OpenAL?

[Chris]

Hello.

I apologize in advance if this is rude of me. I didn't find anything in a search of the forum.

I'm a developer of OpenAL Soft, an open source, cross-platform software implementation of OpenAL. I ran across Nexuiz recently after it was mentioned in a Linux Gamers article for free software shooters. Though I've actually heard about it before, I didn't really try it until recently. I must say I'm pretty impressed.. I generally don't go for deathmatch-only type games, but this is quite enjoyable, along with being visually pleasing.

However, I'm curious about the audio. It seems, via a quick glance at the source files, that the engine uses SDL, or in some cases ALSA, OSS, etc, directly. Have you given thought to using OpenAL? In case anyone's unfamiliar with it, it's basically the audio equivilant of OpenGL.. it's an open cross-platform API for playing sounds in a 3d environment, while also having capabilities to stream audio and play localized sound. Many functions are designed after OpenGL's API, it uses a similar naming scheme, and it even uses the same coordinate system. It can even do audio capture.

OSX comes preinstalled with Apple's OpenAL implementation, Windows has drivers from Creative (fully redistributable with apps) or from a few of the cross-platform implementations floating around, and Linux, *BSD, and other Unix-alikes also use said cross-platform implementations. On some systems, it can be hardware accelerated.

As far as OpenAL Soft is concerned, it can be used on just about any system that has ALSA or OSS (it can also be used on Windows via DirectSound). It supports 4 channel surround sound, along with all the basic effects (distance attentuation, doppler shift, directional audio, source cones). Future releases will hopefully see 7.1 surround sound, along with surround sound input (so you can playback 7.1 or 5.1 audio streams). There's even hope on supporting EFX, which can allow for richer environmental audio (it can handle things like air absorbtion, air density, and area environments, using automatic low-pass filters, reverb effects, etc). For those willing to use GIT, the OpenAL Soft repository includes a working branch with (incomplete) 7.1 surround sound and EFX support.


Unfortunately, I don't think I'll be able to do much actual coding for the Nexuiz project, although I'll be more than willing to help teach how the OpenAL API works. You guys have a pretty nice game and engine here, and I think this would be a nice compliment on the 3d audio side.


Once again, I apologize if this comes off as rude. Thanks for your time and attention.

[Qantourisc]

Darkplaces uses some insane quake mixer
Very strange one imo.
With lots of hacking.
The original design was rather clean.
But it still feels like an implementation oriented sound system.

In concern with OpenAL .... i'm also making a sound system. But nothing like OpenAl ... Sound like you have progressed a bit further then me .... /me cursus his time constrains.

As for darkplaces .... it has NO openal thing in it.
What's your timezone plz ? And what means can I use to contact you ?
OpenAL is such a mess imo ...

[nexbender]

desu

[iSuck]

We need audio streaming. Someone said that that's how you voice chat, and we all know that typing isn't exactly the most efficient communication.
I'm still a little confused as to what audio drivers do. What makes one better than another. With graphics it's easy to tell (you see reflections, textures) but with audio, I don't understand what drivers improve.

[KadaverJack]

We need audio streaming. Someone said that that's how you voice chat, and we all know that typing isn't exactly the most efficient communication.

This is not about voice chat. All OpenAL does is output sound or capture input from microphones, it doesn't implement a bandwidth efficient audio codec or any kind of network communication to my knowledge.

I'm still a little confused as to what audio drivers do. What makes one better than another. With graphics it's easy to tell (you see reflections, textures) but with audio, I don't understand what drivers improve.

This is not a driver either. It's sort of an abstaction layer between the application and various drivers that allows developers to write audio code that will work on different platforms. Right now darkplaces uses either sdl (for the sdl binaries, which also use sdl for input & opengl) or an native implementation for each platform (which could be replaced/unified with openal).

[Chris]

We need audio streaming. Someone said that that's how you voice chat, and we all know that typing isn't exactly the most efficient communication.

This is not about voice chat. All OpenAL does is output sound or capture input from microphones, it doesn't implement a bandwidth efficient audio codec or any kind of network communication to my knowledge.

I'm pretty sure IMA ADPCM is available on most known implementations. aLaw and uLaw are also available via FreeALUT, or sometimes directly in OpenAL itself. Although it doesn't provide methods for actually compressing to such formats, nor does it do networking for you.

Lord Havoc is the engine dev and best bet for a discussion of this.

Thanks. Does he read these forums, sometimes? If not, I'll throw an email his way.

[KadaverJack]

Thanks. Does he read these forums, sometimes? If not, I'll throw an email his way.

From time to time he does... the best method to get a hold of him is via irc (#darkplaces or #alientrap on irc.anynet.org)

[Qantourisc]

We need audio streaming. Someone said that that's how you voice chat, and we all know that typing isn't exactly the most efficient communication.
I'm still a little confused as to what audio drivers do. What makes one better than another. With graphics it's easy to tell (you see reflections, textures) but with audio, I don't understand what drivers improve.

There are several things that soundscards COULD do, some do, some dont, and some do this is software, and other do this in hardware.
And you are right there is no improvement ... someone made a monopoly ... (not naming names).
But what sounds cards COULD do:
- doppler / pitch shift
- sound reflections
- good resampeling between differnt mhz
- echos
- sound bleeding
- sound object occlusion
- resonance of sound stuff
And the nice thing of it all ... it takes about 1/100 of the calculation power gpu's have .... So there is no TECHNICAL reason we don't already have this.
That's why I started my sound thing.
I don't know what Chris intends to to entirely yet. But #1 on his list is making a portable OpenAL implementation.