would make no sense
[22:34:18] <KittyCat> yeah. I too get the feeling that the people working on it didn't completely understand what they were doing. but that's how open source projects usually start
[22:34:25] <P__> devs are not attracted by projects which doesn't look active
[22:34:27] <KittyCat> you learn as you go
[22:35:06] <P__> of course
[22:35:21] <LtJax> P__, well yea - but there's not much point to reviving it if noone uses it
[22:35:32] <LtJax> not many apps are really using it
[22:35:35] <P__> but a bit of structure and organisation is needed, is was probably existing else it would not have reach this state, but nowdays it looks different for sure
[22:35:47] <P__> LtJax, i think a lot of games are using it
[22:36:04] <P__> but commercial companies which do their stuff in a corner without sharing
[22:36:10] <KittyCat> people do use it. especially on windows, where both the Doom3 engine, and the newest Unreal engine both use it. wouldn't doubt id Tech5 does, too
[22:36:15] <LtJax> just Epic as far as I know
[22:36:30] <LtJax> oh, Id too
[22:36:42] <LtJax> didn't know that
[22:37:35] <LtJax> so is it the only engine available for those on windows? because it was my impression it was usually used as backend for linux stuff... and audio is kinda... different there anyways
[22:38:14] <P__> with vista openal becomes even more interesting since there are no equivalent, for free at least
[22:38:29] <LtJax> yea, I read about that
[22:38:33] <KittyCat> I think I may just take the old windows code and port it to alsa/oss from dsound (taking out the dsound3d stuff). it'd at least give a nice common base for it. though unfortunately it'd lose 4-channel surround (one of the few things the current SI has right)
[22:39:07] <P__> KittyCat, you are tying to put a bit of time in it often ?
[22:39:12] <P__> time and effort i meant
[22:39:56] <KittyCat> yeah
[22:40:18] <KittyCat> the only reason I'm not doing anything right now is because I'm waiting for jason to check the new capture code I wrote
[22:40:54] <KittyCat> when that got in, I intended to put together a release, then work on cleaning out a lot of the useless gunk, and reworking the internals
[22:44:17] <P__> ok
[22:47:53] <P__> i m looking at the code, the si looks quite clean, altought i'm more used to C++ than C so it's hard to judge
[22:48:43] <KittyCat> a lot of the locking is confusing, and the actual mixing is horrid
[22:48:54] <P__> i think alc and al should be merge
[22:49:04] <P__> merged
[22:49:11] <KittyCat> there's a lot of useless tracking, and it can't properly handle multiple devices or contexts
[22:52:39] *** jvalenzu has joined #openal
[23:01:22] <P__> well, thanks for the discussion, always learning a lot with you !
[23:01:28] <P__> i'm off, see you
[23:01:31] *** P__ has quit IRC
[23:19:16] *** prophile has quit IRC
[23:38:30] *** jvalenzu has quit IRC

top