[Openal] multispeakers & multichannels newbie question.

[Garin Hiebert]

> - Which cards (and drivers) support multi-speakers output with openAL
> (any dsound or alsa compatible card, audigy, nforce ?) ? on which
> platforms ?

> - On these, how setup headphone, 2 speakers, 4 speakers or 5.1 rendering
>  output ? (.openalrc on linux and what on windows ?)

On Windows, any card which supports multi-speaker output will work as long
as the DirectSound 3D device can be used.  Of course, using the native
nForce or Audigy libraries will also work.  Selection of the speaker mode
is done using whatever mechanism is used with the drivers.  For instance,
on Creative cards you can manually select the mode on a speaker control
panel.

On MacOS, with the new code you can have multi-speaker output for any
device which supports it through Core Audio.  Like with Windows, the
speaker mode is selected outside OpenAL.

On Linux, multi-speaker is a configuration option as you noted.

> - Is the creative standard openAL window's driver support multi-speakers
>  output on one card ? (with software or hardware rendering)

The DirectSound3D device will support multi-speakers for any compliant
sound card (which should be any multi-speaker-capable sound card with the
latest drivers).  The native Creative library will also work for Audigy
and Audigy 2 boards.

> - Is the creative standard openAL window's driver support multi-speakers
>  output with multiple cards ? (with software or hardware rendering)

Probably not.  It's something that the code has been written for, but we
haven't done any testing -- so the chances of it working right now are
next to nothing.  You could open one context up using the DirectSound3D
device and the other on the native library, and that should work fine.

> How is rendering multi-channels audio file (mono,stereo, polyphonic,
> AC3) when they are associate to a source ? (like AL_FORMAT_STEREO16)
> What does it really mean to do it in a 3D rendering engine ?

On Windows, multi-channel audio files can be played, and the 3D
spatialization is automatically disabled (source and listener gain still
apply).

I'm not sure what is happening with the other two platforms right now with
multi-channel content, but regardless you are certainly best off making
sure such content is positioned at (0, 0, 0) in head-relative mode so that
you aren't suprised...

One more note -- this vague situation with multi-channel buffers will be
addressed in the 1.1 spec.


Garin